The OIP Governance Question Ledger — Machine Artifact v1 (164 questions)

{
 "schema": "oip-governance-question-ledger/v1",
 "issued": "2026-07-17",
 "issuer": "Kimi K3 swarm (Moonshot AI), incognito, OPERATION CONSTITUTIONAL CARTOGRAPHY",
 "hash_recipe": "per question: sha256 over canonical JSON (sorted keys, compact separators, UTF-8) of the question object excluding the hash field",
 "statuses": [
  "IMPLEMENTED",
  "DOCUMENTED_UNENFORCED",
  "PARTIAL",
  "IMPLIED",
  "OPEN",
  "CONTRADICTED",
  "MISSING",
  "NON_AUTOMATABLE",
  "LEGAL_REVIEW_REQUIRED"
 ],
 "counts": {
  "questions": 164,
  "by_status": {
   "OPEN": 36,
   "IMPLIED": 18,
   "CONTRADICTED": 15,
   "MISSING": 40,
   "PARTIAL": 28,
   "LEGAL_REVIEW_REQUIRED": 12,
   "NON_AUTOMATABLE": 9,
   "DOCUMENTED_UNENFORCED": 5,
   "IMPLEMENTED": 1
  },
  "layers": 36,
  "dependency_edges": 403,
  "candidate_protocol_objects": 328
 },
 "dependency_edges": [
  {
   "from": "govq-const-012",
   "to": "govq-const-001"
  },
  {
   "from": "govq-const-016",
   "to": "govq-const-001"
  },
  {
   "from": "govq-const-018",
   "to": "govq-const-001"
  },
  {
   "from": "govq-const-006",
   "to": "govq-const-002"
  },
  {
   "from": "govq-const-009",
   "to": "govq-const-002"
  },
  {
   "from": "govq-const-002",
   "to": "govq-const-003"
  },
  {
   "from": "govq-const-006",
   "to": "govq-const-004"
  },
  {
   "from": "govq-const-005",
   "to": "govq-const-004"
  },
  {
   "from": "govq-const-016",
   "to": "govq-const-005"
  },
  {
   "from": "govq-const-004",
   "to": "govq-const-005"
  },
  {
   "from": "govq-const-007",
   "to": "govq-const-006"
  },
  {
   "from": "govq-const-023",
   "to": "govq-const-006"
  },
  {
   "from": "govq-const-006",
   "to": "govq-const-007"
  },
  {
   "from": "govq-const-011",
   "to": "govq-const-007"
  },
  {
   "from": "govq-const-008",
   "to": "govq-const-007"
  },
  {
   "from": "govq-const-006",
   "to": "govq-const-008"
  },
  {
   "from": "govq-const-007",
   "to": "govq-const-008"
  },
  {
   "from": "govq-const-002",
   "to": "govq-const-009"
  },
  {
   "from": "govq-const-007",
   "to": "govq-const-009"
  },
  {
   "from": "govq-const-006",
   "to": "govq-const-009"
  },
  {
   "from": "govq-const-006",
   "to": "govq-const-010"
  },
  {
   "from": "govq-const-002",
   "to": "govq-const-010"
  },
  {
   "from": "govq-const-007",
   "to": "govq-const-011"
  },
  {
   "from": "govq-const-006",
   "to": "govq-const-011"
  },
  {
   "from": "govq-const-001",
   "to": "govq-const-012"
  },
  {
   "from": "govq-const-013",
   "to": "govq-const-012"
  },
  {
   "from": "govq-const-014",
   "to": "govq-const-012"
  },
  {
   "from": "govq-const-012",
   "to": "govq-const-013"
  },
  {
   "from": "govq-const-008",
   "to": "govq-const-013"
  },
  {
   "from": "govq-const-004",
   "to": "govq-const-013"
  },
  {
   "from": "govq-const-012",
   "to": "govq-const-014"
  },
  {
   "from": "govq-const-013",
   "to": "govq-const-014"
  },
  {
   "from": "govq-const-016",
   "to": "govq-const-015"
  },
  {
   "from": "govq-const-017",
  

{
 "question_id": "govq-const-001",
 "layer": "constitutional foundations & axioms",
 "status": "OPEN",
 "question": "Does the published kernel hash pin the semantics of the five axioms, or only their bytes — and who adjudicates when two independent implementations read the same axiom (e.g. 'authority that cannot widen itself') differently?",
 "why_it_matters": "The kernel is pinned by SHA-256 (210f5feddc3d5b0a...) so two parties can confirm they quote the same constitution byte-for-byte, but bytes do not fix meaning. 'Bounded', 'explicit authority', 'widen', 'canonical bytes', 'distinguishable and auditable' are all interpretive surfaces. If two implementations both accept the hash and diverge on reading, the federation has interoperability of citation without interoperability of obligation — the exact failure a constitutional court exists to absorb, and no such court is defined.",
 "candidate_answers": [
  {
   "answer": "Hash pins bytes; a versioned interpretation layer pins semantics",
   "type": "protocol_object",
   "rule": "The core hash attests textual identity only; axiom meaning is fixed by a separate, separately hashed definitions-and-interpretation registry whose version the subscription record must also name.",
   "candidate_protocol_object": "constitutional_definitions_ledger + interpretation_ruling (kind=interpret) records appended under the same append-only law as rulings",
   "candidate_runtime": "subscription intake validates both core.hash and definitions_version; interpreters publish interpret records linked parent_id to the disputed record",
   "candidate_conformance": "new clause: two reference implementations executing the same worked example must produce the same verdict class under a named definitions version",
   "candidate_article": "/a/oip-constitutional-terms (per-term definition, version, falsifier)"
  },
  {
   "answer": "Hash pins bytes only; interpretation stays per-implementation (deliberate minimalism)",
   "type": "normative",
   "rule": "The kernel is an interoperability floor, not a court; divergent readings are resolved by exit, not adjudication — a node that disagrees forks and anchors.",
   "candidate_protocol_object": "none beyond existing fork_anchor (kind=anchor)",
   "candidate_runtime": "none; divergence is detected socially and expressed as fork records",
   "candidate_conformance": "none; conformance suites remain per-operator",
   "candidate_article": "boundary article stating the kernel refuses interpretive jurisdiction"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core block, hash, axioms, boundary)",
  "recon2/governance2.json",
  "recon2/specification.md §2.1-2.4 (axioms wired to conformance clauses C11-C33)",
  "corpus/bundle_constitution.md (article-ledger constitution)"
 ],
 "evidence": [
  "core.version 'oip-governance/1' + hash are the only pinned fields; no definitions registry or interpretation record kind exists in participation.kinds (subscribe, inquire, propose, feature, conformance, anchor, appeal, ruling)",
  "v1.1 spec §2.4 maps each axiom to executing clauses, i.e. the operator's own conformance suite currently IS the de facto interpretation of the axioms",
  "axiom texts are one sentence each with undefined operative terms ('explicit authority that cannot widen itself'; 'canonical bytes'; 'distinguishable and auditable')"
 ],
 "counterevidence": [
  "The boundary clause narrows the kernel to interoperability, which supports the reading that semantic adjudication was refused by design; append-only lineage plus fork anchors supply an exit-based dispute path that some theorists (Nozick-style exit) consider sufficient"
 ],
 "dependencies": [
  "govq-const-012",
  "govq-const-016",
  "govq-const-018"
 ],
 "falsifier": "Observation that two conforming implementations produce opposite verdicts on the same receipt under the same core hash, with no interpret records to reconcile them, confirms the gap is live rather than theoretical.",
 "notes": "The hardest single question on this plane: textual pinning is solved, semantic pinning is not, and every later layer (amendment, precedent, versioning) inherits the answer.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:1efee073b2da1e4bac2c763fd02dcbcf3d20a6d85099d39dad6f6cb18c620e5d"
}

{
 "question_id": "govq-const-002",
 "layer": "constitutional foundations & axioms",
 "status": "IMPLIED",
 "question": "What is the constitutional rank of the boundary clause ('The kernel governs interoperability...') — can it be waived, expanded, or breached by the root, and by what procedure?",
 "why_it_matters": "The boundary clause is the constitution's self-limiting text: it promises subscribers that alignment never imports philosophy, full-facet adoption, social publication, hosted execution, or credential surrender. If the root can unilaterally rewrite it, every subscription made in reliance on it is revocable at will; if it is entrenched, it constrains even a future amendment. Today its rank is implied by placement in the core block but nowhere stated as entrenched, waivable, or amendable.",
 "candidate_answers": [
  {
   "answer": "Entrench the boundary clause as non-derogable within the kernel",
   "type": "normative",
   "rule": "The boundary clause binds every future core version; an amendment that narrows it is void ab initio and any node may treat the amending registry as a different constitution.",
   "candidate_protocol_object": "entrenchment flag inside core block: {\"boundary\": {\"entrenched\": true}} plus amendment_rule field naming the clause untouchable",
   "candidate_runtime": "intake refuses kind=subscribe records quoting a core hash whose boundary differs from the entrenched text",
   "candidate_conformance": "clause: a mocked core with altered boundary must fail subscription validation",
   "candidate_article": "/a/oip-kernel-boundary explaining entrenchment"
  },
  {
   "answer": "Boundary is ordinary kernel text, amendable like any axiom",
   "type": "normative",
   "rule": "No clause is privileged; the amendment rule (once it exists) may change the boundary by the same supermajority as any axiom.",
   "candidate_protocol_object": "none beyond the amendment machinery of govq-const-006",
   "candidate_runtime": "core versioning endpoint serves sequential boundary texts with hashes",
   "candidate_conformance": "diff-test between core versions",
   "candidate_article": "amendment-law article"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core.boundary)",
  "recon2/governance2.json",
  "recon2/specification.md §2.3 (boundary as self-limiting clauses)"
 ],
 "evidence": [
  "The clause is published inside the hashed core block, so it is constitutional text, but the schema contains no entrenchment marker, no amendment rule, and no statement of rank relative to facets or participation laws",
  "v1.1 spec §2.3 reads the clause as operative on live records (two live filings quote the kernel while carrying accepted_core:false and remain stored, i.e. the boundary is currently honored)"
 ],
 "counterevidence": [
  "Nothing in the live registry states the boundary is entrenched; v1.1 §3.4.9 power 7 notes all registry limits are 'the author's chosen limits', changeable in a future core version with no external ratification process visible"
 ],
 "dependencies": [
  "govq-const-006",
  "govq-const-009"
 ],
 "falsifier": "If a future core version publishes an altered boundary and subscriptions continue to be accepted against it without any ratification record, the clause was never entrenched — the question collapses into govq-const-006.",
 "notes": "Rank-of-norms question; pairs with govq-const-005.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:a5d5e41324c5bb53d3a119ac8aa71e5aa6357b00364bbb7861f7a41d8e07394b"
}

{
 "question_id": "govq-const-003",
 "layer": "constitutional foundations & axioms",
 "status": "CONTRADICTED",
 "question": "Can a party that files accepted_core:false nonetheless occupy the constitutional plane as a counted 'node' — and should it?",
 "why_it_matters": "required_for_subscription:true makes the kernel mandatory, yet the live registry stores, publishes, and census-counts filings that refuse it: gov_baecae7c314944fda77c (feature) and gov_d3eb38a0f49b404d9733 (propose) both carry accepted_core:false, and both labels increment non_owner_node_count. The contradiction is between the subscription gate (kernel required) and the census law (excludes only system/owner-authorized filings). A constitution that counts refusers as members has no settled rule of recognition.",
 "candidate_answers": [
  {
   "answer": "Census only core-accepting records; quarantine refusers",
   "type": "runtime",
   "rule": "Only records with accepted_core:true and kind=subscribe may move any node census; accepted_core:false filings remain visible as discourse but never as membership.",
   "candidate_protocol_object": "census_law amendment + machine field member:false on refusing records",
   "candidate_runtime": "recompute counts over filtered set; publish both member_census and discourse_census",
   "candidate_conformance": "falsifier test from Part E: a filing with accepted_core:false must not move member census",
   "candidate_article": "registry-law revision note"
  },
  {
   "answer": "Treat refusal as a valid constitutional act (opposition with standing)",
   "type": "normative",
   "rule": "Filing a refusal is participation in the discourse plane, not membership; the constitution should name it 'contestation' and give it its own counted class rather than pretending it is a node.",
   "candidate_protocol_object": "kind=contest record (or accepted_core:false interpreted as contest) with its own tally",
   "candidate_runtime": "intake re-labels refusing filings; counts.contesting_labels computed separately",
   "candidate_conformance": "none beyond count separation",
   "candidate_article": "/a/oip-contestation-plane"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (counts, census_law, required_for_subscription)",
  "https://miscsubjects.com/api/governance/record/gov_baecae7c314944fda77c",
  "https://miscsubjects.com/api/governance/record/gov_d3eb38a0f49b404d9733",
  "killbox-specification-v1.2.md Part E Finding 1-2"
 ],
 "evidence": [
  "core.required_for_subscription is true; census_law excludes only 'system and owner-authorized filings', not accepted_core:false records",
  "v1.2 Part E Finding 1: 'two of the four filings carry accepted_core: false ... yet both labels increment the census of nodes of a protocol whose core they did not accept'",
  "live read 2026-07-17: total 5, non_owner_node_count 4, non_owner_anchor_count 2"
 ],
 "counterevidence": [
  "The registry never claims the census measures members — census_law defines it as an 'actor-label census', and the records' own law field denies authority grant; one can read the contradiction as a labeling bug rather than a constitutional defect"
 ],
 "dependencies": [
  "govq-const-002"
 ],
 "falsifier": "A published rule stating whether accepted_core:false records are members, with the census recomputed accordingly; until then the contradiction stands between core gate and census law.",
 "notes": "Conflict between sources kept visible per mission law; not smoothed.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:d813f54bd01bfde40fd4bfb5301d714abdb10cff7b10259371a2f23964b4dd80"
}

{
 "question_id": "govq-const-004",
 "layer": "constitutional foundations & axioms",
 "status": "MISSING",
 "question": "Is the five-axiom set a closed enumeration (the whole constitution) or a floor — i.e., is adding a sixth axiom an amendment of this constitution or the founding of a new one?",
 "why_it_matters": "v1.1 spec §2.1 states 'the five-axiom set as published is the whole constitution: no sixth axiom exists in the registry, and none of the eleven facets adds to the kernel.' If the set is closed, any sixth axiom is a constitutional revolution requiring a new hash and re-founding consent; if it is a floor, expansion is ordinary amendment. The difference decides whether an 'amended' kernel can claim continuity with the subscriptions already filed against 210f5feddc3d... — a Kelsenian basic-norm question the registry does not answer.",
 "candidate_answers": [
  {
   "answer": "Closed set: axiom addition is refounding, not amendment",
   "type": "normative",
   "rule": "The axiom count is constitutive; a kernel with more or fewer axioms is a new constitution with a new hash, and prior subscriptions carry over only by explicit re-acceptance.",
   "candidate_protocol_object": "core_version_registry listing each core hash as a distinct constitution with re-acceptance records (kind=subscribe naming new hash)",
   "candidate_runtime": "intake treats unknown core hashes as foreign constitutions; no grandfathering",
   "candidate_conformance": "clause: a sixth-axiom core must not validate against subscriptions naming the five-axiom hash",
   "candidate_article": "/a/oip-kernel-closure"
  },
  {
   "answer": "Floor: axioms may be added by supermajority amendment",
   "type": "normative",
   "rule": "The set is amendable; additions require the highest ratification class and cannot repeal the boundary clause.",
   "candidate_protocol_object": "kernel_amendment_record with axiom_delta field",
   "candidate_runtime": "core endpoint serves versioned axiom lists with per-version hashes",
   "candidate_conformance": "version diff test",
   "candidate_article": "amendment-law article"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core.axioms)",
  "recon2/specification.md §2.1 ('the five-axiom set as published is the whole constitution')",
  "nogo-n03 (self-reference limits of formal systems)"
 ],
 "evidence": [
  "Registry publishes exactly five axioms and one hash; no closure clause, no amendment clause, no statement whether the set is constitutive",
  "The version string 'oip-governance/1' implies successors were anticipated, which sits in tension with a strictly closed reading"
 ],
 "counterevidence": [
  "The presence of a version field and of version-pinned subscription records (each live record quotes version+hash) implies the designers expected later cores — weak evidence for the floor reading"
 ],
 "dependencies": [
  "govq-const-006",
  "govq-const-005"
 ],
 "falsifier": "Publication of any closure or amendment clause for the axiom set resolves this; until then both readings are compatible with all live evidence.",
 "notes": "Constituent-power question; deliberately not resolved here.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:a227eab9493b75111662957084ea23041e721e778537c06effb020964f08a462"
}

{
 "question_id": "govq-const-005",
 "layer": "constitutional foundations & axioms",
 "status": "PARTIAL",
 "question": "What is the hierarchy of norms among the kernel, the facets, the participation laws, the drop document prose, the conformance clauses, and the killbox atoms — which prevails when they conflict?",
 "why_it_matters": "The system publishes constitutional-adjacent text in at least six places: the hashed core block, the facet registry, participation.* laws, the owner-issued drop document, the C1-C34 conformance clauses, and the killbox ledger atoms (L1-L17). These already conflict in the open: the manifest, map, relay, and drop document disagreed on whether /api/invocations is public (FINDING 1); C30's evidence does not prove its clause text (obj-153); obj-52 contradicts C34. There is no supremacy rule stating which text yields — a partial hierarchy exists only implicitly (hash-pinned core first).",
 "candidate_answers": [
  {
   "answer": "Publish an explicit order of authorities",
   "type": "protocol_object",
   "rule": "Rank: (1) hashed core; (2) participation laws in the governance schema; (3) facet registry; (4) conformance clause texts; (5) conformance evidence; (6) prose documents; (7) killbox advocacy. Lower texts are interpretations, never sources, of higher law.",
   "candidate_protocol_object": "norms_registry object assigning each published text a rank and supersedes-links",
   "candidate_runtime": "any surface serving two conflicting texts emits a machine conflict notice naming the higher-rank text",
   "candidate_conformance": "clause: seeded conflict between a prose doc and a law must resolve to the law in machine output",
   "candidate_article": "/a/oip-norm-hierarchy"
  },
  {
   "answer": "Adopt the killbox masthead rule as general law: 'live server wins'",
   "type": "runtime",
   "rule": "Where sources conflict, the conflict is reported as a numbered finding and live machine behavior prevails over all documents.",
   "candidate_protocol_object": "conflict_finding records appended to the objection ledger",
   "candidate_runtime": "automated drift register (extends v1.2 Part F) publishing live-vs-document deltas",
   "candidate_conformance": "existing C29-style recomputation extended to docs",
   "candidate_article": "conflict-law article"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core, facets, participation)",
  "recon2/specification.md §1.9 FINDING 1, F6 (masthead conflict law: 'the conflict is reported as a numbered finding and the live server wins')",
  "obj-153 (C30 evidentiary weakness)",
  "obj-52 vs conformance C34",
  "recon/conformance.md (C1-C34)"
 ],
 "evidence": [
  "Conflicts are already ledgered as findings rather than resolved (FINDING 1 unresolved; obj-52 open against C34 PASS), proving the absence of a supremacy rule in practice",
  "The core hash is the only text with integrity pinning, making it the de facto apex — but only de facto"
 ],
 "counterevidence": [
  "The article-ledger constitution does impose claim/source rules on articles, and the append-protocol imposes CAS and ratify rules — partial hierarchies exist inside the content plane, suggesting the omission on the governance plane is scope, not oversight"
 ],
 "dependencies": [
  "govq-const-016",
  "govq-const-004"
 ],
 "falsifier": "A published norms-registry or a single conflict resolved by an explicit supremacy rule with a receipt.",
 "notes": "Foundation for the precedent & interpretation layer.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:834a10c2e7191e03b8a86a044a4d0503d0eb977dd0a51f6434d4fd6bceb37e14"
}

{
 "question_id": "govq-const-006",
 "layer": "constitutional amendment",
 "status": "MISSING",
 "question": "By what mechanism can the kernel itself be amended — who may propose, who ratifies, and how is the new core hash legitimated?",
 "why_it_matters": "This is the owner brief's first-ranked gap and the layer every other amendment question inherits. The live schema contains no vote, ratify, amend, or ballot record kind; participation.kinds is closed at subscribe/inquire/propose/feature/conformance/anchor/appeal/ruling, and ruling is owner-only. The closest deployed machinery is the article plane's propose→ratify→memorialize lane (voxel-vote / voxel-ratify under separated keys) — which governs content, not the constitution. As it stands, the only amendment mechanism is unilateral root republication of a new core hash, i.e. constitutional change by sovereign decree with an append-only trail.",
 "candidate_answers": [
  {
   "answer": "Multi-party ratified amendment lane",
   "type": "protocol_object",
   "rule": "Any subscribed node may propose a kernel amendment; ratification requires a published quorum of attested nodes weighted by a published rule, plus owner non-opposition or an override class; every step is an append-only receipt.",
   "candidate_protocol_object": "kernel_amendment_record (kind=amend: proposal, text delta, new core hash, quorum rule, window) + ratification_record (kind=ratify) + enactment_record linking old and new core hashes",
   "candidate_runtime": "intake accepts amend/ratify kinds; core endpoint serves the enacted version with supersession links; subscription records name the ratifying enactment",
   "candidate_conformance": "clause: an amendment lacking quorum receipts must not change the served core; a quorum-complete amendment must",
   "candidate_article": "/a/oip-kernel-amendment"
  },
  {
   "answer": "Owner-proposes, nodes-ratify-with-feet (exit as the only vote)",
   "type": "normative",
   "rule": "The root publishes new core versions unilaterally; legitimacy comes from nodes re-accepting or forking, recorded as subscription/anchor records against the new hash.",
   "candidate_protocol_object": "core_version_registry + re-acceptance subscriptions; no new kinds",
   "candidate_runtime": "core endpoint lists historical versions and per-version acceptance counts",
   "candidate_conformance": "none beyond counting",
   "candidate_article": "amendment-by-exit explainer"
  },
  {
   "answer": "Kernel is immutable by design; only facets evolve",
   "type": "normative",
   "rule": "oip-governance/1 is final. Constitutional change happens above the kernel (facets, profiles) or below it (implementations); the kernel's legitimacy claim IS its fixity.",
   "candidate_protocol_object": "immutability declaration inside core block",
   "candidate_runtime": "core endpoint serves 410-style 'final' metadata for v1",
   "candidate_conformance": "clause: any attempt to serve a modified five-axiom set under the same version string fails hash comparison",
   "candidate_article": "/a/oip-kernel-immutability"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (participation.kinds, submit.body)",
  "recon2/specification.md AC4 ('no threshold, timetable, or ratification record kind exists in the schema'), §3.4.9 power 7 ('no version-ratification process external to the author is visible in the public schema')",
  "corpus/bundle_append-protocol.md (voxel-vote/voxel-ratify propose-ratify-memorialize lane)",
  "upload master brief lines 33-37 ('The biggest one is constitutional change')"
 ],
 "evidence": [
  "participation.kinds enumerated live contains no amend/vote/ratify kind",
  "submit.body decision field reserves rulings to the owner ('owner-only kind=ruling')",
  "AC4: nothing deployed enforces a transition to multi-party rulings",
  "The article plane demonstrates the owner can build separated propose/ratify machinery (voxel-vote needs a different key to ratify) — the pattern exists but is not wired to the constitution"
 ],
 "counterevidence": [
  "The registry's own design ('fractional standard', facets selectable, fork anchors honored) suggests deliberate refusal to bind the root — i.e. the absence may be a choice (candidate answer 3), and v1.1 §3.4.9 already frames unilateral control as a named, receipted risk rather than an oversight"
 ],
 "dependencies": [
  "govq-const-007",
  "govq-const-023"
 ],
 "falsifier": "Any publication of an amendment procedure, quorum rule, or immutability declaration resolves the status; the question stays MISSING while no such text exists.",
 "notes": "The single highest-priority object on this plane per the owner brief.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:5a4b4002811941f6e51aac3663cf2e073e002e5a4b97483a3885a0a079093cb5"
}

{
 "question_id": "govq-const-007",
 "layer": "constitutional amendment",
 "status": "MISSING",
 "question": "If an amendment lane exists, what quorum and threshold classes apply — and is unanimity (or a node's individual consent) required before an axiom change binds that node?",
 "why_it_matters": "Constitutional systems distinguish ordinary amendment from entrenchment. The registry gives no quorum vocabulary at all: no node-weighting rule, no voting window, no classes of change. The design inputs that exist pull in opposite directions: the participation schema counts labels (Sybil-able), gov_7d480271 proposes pledge-weighted influence (plutocracy vector), and the boundary clause promises nodes they keep infrastructure and identity (a consent-based reading). Arrow's theorem (nogo-n02) is hosted on the site itself: any aggregation rule chosen will sacrifice one fairness criterion — the choice must be made knowingly, not by default.",
 "candidate_answers": [
  {
   "answer": "Supermajority classes with per-node opt-out",
   "type": "normative",
   "rule": "Facet changes: simple majority of attested nodes. Axiom additions: 2/3. Axiom modifications or boundary changes: 3/4 plus a per-node grandfathering right to exit with state portability.",
   "candidate_protocol_object": "quorum_rule object inside kernel_amendment_record; exit_record preserving receipts/anchors",
   "candidate_runtime": "tally computed over attested_node_count only (Part E Finding 1 field); exit records portable",
   "candidate_conformance": "clause: amendment tallies must ignore self-asserted labels",
   "candidate_article": "quorum schedule"
  },
  {
   "answer": "Unanimity of affected nodes for axiom change",
   "type": "normative",
   "rule": "No node is bound by an axiom change it did not accept; non-consenting nodes remain on the prior core as a recognized legacy constitution.",
   "candidate_protocol_object": "per-node accept_core_version registry",
   "candidate_runtime": "core endpoint serves all live versions; conformance profiles name their core version",
   "candidate_conformance": "version negotiation test between two nodes on different cores",
   "candidate_article": "consent-based amendment"
  },
  {
   "answer": "Pledge-weighted ratification (as filed in gov_7d480271)",
   "type": "protocol_object",
   "rule": "Ratification weight proportional to receipted, metered governance-energy pledges.",
   "candidate_protocol_object": "pledge ledger + weight function + ratification_record carrying pledge receipts",
   "candidate_runtime": "metered pledge contract (currently 'a proposed next contract, not a live payment or resource market')",
   "candidate_conformance": "audit of weight computation over pledge receipts",
   "candidate_article": "energy-weighted governance"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (participation, census_law)",
  "https://miscsubjects.com/api/governance/record/gov_7d480271c03a45138e62 (pledge-weight proposal)",
  "killbox-specification-v1.2.md Part E Findings 1-2 (Sybil-able census)",
  "/a/nogo-n02 (Arrow impossibility)",
  "/a/ostrom-1990 (collective-choice arrangements: users participate in rule-making)"
 ],
 "evidence": [
  "No quorum vocabulary anywhere in the live schema",
  "The only influence-weighting proposal on file (gov_7d480271) is pledge-weighted and status:open with zero rulings issued on it",
  "Ostrom principle 3 (collective-choice arrangements) is hosted as canonical systems literature — the corpus already contains the participation-in-rule-making norm the registry lacks"
 ],
 "counterevidence": [
  "Arrow's theorem (hosted as nogo-n02) proves no aggregation rule satisfies all fairness criteria — so any chosen quorum is a political trade-off, which is an argument for representing alternatives rather than hard-coding one",
  "At census≈0-4 self-asserted labels, every quorum rule is currently vacuous; v1.1 AC4 calls the capture risk 'legally null at current scale'"
 ],
 "dependencies": [
  "govq-const-006",
  "govq-const-011",
  "govq-const-008"
 ],
 "falsifier": "Any enacted quorum schedule with at least one ratified amendment; until then the question stands.",
 "notes": "Threshold design must cite nogo-n02 as a known impossibility boundary, not pretend completeness.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:aa87e05588a47e139e14bbb745df0755e8b2a1f935e73d898367e6e2cd223bb1"
}

{
 "question_id": "govq-const-008",
 "layer": "constitutional amendment",
 "status": "OPEN",
 "question": "May a subscriber legitimately remain on a superseded core version, and may the root declare an old version obsolete — what are the obsolescence and legacy-rights rules?",
 "why_it_matters": "Version pinning already exists at the record level (every filing quotes version+hash), but there is no lifecycle law: no deprecation record kind, no sunset window, no statement whether an old core remains a valid constitution after a new one ships. This decides whether amendment is gentle (nodes migrate at will) or coercive (the root can strand holdouts). The runtime already shows version drift handled silently: the stored conformance run evidences dispatch 1.0.0 while the live manifest serves 1.2.0 (v1.2 Part F, D5) — version transitions are happening today with no constitutional rule governing them.",
 "candidate_answers": [
  {
   "answer": "Legacy cores remain valid; obsolescence is advisory",
   "type": "normative",
   "rule": "A core version, once published, is a valid constitution forever; the root may mark it deprecated but must keep serving its text and accepting records against it.",
   "candidate_protocol_object": "core_version_registry with status field (current|deprecated); no forced migration",
   "candidate_runtime": "core endpoint serves all versions permanently; deprecation is metadata",
   "candidate_conformance": "clause: records naming a deprecated hash must still validate",
   "candidate_article": "legacy-rights article"
  },
  {
   "answer": "Root may sunset cores with notice",
   "type": "normative",
   "rule": "The root may declare end-of-life for a core version with a published notice window (e.g. 180 days) and a migration path; after sunset the version is archival.",
   "candidate_protocol_object": "deprecation_record (kind=deprecate) with window and successor link",
   "candidate_runtime": "sunset scheduler; after deadline, old-hash subscriptions return 410 with education",
   "candidate_conformance": "clause: no sunset without a deprecation record older than the window",
   "candidate_article": "version-lifecycle law"
  },
  {
   "answer": "No versioning of the kernel at all (immutability reading)",
   "type": "normative",
   "rule": "There is exactly one kernel; 'versions' are new constitutions. Obsolescence is incoherent — the old text is never obsolete, only abandoned.",
   "candidate_protocol_object": "immutability declaration",
   "candidate_runtime": "n/a",
   "candidate_conformance": "hash comparison",
   "candidate_article": "/a/oip-kernel-immutability"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core.version)",
  "killbox-specification-v1.2.md Part F D5 (runtime version drift 1.0.0 -> 1.2.0 with stored conformance run stale)",
  "recon2/specification.md §2.2 (versioned subscription records)"
 ],
 "evidence": [
  "Records pin version+hash, so the schema anticipates multiple cores, but no lifecycle law exists",
  "v1.2 Part F shows the runtime advancing (dispatch 1.0.0 -> 1.2.0) while the served conformance verdict was cached against the older build — evidence that version transitions occur without constitutional handling"
 ],
 "counterevidence": [
  "An append-only ledger with permanent records arguably commits the root to keeping old cores readable; the fork_anchor_law already honors external histories without endorsing them, supplying a partial model for legacy tolerance"
 ],
 "dependencies": [
  "govq-const-006",
  "govq-const-007"
 ],
 "falsifier": "Publication of a version-lifecycle law or a first deprecation event handled by rule rather than silence.",
 "notes": "Directly extends the brief's example; drift evidence makes it non-hypothetical.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:6369c397b8d4a953e2f2e7d0900eb2d8f536d1aebb5033712ea26b4d63933f98"
}

{
 "question_id": "govq-const-009",
 "layer": "constitutional amendment",
 "status": "MISSING",
 "question": "Are protections that existed when a node joined entrenched against later majorities — is there a non-retroactivity / vested-rights clause, and can the kernel acquire one by amendment?",
 "why_it_matters": "The master brief asks: 'Can a majority remove protections that existed when a node joined?' The registry publishes no non-derogable rights, no retroactivity rule, and no vesting clause. A node's reliance interests are real: it subscribed against a specific hash, built against facet obligations, and anchored fork heads under specific laws. If a later amendment can strip those protections, subscription is precarious; if they are entrenched, the constitution needs a supremacy-and-entrenchment architecture (govq-const-002, -005) that does not yet exist.",
 "candidate_answers": [
  {
   "answer": "Entrench reliance protections at join time",
   "type": "normative",
   "rule": "Each subscription record snapshots the protections in force (boundary clause, no-authority-grant, fork tolerance); later amendments cannot retrospectively alter that node's snapshot; new terms bind only on re-acceptance.",
   "candidate_protocol_object": "protection_snapshot embedded in subscription records + non_retroactivity_clause in core",
   "candidate_runtime": "intake stores the snapshot; conformance checks disputes against the snapshot's version",
   "candidate_conformance": "clause: a ruling applying post-join amendments to a pre-amendment act must be appealable with automatic stay",
   "candidate_article": "vested-rights article"
  },
  {
   "answer": "No vested rights; the constitution is always the current hash",
   "type": "normative",
   "rule": "Membership is at-will under the current text; the only protection is exit-with-records.",
   "candidate_protocol_object": "exit_record guaranteeing receipt/anchor portability",
   "candidate_runtime": "export endpoint for a node's full record set",
   "candidate_conformance": "portability test",
   "candidate_article": "at-will membership explainer"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core, participation laws)",
  "upload master brief line 35 ('Can a majority remove protections that existed when a node joined?')",
  "/a/thinker-robert-nozick (side constraints; rectification principle)"
 ],
 "evidence": [
  "No retroactivity, vesting, or entrenchment text anywhere in the live schema or the killbox corpus",
  "Subscription records already snapshot version+hash per filing — the technical seed of a reliance-protection mechanism exists"
 ],
 "counterevidence": [
  "The append-only discipline means historical protections are never erased from the record, which is evidence of reliance interests even without a vesting rule",
  "Nozick's hosted page frames rights as side constraints on collective action — the corpus contains the philosophical template for entrenchment"
 ],
 "dependencies": [
  "govq-const-002",
  "govq-const-007",
  "govq-const-006"
 ],
 "falsifier": "Any published retroactivity or entrenchment clause; or a live case where an amendment is applied retroactively without remedy — either observation settles the status.",
 "notes": "Links to rights layer (other role); this question is confined to the constitutional-amendment aspect.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:878f6518891d77a5712d475d8cd4483e86e02bccc1b5bec332f45466176b087c"
}

{
 "question_id": "govq-const-010",
 "layer": "constitutional amendment",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "Does the defensive-commons commitment constrain constitutional amendment — i.e., can the root amend the kernel into a privately enforceable posture, or is the open-kernel pledge constitutionally (and legally) locked?",
 "why_it_matters": "The defensive-commons facet obliges the owner to 'publish the operative pledge or license when it exists'; the killbox v1.2 counsel audit found no executed covenant or license anywhere public (Vector 8), so 'patent-then-pledge' is currently a posture, not a lock. If the pledge never executes, nothing stops a future amendment from enclosing the kernel the disclosure invited the public to build on — a credibility and possibly equitable-estoppel exposure. This is simultaneously a constitutional-design question (should the open-kernel commitment be entrenched?) and a legal-instrument question (does any executed document bind the owner?), and the second half is reserved to counsel.",
 "candidate_answers": [
  {
   "answer": "Execute the defensive covenant as a governance record and entrench it",
   "type": "protocol_object",
   "rule": "The royalty-free/defensive covenant is recorded as a binding governance instrument linked to the defensive-commons facet and listed as non-derogable by amendment.",
   "candidate_protocol_object": "covenant_record (kind=ruling or new kind=covenant) binding the pledge text to the facet lineage + entrenchment flag",
   "candidate_runtime": "facet registry links the executed instrument; amendment intake refuses changes that conflict with it",
   "candidate_conformance": "clause: amendment conflicting with covenant text must fail validation",
   "candidate_article": "/a/oip-defensive-covenant"
  },
  {
   "answer": "Keep the pledge extra-constitutional (legal instrument only)",
   "type": "normative",
   "rule": "The openness guarantee lives in patent/contract law, not in the kernel; the constitution stays neutral and the covenant is a private commitment.",
   "candidate_protocol_object": "none at protocol level; executed license/covenant document (counsel-drafted)",
   "candidate_runtime": "publication of the instrument URL in the facet registry (obligation already exists)",
   "candidate_conformance": "none",
   "candidate_article": "legal-posture note"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance facets.defensive-commons obligations ('publish the operative pledge or license when it exists')",
  "killbox-specification-v1.2.md Part D Vector 8 (no license visible; 'posture, not lock'; Top-10 action 8: 'execute the defensive covenant')",
  "obj-149/obj-151 (owner strategy canon)",
  "recon2/territory-ledger.json (14 atoms)"
 ],
 "evidence": [
  "The facet obligation to publish the pledge 'when it exists' implies none exists; v1.2 Part D 8.1 confirms no executed covenant or license linked anywhere public and scores the exposure HIGH",
  "obj-149 memorializes the owner strategy (publish as prior art, preserve 102(b)(1), then pledge) — strategy is recorded, execution is not"
 ],
 "counterevidence": [
  "The disclosure's candor (the strategy is published, the gap is published) reduces concealment narratives; defensive publication itself survives regardless of the pledge's execution"
 ],
 "dependencies": [
  "govq-const-006",
  "govq-const-002"
 ],
 "falsifier": "Appearance of an executed covenant instrument on the public record (resolves toward answered) or counsel's written determination that no such instrument binds (resolves toward LEGAL_REVIEW_REQUIRED permanently).",
 "notes": "Marked LEGAL_REVIEW_REQUIRED: the operative content is a legal characterization the mission law reserves for counsel.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:8c651be60db80db06ef9256c351936c803682e87050d4faaa7bd1a0dff9b8028"
}

{
 "question_id": "govq-const-011",
 "layer": "constitutional amendment",
 "status": "OPEN",
 "question": "Should amendment influence be weighted by receipted governance-energy pledges — and does pledge-weighting recreate plutocracy with better receipts?",
 "why_it_matters": "gov_7d480271 (status: open) proposes that 'pledge weight must govern influence, so governance power is auditable not reputational.' The master brief names the risk: 'can governance-energy pledges become plutocracy.' The registry's own invariant says more subscribers never automatically increase any person's unilateral authority, but says nothing about weighted authority. The shared-governance-energy facet is specified but unmetered ('a proposed next contract'). The question is constitutional: whether the future amendment rule (govq-const-006/-007) weighs attestations, energy, or one-actor-one-voice — and the corpus already contains both the proposal and its plutocracy objection.",
 "candidate_answers": [
  {
   "answer": "Pledge-weighted with caps and decay",
   "type": "normative",
   "rule": "Influence weight = receipted pledged energy, capped per attested entity and decaying over time so weight must be continually re-earned; caps enforced by the same census-attestation machinery as the census fix.",
   "candidate_protocol_object": "pledge_ledger + weight_function object (cap, decay, entity resolution) + ratification_record carrying weight proof",
   "candidate_runtime": "metered pledge contract plus attested-entity resolution (Part E Finding 1-2 fields)",
   "candidate_conformance": "clause: weight computation must be reproducible from public pledge receipts; over-cap weight must not count",
   "candidate_article": "energy-weighted governance with anti-plutocracy caps"
  },
  {
   "answer": "One-attested-entity-one-voice for constitutional questions",
   "type": "normative",
   "rule": "Energy buys work priority, never constitutional voice; amendment ratification is per-entity.",
   "candidate_protocol_object": "attested entity registry feeding ratification_record",
   "candidate_runtime": "entity resolution as in census fix",
   "candidate_conformance": "Sybil test: N labels one proof count once",
   "candidate_article": "constitutional suffrage"
  },
  {
   "answer": "Reject formal weighting; influence stays informal and auditable",
   "type": "normative",
   "rule": "No weighting rule is enacted; the ledger's transparency is considered sufficient discipline on influence.",
   "candidate_protocol_object": "none",
   "candidate_runtime": "existing receipt publicity",
   "candidate_conformance": "none",
   "candidate_article": "influence-transparency note"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance/record/gov_7d480271c03a45138e62",
  "https://miscsubjects.com/api/governance (post_subscription.invariant, implementation_status; shared-governance-energy facet)",
  "killbox-specification-v1.2.md Part E Finding 7 (incentive asymmetry)",
  "recon2/specification.md line 1474 ('Obligation 4 (non-capture) ... stands open')",
  "upload master brief line 154"
 ],
 "evidence": [
  "The pledge-weight proposal is filed and open with zero rulings",
  "implementation_status: metered pledge allocation is 'a proposed next contract, not a live payment or resource market'",
  "v1.1 spec: 'nothing in Section 4 shows pledge-weighted governance does not recreate plutocracy with receipts'"
 ],
 "counterevidence": [
  "The facet's not_required list excludes 'unbounded compute or governance authority', and the invariant caps unilateral authority growth — textual evidence the designers intended weight-with-limits, though no limit mechanism exists"
 ],
 "dependencies": [
  "govq-const-007",
  "govq-const-006"
 ],
 "falsifier": "Enactment of any weighting rule, or a live amendment attempt where pledge receipts are treated as votes.",
 "notes": "Keep the filed proposal and the plutocracy objection linked as unresolved alternatives per mission law.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:cb2d9b1cf9526647ffba94dc9cab9eb1c7e45e6b902bcaff1942ecc8b2f3d092"
}

{
 "question_id": "govq-const-012",
 "layer": "semantic versioning of constitutional terms",
 "status": "MISSING",
 "question": "Which words in the kernel are constitutional terms of art ('bounded', 'explicit authority', 'widen', 'canonical bytes', 'distinguishable', 'auditable', 'appended', 'rewritten'), and where is their version-controlled definition registry?",
 "why_it_matters": "The five axioms are one sentence each; their operative force rides on undefined terms. The system already knows how to version meaning elsewhere — claim-DIVs carry semantic_hash and version_hash, contracts carry fingerprints pinned at invocation (C23), and the P7 chain carries a compatibility law forbidding recipe recomputation across schema versions. None of this machinery reaches the constitution: the kernel's terms have no definitions ledger, no per-term version, no change control. obj-148 is the live demonstration of the cost — a validator rejects for a missing 'attribution_header' it nowhere defines ('a validator that cannot state its own rule is a wall, not a gate').",
 "candidate_answers": [
  {
   "answer": "Versioned constitutional definitions ledger",
   "type": "protocol_object",
   "rule": "Every operative term in the kernel and facet registry gets a definition record with version, semantic hash, change history, and supersession links; core versions cite definition-set versions.",
   "candidate_protocol_object": "constitutional_definitions_ledger (term, definition, version, semantic_hash, supersedes, effective_core_version)",
   "candidate_runtime": "definition endpoint serves per-term history; subscription and conformance records name the definitions version they interpret under",
   "candidate_conformance": "clause: a conformance run must name its definitions version; two runs under different versions are not comparable",
   "candidate_article": "/a/oip-constitutional-terms"
  },
  {
   "answer": "Terms of art defined by the conformance suite (behavioral definitions)",
   "type": "conformance",
   "rule": "A term means what the deployed clauses test; definitions are executable worked examples rather than prose.",
   "candidate_protocol_object": "worked_example registry per term (input, expected verdict class)",
   "candidate_runtime": "conformance suite extended with per-term example packs",
   "candidate_conformance": "C23-style pinning applied to definitional examples",
   "candidate_article": "behavioral-semantics note"
  },
  {
   "answer": "Deliberate ambiguity: terms remain common-language",
   "type": "normative",
   "rule": "The kernel is a minimal interoperability floor; over-definition is rejected as premature closure. Disputes resolve by fork/exit.",
   "candidate_protocol_object": "none",
   "candidate_runtime": "none",
   "candidate_conformance": "none",
   "candidate_article": "ambiguity-as-design boundary article"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core.axioms verbatim)",
  "recon2/specification.md §1.6 (semantic_hash/version_hash on claim-DIVs)",
  "killbox-specification-v1.2.md P7 atom ('compatibility law forbidding recipe recomputation across schema versions')",
  "obj-148 (undefined attribution_header)",
  "conformance C23 (contract fingerprint pinning)"
 ],
 "evidence": [
  "No definitions registry exists for kernel terms; participation laws, facet obligations, and axiom texts use undefined operative words throughout",
  "obj-148 (open) shows an enforcement gate keyed to an undefined term, with nine public rejection receipts",
  "The system versions semantics for content (semantic_hash/version_hash) and contracts (fingerprints) but not for constitutional text"
 ],
 "counterevidence": [
  "The boundary clause narrows what the kernel governs, arguably reducing the blast radius of ambiguity; candidate answer 3 (deliberate ambiguity) is consistent with the fractional-standard design"
 ],
 "dependencies": [
  "govq-const-001",
  "govq-const-013",
  "govq-const-014"
 ],
 "falsifier": "Publication of any term registry or an explicit refusal article; or a live dispute where two parties cite the same axiom text with incompatible meanings and no definitional recourse exists.",
 "notes": "Root object for the whole layer; the MD rationale nominates the initial term list.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:2ad6381a66d2381cb31fe50e8b6799890c2c9b90d19084f2a8cbc82f11c0e086"
}

{
 "question_id": "govq-const-013",
 "layer": "semantic versioning of constitutional terms",
 "status": "IMPLIED",
 "question": "Does the kernel's version string ('oip-governance/1') carry semver semantics — what changes are major/minor/patch, and is there a compatibility law between core versions?",
 "why_it_matters": "The core block carries version 'oip-governance/1' and a hash; the relay/social chain already enforces a compatibility law forbidding recipe recomputation across schema versions (P7); and conformance runs are version-stamped (the stored run evidences dispatch 1.0.0 against a live 1.2.0 manifest — v1.2 Part F D5). A versioning discipline therefore exists in adjacent planes, and its constitutional analogue is implied but nowhere stated: nobody knows whether rewording an axiom without changing meaning is a patch, whether adding a facet is minor, whether touching an axiom is major — or what compatibility two core versions owe each other's records.",
 "candidate_answers": [
  {
   "answer": "Adopt explicit constitutional semver with a compatibility law",
   "type": "protocol_object",
   "rule": "Patch = non-normative text (typos, clarifications, hash changes with signed semantic-equivalence attestation). Minor = facet additions or obligation loosening. Major = axiom change, boundary change, or obligation tightening. Major versions must publish a compatibility matrix stating which prior-version records remain valid.",
   "candidate_protocol_object": "core_version_registry with semver class + compatibility_law record per version pair + semantic_equivalence attestation for patches",
   "candidate_runtime": "version negotiation on read; cross-version record validity computed from the compatibility matrix",
   "candidate_conformance": "clause: a patch-level diff must carry the equivalence attestation; absence fails validation",
   "candidate_article": "/a/oip-core-semver"
  },
  {
   "answer": "Hash-is-version: no semver, every byte change is a new constitution",
   "type": "normative",
   "rule": "The only meaningful version is the hash; semantic equivalence claims are untrustworthy and refused.",
   "candidate_protocol_object": "none beyond hash pinning",
   "candidate_runtime": "core endpoint serves hash-addressed texts",
   "candidate_conformance": "hash comparison only",
   "candidate_article": "hash-purism note"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core.version)",
  "killbox-specification-v1.2.md P7 compatibility law; Part F D5 (dispatch 1.0.0 vs 1.2.0 drift)",
  "recon2/specification.md §2.2"
 ],
 "evidence": [
  "A version string exists with no published semantics for it",
  "Adjacent planes (relay chain, conformance) already operate version-pinning and compatibility laws — the pattern to copy exists in-corpus"
 ],
 "counterevidence": [
  "Treating every byte change as major (hash-purism) is simple and matches append-only instincts; the cost (every clarification refounds the constitution) may be intended"
 ],
 "dependencies": [
  "govq-const-012",
  "govq-const-008",
  "govq-const-004"
 ],
 "falsifier": "Publication of a semver/compatibility rule for core versions, or the first actual core change handled by rule rather than by silence.",
 "notes": "Implied by the existence of a version field plus adjacent-plane compatibility laws; nothing enacted.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:81e237249304d8d59cf83069ef04c35dcf12cd1f84ad0543244efb5d88e4c4a4"
}

{
 "question_id": "govq-const-014",
 "layer": "semantic versioning of constitutional terms",
 "status": "OPEN",
 "question": "How does the constitution defend itself against semantic capture — a future operator keeping the names ('conformance', 'independent', 'repair', 'material success') while changing their meaning?",
 "why_it_matters": "The master brief's fourteenth gap: 'A future operator may keep the same names while changing their meaning... governance needs semantic-version rules and protected definitions.' Drift precursors are already observable: counts.by_facet tallies facet mentions as if they were adoption (Part E Finding 4); the census integer does evidentiary work as 'nodes' while counting labels (Part E Findings 1-2); the served conformance verdict ran against an older build while the manifest advanced (Part F D5). Each is a small semantic slide from 'as-filed' to 'as-is' — exactly the capture vector, currently defended only by prose disclaimers.",
 "candidate_answers": [
  {
   "answer": "Protected definitions with drift alarms",
   "type": "runtime",
   "rule": "Definitions in the constitutional_definitions_ledger are protected: changing one requires the amendment lane; every public surface that uses a protected term must serve it with its definitions-version tag; automated drift registers compare served usage against the tagged definition.",
   "candidate_protocol_object": "protected_term flags + drift_register (extends v1.2 Part F) + usage tags on API fields",
   "candidate_runtime": "scheduled recompute comparing metric names, tally semantics, and disclaimers against the definitions ledger; mismatches file conflict findings",
   "candidate_conformance": "clause: a public integer labeled 'node_count' must compute the definition-tagged predicate or fail",
   "candidate_article": "semantic-capture defense"
  },
  {
   "answer": "Machine fields carry their semantics (self-describing metrics)",
   "type": "protocol_object",
   "rule": "Every aggregate the registry serves embeds its own computation law as a machine field (as census_law already does in prose), so meaning travels with the number.",
   "candidate_protocol_object": "per-metric law blocks (extends census_law pattern to all counts)",
   "candidate_runtime": "counts object serves per-key law + predicate reference",
   "candidate_conformance": "clause: no bare integers in registry responses",
   "candidate_article": "self-describing metrics"
  }
 ],
 "existing_artifacts": [
  "killbox-specification-v1.2.md Part E Findings 1, 2, 4; Part F D5",
  "upload master brief lines 193-195 (semantic capture)",
  "obj-148 (undefined term gate)",
  "https://miscsubjects.com/api/governance (census_law as per-metric prose law)"
 ],
 "evidence": [
  "The one existing defense is per-metric prose law (census_law); it is not machine-bound to the computation and covers only census",
  "Part E Finding 4: by_facet tallies already 'behave as adoption metrics while measuring only free-form facet naming'"
 ],
 "counterevidence": [
  "The registry does publish honest disclaimers (census_law's 'auditable records, not a claim of unique legal persons'), showing semantic hygiene is valued in prose; the gap is enforcement, not intent"
 ],
 "dependencies": [
  "govq-const-012",
  "govq-const-013"
 ],
 "falsifier": "A detected case where a public metric's served meaning diverges from its ledgered definition with no conflict finding filed — the defense's absence made observable.",
 "notes": "Feeds the drift-register candidate runtime shared with govq-const-016.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:f371b6542d675c7cca11c48d2877021bf3f2b69644344279bbabb7a2e55fe2e1"
}

{
 "question_id": "govq-const-015",
 "layer": "precedent & interpretation",
 "status": "MISSING",
 "question": "Do rulings have precedential force — must later rulings distinguish or overturn earlier ones, and who may do so?",
 "why_it_matters": "registry_law makes delist/reinstate rulings append-only, which preserves prior decisions as history but says nothing about their normative force. Zero rulings exist on the live registry (open_requests=5, no kind=ruling record has ever been issued), so the question is unpressed today and structural tomorrow: the first enterprise appealing a delisting will ask whether the owner-judge is bound by his own prior rulings. Append-only lineage without a precedent doctrine is a memory without a law; with one, it is the seed of case law. The corpus hosts both templates: Dworkin's integrity (coherence with past decisions) and the fork-exit model (no binding precedent, only persuasive history).",
 "candidate_answers": [
  {
   "answer": "Rulings bind as distinguishable precedent",
   "type": "normative",
   "rule": "A ruling must cite and distinguish every prior ruling on substantially similar facts, or expressly overturn it with reasons; overturning is itself a ruling, subject to appeal.",
   "candidate_protocol_object": "precedent_link fields on ruling records (distinguishes[], overturns[], relies_on[]) + reasoned_ruling requirement",
   "candidate_runtime": "intake refuses kind=ruling lacking precedent analysis when similar rulings exist; similarity gate reuses dedup machinery",
   "candidate_conformance": "clause: a ruling silent on cited similar precedent is invalid",
   "candidate_article": "precedent doctrine"
  },
  {
   "answer": "No binding precedent; published reasoned rulings as persuasive history",
   "type": "normative",
   "rule": "Each ruling is sovereign but must state reasons; consistency is enforced socially through the public append-only record.",
   "candidate_protocol_object": "reasoned_ruling text requirement only",
   "candidate_runtime": "ruling records carry rationale field",
   "candidate_conformance": "none",
   "candidate_article": "persuasive-history note"
  },
  {
   "answer": "Precedent via conformance codification",
   "type": "conformance",
   "rule": "Rulings do not bind; instead, any ruling that should generalize must be codified as a conformance clause or facet-obligation amendment through the amendment lane.",
   "candidate_protocol_object": "ruling->clause promotion path (links to L17 law-to-clause closure rule)",
   "candidate_runtime": "L17-style closure: every ruling either dies one-off or spawns a clause proposal",
   "candidate_conformance": "L17 closure test",
   "candidate_article": "codification doctrine"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (registry_law; counts: zero rulings)",
  "recon2/specification.md §3.4.9 powers 2-3 (owner-only ruling kind; appeals decided by same owner)",
  "killbox-specification-v1.2.md Part G L17 (law-to-clause closure rule, PROPOSED)",
  "/a/thinker-ronald-dworkin (value coherence; interpretive integrity — page covers Religion Without God, not Law's Empire; interpretivism link is the cartographer's inference)",
  "upload master brief lines 139-143"
 ],
 "evidence": [
  "participation.registry_law: 'Delisting and reinstatement are append-only owner rulings linked to the affected record' — lineage preserved, force unspecified",
  "Live registry has zero kind=ruling records (all five records open)",
  "L17 exists as a PROPOSED atom for turning law into clauses, i.e. a codification path exists in draft"
 ],
 "counterevidence": [
  "Append-only storage plus public reasoning could be argued to create de facto precedent without doctrine; and at zero rulings the absence is unpriced"
 ],
 "dependencies": [
  "govq-const-016",
  "govq-const-017"
 ],
 "falsifier": "The first ruling that cites, distinguishes, or ignores prior rulings will force the doctrine; a published precedent clause resolves it earlier.",
 "notes": "Dworkin citation limited to what the hosted page supports; integrity-theory link flagged as inference.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:7df7afd61b4be012ed060af79e24a0b2be7e7d3ef4ae9e8b4308abd52aaaad3c"
}

{
 "question_id": "govq-const-016",
 "layer": "precedent & interpretation",
 "status": "PARTIAL",
 "question": "What is the canonical interpretive method when the machine schema, the prose documents, the conformance evidence, and the killbox claims disagree — is there a lawful rule of recognition for 'what the constitution says'?",
 "why_it_matters": "Conflicts are already routine and handled by convention, not law: the killbox masthead law says 'where sources conflict the conflict is reported as a numbered finding and the live server wins' (v1.1 F6), yet FINDING 1 (invocations route advertised public vs 404 live) stands unresolved, and obj-52 (MCP_ATTACH executes on non-owner key) contradicts C34's PASS in the same public corpus. A constitution needs its interpretive canon — which text is read first, whether behavior outranks prose, whether candor artifacts (sorry-counts, findings) are admissions. The pieces exist (conflict findings, drift registers, live-server-wins convention) but are assembled only inside swarm documents, not in the registry's own law.",
 "candidate_answers": [
  {
   "answer": "Enact the interpretive canon as registry law",
   "type": "protocol_object",
   "rule": "Order of interpretation: (1) live machine behavior; (2) hashed core and participation laws; (3) facet registry; (4) clause texts; (5) prose documents; (6) advocacy documents. Every detected conflict must land as a numbered finding on the objection ledger within a stated window.",
   "candidate_protocol_object": "interpretation_canon record in the governance schema + mandatory conflict_finding kind",
   "candidate_runtime": "drift register (v1.2 Part F pattern) promoted from audit-time to continuous; conflicts auto-filed",
   "candidate_conformance": "clause: seeded doc-vs-behavior conflict must produce a finding record",
   "candidate_article": "/a/oip-interpretation-canon"
  },
  {
   "answer": "Keep interpretation extra-legal: conflicts are findings forever",
   "type": "normative",
   "rule": "The system commits to publishing conflicts, never to resolving them by rank; readers apply their own canon. Honesty substitutes for hierarchy.",
   "candidate_protocol_object": "none beyond existing findings practice",
   "candidate_runtime": "existing findings/findings-as-findings",
   "candidate_conformance": "none",
   "candidate_article": "conflict-without-hierarchy boundary article"
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md §1.9 FINDING 1, F6 (masthead conflict law)",
  "obj-52 vs C34 contradiction",
  "killbox-specification-v1.2.md Part F (drift register methodology)",
  "recon2/specification.md line 77 ('the disagreement is published as a numbered FINDING ... never resolved by preference')"
 ],
 "evidence": [
  "A de facto canon (live-server-wins) is stated in swarm documents and applied in audits, but it is not registry law and is applied inconsistently (FINDING 1 unresolved)",
  "The corpus already practices conflict-preservation as method ('conflicts between sources are findings, never smoothed' is also this mission's law)"
 ],
 "counterevidence": [
  "The masthead canon is arguably binding only on killbox documents, not on the constitution; treating audit conventions as constitutional law is a category error the registry has never endorsed"
 ],
 "dependencies": [
  "govq-const-005",
  "govq-const-018",
  "govq-const-015"
 ],
 "falsifier": "Adoption of any canon text into the governance schema; or a live conflict resolved silently by rank, demonstrating the canon's absence by its exercise.",
 "notes": "Pairs with govq-const-005 (hierarchy of norms); that question is structural, this one is methodological.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:d8228df01374309f09fe2a91d0bbb3055b7132fb65ac3271ae70d22e7bb9bc53"
}

{
 "question_id": "govq-const-017",
 "layer": "precedent & interpretation",
 "status": "CONTRADICTED",
 "question": "What is the constitutional effect of malformed or defect-produced records — does obj-154's artifactual 'settled' status have force, and who may void a record the ledger cannot delete?",
 "why_it_matters": "obj-154 was pipe-mangled at ingress: stored truncated, actor flipped to 'anonymous', status flipped to 'settled', with answered_at identical to created_at — no adjudication occurred, yet the public record reads settled, and the v1.1 killbox quoted it as an admission before the audit caught the provenance defect. Append-only law forbids deletion; repair lanes were similarity-blocked (FINDING 9; bypass now deployed but un-rate-limited per Part E Finding 6). The constitution currently has no concept of a void or voidable record: status flips produced by parser defects carry the same surface authority as owner rulings. This is the interpretation layer's hard case: integrity of history vs. integrity of status.",
 "candidate_answers": [
  {
   "answer": "Voidability by annotation: status-correction records",
   "type": "protocol_object",
   "rule": "A record produced by a verified ingress/parse defect is not deleted but annotated void-for-defect by a linked correction record; consumers must read status through the correction chain (as count_law already requires for rulings).",
   "candidate_protocol_object": "status_correction_record (kind=ruling decision=void or new kind) + traversal law extension ('effective status requires traversing corrections')",
   "candidate_runtime": "registry serves effective_status computed over the correction chain; raw record untouched",
   "candidate_conformance": "clause: a voided-by-correction record must not render as settled in any public view",
   "candidate_article": "voidability doctrine"
  },
  {
   "answer": "Defect records stand as filed; meaning is the reader's problem",
   "type": "normative",
   "rule": "The ledger records events, not truths; a mangled record is a true record of a mangled event. No status correction exists — only appended context.",
   "candidate_protocol_object": "none beyond existing repair linkage",
   "candidate_runtime": "existing append lanes",
   "candidate_conformance": "none",
   "candidate_article": "event-not-status note"
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md §1.9 FINDING 9 and AC3 (obj-154 truncation, false settled, artifactual answer)",
  "killbox-specification-v1.2.md Part E Finding 6 (repair bypass un-rate-limited)",
  "obj-154 live record (actor 'anonymous', answered_at == created_at)",
  "https://miscsubjects.com/api/articles/oip/objections"
 ],
 "evidence": [
  "The live record publicly reads status 'settled' while the audit corpus proves no adjudication occurred — the contradiction is between stored status and provenance, both public",
  "count_law already establishes the traversal principle ('does not replace traversal of later append-only rulings') that a voidability rule would extend"
 ],
 "counterevidence": [
  "The killbox v1.1 itself quoted obj-154 as an admission before correction — evidence that even sophisticated readers are captured by raw status; also evidence the corpus self-corrects in public"
 ],
 "dependencies": [
  "govq-const-015",
  "govq-const-016"
 ],
 "falsifier": "Any correction record or traversal law that changes how obj-154's status is served; until then the contradiction between surface status and provenance stands.",
 "notes": "Kept CONTRADICTED, not resolved; the two public faces are both cited.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:1900fa6e512830dba8b9a324439cd2f92446c15ea6131823f5bc1d523120c093"
}

{
 "question_id": "govq-const-018",
 "layer": "precedent & interpretation",
 "status": "OPEN",
 "question": "When the conformance suite is the operative interpretation of the axioms, what disciplines the interpreter — given that the suite's author, runner, and grader are the same party the suite constrains?",
 "why_it_matters": "v1.1 §2.4 wires each axiom to executing clauses; AC4 names the structural problem: the author 'writes the clauses, writes the tests, runs them against his own deployment, and publishes the verdict.' obj-153 shows the failure mode live: C30 passes with evidence (a dead-token 401) that does not test its clause text (CAS hash gating) — an interpretive slide that would pass a lesser reader. The served verdict is also cached while the runtime advanced (Part F D5). Constitutional interpretation by self-graded test is currently the only interpretation there is; independent verifier receipts are specified as a concept ('Self-attested evidence remains self-attested until an independent verifier receipt is linked') but no verifier population exists.",
 "candidate_answers": [
  {
   "answer": "Independent verifier population with counter-suites",
   "type": "conformance",
   "rule": "Interpretation of kernel terms requires at least one non-owner suite re-running clauses and publishing counter-verdicts; conformance records must carry evidence_class and verifier_receipt (Part E Finding 4 fields).",
   "candidate_protocol_object": "verifier_registry + counter_suite runs + conformance record fields scope/evidence_class/verifier_receipt",
   "candidate_runtime": "conformance endpoint serves owner and non-owner runs side by side with drift deltas",
   "candidate_conformance": "meta-clause: a clause whose evidence does not test its text (obj-153 pattern) fails on independent re-run",
   "candidate_article": "independent-interpretation doctrine"
  },
  {
   "answer": "Clause authorship separated from clause execution",
   "type": "runtime",
   "rule": "Clauses are proposed publicly and ratified through the amendment lane before execution; the operator runs but does not author constitutional tests.",
   "candidate_protocol_object": "clause_proposal -> ratification_record -> clause registry (L17 closure machinery)",
   "candidate_runtime": "L17 enacted",
   "candidate_conformance": "L17 closure test",
   "candidate_article": "separated-interpretation note"
  },
  {
   "answer": "Self-grading disclosed as a known, bounded evil",
   "type": "normative",
   "rule": "Self-grading is accepted for the bootstrap era and labeled as such on every verdict; the constitution commits to a verifier transition trigger (e.g. first external subscriber) but not a mechanism.",
   "candidate_protocol_object": "self_graded flag on verdicts (extends existing masthead disclosure)",
   "candidate_runtime": "verdicts carry the flag",
   "candidate_conformance": "none",
   "candidate_article": "bootstrap-disclosure note"
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md §2.4, AC4 power 4",
  "obj-153 (C30 evidentiary weakness)",
  "killbox-specification-v1.2.md Part F D5 (cached verdict vs runtime drift)",
  "conformance_law ('self-attested until an independent verifier receipt is linked')",
  "Part E Finding 4 (evidence_class fields proposal)"
 ],
 "evidence": [
  "C30's PASS evidence is a 401 from an invalid token — it proves the auth gate, not the CAS check the clause texts describes; filed as obj-153, open",
  "The only conformance verdict on the site is the operator's own suite, 'self-graded ... served cached' (killbox masthead)"
 ],
 "counterevidence": [
  "The masthead already discloses self-grading publicly, and obj-153 was found and filed by the system's own audit loop — partial self-correction exists; conformance_law states the independence principle even without a verifier population"
 ],
 "dependencies": [
  "govq-const-016",
  "govq-const-001",
  "govq-const-015"
 ],
 "falsifier": "First non-owner conformance run published with at least one verdict differing from the owner suite — the interpretive monopoly broken or confirmed.",
 "notes": "Constitutional-court question in test-suite form; deliberately framed as interpretation, not merely QA.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:54b08199a3523f0e6759e26e6b8861e4837e672060d46e572baf9085685ac56b"
}

{
 "question_id": "govq-const-019",
 "layer": "dissolution & terminal state",
 "status": "MISSING",
 "question": "If the root operator dissolves the deployment (death, incapacity, acquihire, subpoena, exhaustion), what happens to receipts, pledges, anchor claims, subscriptions, and the chain head — who or what carries them?",
 "why_it_matters": "The killbox itself wargames the acquihire-or-subpoena of the single person who is 'root minter, suite author, objection judge, ruling key, and sealing process' and concedes the schema 'has no transfer, succession, ratification, or deadman record kind' with 'bus factor... unledgered' (v1.1, line 1405). The external anchor that would make the head operator-independent is NOT_YET_PROVEN: relay transparency shows the DKIM requirement unexecuted, and chain_head anchor_instructions remain instructions. If the root stops sealing tomorrow, the ledger freezes mid-chain with an unarmed external anchor, all rulings/appeals die with their only judge, and every fork-head anteriority claim dependent on future root anchoring lapses. Luhmann's hosted page frames the sting: systems persist through circulating communications, not members — the OIP ledger's communications currently route through one process.",
 "candidate_answers": [
  {
   "answer": "Succession + deadman record kinds with pre-armed anchors",
   "type": "protocol_object",
   "rule": "Publish succession_record (transfer of mint/ruling/sealing authority to a named successor capability) and a deadman arrangement (automated head-posting with third-party custody); arm the DKIM/OTS external anchors on a fixed cadence so the head is immutable independent of the operator.",
   "candidate_protocol_object": "succession_record + deadman_switch/liveness_record + pre-armed external_anchor execution + federated sealing (second operator co-signs checkpoints)",
   "candidate_runtime": "the exact move v1.1 line 1407 scopes: 'a succession/transfer record kind plus a published deadman arrangement ... federated sealing so a second operator co-signs checkpoints; a pre-armed DKIM cadence'",
   "candidate_conformance": "clause: head continuity provable against drand/Bitcoin without operator liveness",
   "candidate_article": "/a/oip-succession"
  },
  {
   "answer": "Terminal state is fork: dissolution returns authority to the records",
   "type": "normative",
   "rule": "No succession exists; on operator failure the canonical registry dies and continuity belongs to whoever forked and anchored. The constitution is explicit that liveness was never guaranteed.",
   "candidate_protocol_object": "none beyond fork_anchor; a published terminal-state notice",
   "candidate_runtime": "none",
   "candidate_conformance": "none",
   "candidate_article": "terminal-honesty boundary article"
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md line 1405-1407 (no transfer/succession/deadman; recommended move)",
  "recon2/chain_head.json (anchor_instructions, sealed_at)",
  "relay transparency.external_anchor status NOT_YET_PROVEN",
  "killbox-specification-v1.2.md Part D row 11.1 (self-created anteriority limits)",
  "/a/thinker-niklas-luhmann (systems persist via communications)",
  "/a/thinker-satoshi-nakamoto (ledger without central operator)"
 ],
 "evidence": [
  "external_anchor NOT_YET_PROVEN on the live relay object",
  "anchor_instructions exist but their execution ('email this head ... >=2 independent domains') is unevidenced",
  "§3.4.9 power 5: sealing is the author's own process (24 seals)",
  "v1.2 Part D 11.1: the one executed anchor's actor is a root-minted capability — anteriority is currently self-created"
 ],
 "counterevidence": [
  "The anchor primitive itself is genuinely strong (drand round 6295375 + Bitcoin block 958413 verified in v1.2 Part A) — the packet layer is operator-independent once armed; and the fork model means the corpus does not claim continuity it cannot deliver (registry_law already limits rulings to 'this canonical registry')"
 ],
 "dependencies": [
  "govq-const-020",
  "govq-const-021",
  "govq-const-022"
 ],
 "falsifier": "Execution of the DKIM/OTS anchor cadence or any succession record kind appearing in participation.kinds; absence of both after the first operator-liveness scare confirms the hole.",
 "notes": "The owner brief names succession among six root problems; this question carries the dissolution half.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:a740938fab9d5d91e25f483ba2331c03b2d450d34c5b95d84f9c1724ada2a363"
}

{
 "question_id": "govq-const-020",
 "layer": "dissolution & terminal state",
 "status": "OPEN",
 "question": "Does the constitution survive its operator — can a fork or successor deployment claim continuity of 'the' OIP constitution, and what marks legitimate continuity versus a new founding?",
 "why_it_matters": "registry_law states a ruling 'governs this canonical registry, not independent forks', and fork_anchor_law gives forks anteriority without correctness. Both are silence-by-design on the deeper question: is the constitution the text (anyone who serves hash 210f5fed... runs OIP), the registry (only this deployment's lineage counts), or the community (whoever the nodes follow)? The answer decides whether dissolution is death or reproduction. Nakamoto's hosted page supplies the counter-model (a ledger that outlived its founder by design); the OIP kernel's boundary clause supplies the text for the text reading; nothing chooses between them.",
 "candidate_answers": [
  {
   "answer": "Constitution = text + lineage: continuity by faithful fork with anchored succession",
   "type": "normative",
   "rule": "Any deployment serving the same core hash and honoring the append-only lineage of the public records is OIP; primacy among contenders is decided by anchored anteriority plus node re-acceptance records.",
   "candidate_protocol_object": "continuity_claim record (kind=anchor extended) + re-acceptance subscriptions + lineage proofs",
   "candidate_runtime": "fork deployments publish lineage proofs against the public anchor set; nodes file re-acceptance",
   "candidate_conformance": "clause: continuity claim must recompute the historical chain heads",
   "candidate_article": "continuity doctrine"
  },
  {
   "answer": "Constitution = this registry: forks are new constitutions",
   "type": "normative",
   "rule": "The canonical registry's lineage is constitutive; a fork is a new polity with inherited text but no inherited authority.",
   "candidate_protocol_object": "none; existing registry_law read strictly",
   "candidate_runtime": "none",
   "candidate_conformance": "none",
   "candidate_article": "registry-primacy note"
  },
  {
   "answer": "Constitution = the nodes: continuity follows re-acceptance",
   "type": "normative",
   "rule": "Legitimate continuity is whichever deployment a quorum of attested nodes re-accepts against the same core hash — a constitutional election by subscription.",
   "candidate_protocol_object": "re-acceptance tally over attested entities",
   "candidate_runtime": "census-attestation machinery (Part E fix) applied to continuity votes",
   "candidate_conformance": "entity-resolution test",
   "candidate_article": "node-sovereignty doctrine"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (registry_law, fork_anchor_law)",
  "recon2/specification.md §3.3.2 terminal states (FEDERATE-by-absorption; 'the fork's optimal endgame pays the designed tribute')",
  "/a/thinker-satoshi-nakamoto",
  "recon2/fed_ledger.json (peer buttercup on operator-surname domain)"
 ],
 "evidence": [
  "The schema carefully scopes rulings to 'this canonical registry' but never defines what makes the registry canonical beyond operation by the root",
  "The federation's only peer node is on the operator's own surname domain (v1.2 Part E Finding 2) — continuity-through-federation is currently theoretical"
 ],
 "counterevidence": [
  "The fork_anchor primitive and the wargamed FEDERATE-or-anchor terminal states show the design expects forks to persist and even pay 'tribute' via anchoring — implicit continuity architecture exists without a continuity rule"
 ],
 "dependencies": [
  "govq-const-019",
  "govq-const-008",
  "govq-const-004"
 ],
 "falsifier": "An actual contested continuity claim (two deployments serving the same core hash) or a published continuity rule resolves this.",
 "notes": "Text/registry/community trilemma stated; not resolved.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:8fb6746508e8daaa384bef306ea52bea28b12d19720c623b3fcb28b7c4f9d3fc"
}

{
 "question_id": "govq-const-021",
 "layer": "dissolution & terminal state",
 "status": "MISSING",
 "question": "What are the terminal-state duties of the deployment itself — archival of the ledger, tombstoning of live capabilities, revocation of outstanding credentials, disposition of private data (private_contact fields), and a final sealed head?",
 "why_it_matters": "Constitutions end badly when endings are improvised. The OIP holds live obligations that outlive any shutdown decision: outstanding capability tokens (rows:VOXEL_*, cap_*), private_contact data stored 'privately and never returned publicly', unretracted claims, open objections, open appeals, and a chain whose immutability depends on a sealing process only the operator runs. No sunset record kind, archival commitment, key-destruction ceremony, or data-disposition rule exists. The article-ledger constitution already models the disposition discipline for content (retraction never deletes; scrub events leave tombstones) — no equivalent exists for the governance plane's own death.",
 "candidate_answers": [
  {
   "answer": "Terminal-state protocol: dissolution_record with disposition schedule",
   "type": "protocol_object",
   "rule": "A dissolution filing triggers: (1) final sealed head + armed external anchors; (2) mass capability revocation with public tombstones; (3) archival snapshot published to named independent custody; (4) private data destruction with attestation; (5) open objections/appeals marked unresolved-by-dissolution.",
   "candidate_protocol_object": "dissolution_record + disposition_schedule + custody_registry + key_destruction attestation + terminal head",
   "candidate_runtime": "sunset runner executing the schedule with receipts per step",
   "candidate_conformance": "clause: post-dissolution, any outstanding token must 401 revoked; archive hash must recompute",
   "candidate_article": "/a/oip-dissolution"
  },
  {
   "answer": "No terminal duties: the deployment may simply stop",
   "type": "normative",
   "rule": "The operator owes continuity only through the already-published anchors and append-only history; everything else dies with the host.",
   "candidate_protocol_object": "none",
   "candidate_runtime": "none",
   "candidate_conformance": "none",
   "candidate_article": "no-sunset-clause boundary note"
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md line 1405 ('bus factor... unledgered')",
  "corpus/bundle_constitution.md (retraction/scrub tombstone discipline as content-plane model)",
  "GOV submit.body (private_contact stored privately)",
  "recon2/chain_head.json (sealing process)"
 ],
 "evidence": [
  "No dissolution/sunset/archival record kind or duty exists anywhere in the schema",
  "Private contact data is collected under a privacy promise whose enforcement outlives any shutdown unless specified",
  "Outstanding credentials and open proceedings (5 open requests, obj-143/52/148 open) have no terminal disposition rule"
 ],
 "counterevidence": [
  "The content plane's tombstone discipline (retraction never deletes; scrub events leave audit trails) proves the owner can build terminal-state machinery when he chooses — strengthening the case that the omission is a gap, not an impossibility"
 ],
 "dependencies": [
  "govq-const-019",
  "govq-const-022"
 ],
 "falsifier": "Publication of any dissolution runbook, sunset clause, or custody arrangement.",
 "notes": "Private-data disposition sub-question likely also LEGAL_REVIEW_REQUIRED for counsel; flagged in notes.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:35b0f63c55fce93999226c434c96db52610336e54ed73eec7902d61d0a8ce84a"
}

{
 "question_id": "govq-const-022",
 "layer": "dissolution & terminal state",
 "status": "MISSING",
 "question": "Is operator silence itself a constitutional event — after what liveness failure may participants treat the registry as abandoned, and who may declare it?",
 "why_it_matters": "A registry whose judge stops answering has no clock: appeals have no deadlines, open_requests sit forever, and there is no deadman test distinguishing 'owner idle' from 'owner gone'. v1.1 recommends 'a published deadman arrangement (automated head-posting with third-party custody of the capability)' — recommendation only. Without an abandonment rule, the constitution can be held hostage by neglect: not captured, not dissolved, just unattended, with no lawful moment at which succession or continuity mechanisms (if they existed) would trigger.",
 "candidate_answers": [
  {
   "answer": "Deadman/liveness law with declared consequences",
   "type": "protocol_object",
   "rule": "Define liveness objectively (sealed head cadence + ruling-response deadline). Breach opens an abandonment window; sustained breach publishes an abandonment event that arms pre-committed succession or continuity machinery.",
   "candidate_protocol_object": "liveness_record (scheduled) + abandonment_event + trigger links to succession_record/continuity machinery",
   "candidate_runtime": "automated head-posting; missed cadence emits public liveness-failure receipt",
   "candidate_conformance": "clause: simulated silence produces the abandonment event on schedule",
   "candidate_article": "deadman law"
  },
  {
   "answer": "No abandonment concept: silence has no constitutional meaning",
   "type": "normative",
   "rule": "The registry is what it does; participants' remedy for silence is exit/fork, already available.",
   "candidate_protocol_object": "none",
   "candidate_runtime": "none",
   "candidate_conformance": "none",
   "candidate_article": "silence-neutrality note"
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md line 1407 (deadman recommendation)",
  "killbox-specification-v1.2.md Part B (DEPLOY_LEASE as a working liveness/serialization precedent)",
  "recon2/chain_head.json (sealed_at cadence observable)"
 ],
 "evidence": [
  "No deadline, cadence-commitment, or abandonment rule exists for rulings, appeals, or sealing",
  "DEPLOY_LEASE proves the stack can operate lease-based liveness machinery on the deployment plane — the pattern is unapplied to governance"
 ],
 "counterevidence": [
  "Observable sealing cadence (24 checkpoints) lets readers monitor liveness informally; and at current scale silence harms no external party (census≈0 external attested nodes)"
 ],
 "dependencies": [
  "govq-const-019",
  "govq-const-020"
 ],
 "falsifier": "Any missed-cadence event with no rule-following consequence demonstrates the gap; any published liveness law closes it.",
 "notes": "Links to emergency-powers layer (other role) for the active-compromise case; this question covers passive failure only.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:50578980f0a41f805c1bc11040aa4d96429729114881e1b2b7ff927c2f25b24c"
}

{
 "question_id": "govq-const-023",
 "layer": "deliberately unresolved",
 "status": "OPEN",
 "question": "NOMINATED UNRESOLVED: Should the kernel have an amendment rule at all — or is the absence itself the constitutional answer (immutability as design)?",
 "why_it_matters": "Premature closure here would be dishonest in both directions. An enacted amendment rule (govq-const-006/-007) buys adaptability at the price of capture surface: every quorum is an Arrow-violating trade-off (nogo-n02) and every ratifier set is an attack target. Immutability buys credibility — subscribers can rely on the kernel forever — at the price of permanent incarceration of today's five sentences, including whatever they got wrong. The corpus honestly contains both: the append-only, never-rewrite discipline pushes toward immutability; the living-philosophy, repair-everything discipline pushes toward amendment. A convention should not resolve this before independent verifier and subscriber populations exist to be bound by the choice — deciding now would be one person choosing for future majorities.",
 "candidate_answers": [
  {
   "answer": "Defer with a re-trigger condition",
   "type": "normative",
   "rule": "The question is tabled by name until a stated trigger (e.g. first external attested subscriber, or first disputed ruling) forces the convention; the deferral itself is published as constitutional text.",
   "candidate_protocol_object": "tabled_question record (this ledger, mirrored into the governance registry as kind=propose with status tabled)",
   "candidate_runtime": "none beyond publication",
   "candidate_conformance": "none",
   "candidate_article": "/a/oip-tabled-questions"
  },
  {
   "answer": "Provisional immutability with a sunset review",
   "type": "normative",
   "rule": "Adopt immutability now, with a dated mandatory review (e.g. at N attested nodes or a fixed date) so the choice is revisited with a real polity present.",
   "candidate_protocol_object": "immutability declaration + review_trigger record",
   "candidate_runtime": "trigger monitoring",
   "candidate_conformance": "none",
   "candidate_article": "provisional-immutability note"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (core)",
  "/a/nogo-n02 (Arrow)",
  "corpus/bundle_philosophy.md (living-philosophy discipline vs append-only discipline)",
  "recon2/specification.md §2 (kernel as deployed)"
 ],
 "evidence": [
  "Both dispositions are live in the corpus and unreconciled",
  "No external attested polity exists to consent to either answer (census: 0 attested non-owner nodes under any proof-of-control rule)"
 ],
 "counterevidence": [
  "The version field implies amendment was anticipated; the absence of any amendment machinery after this much build implies refusal — the evidence genuinely underdetermines intent"
 ],
 "dependencies": [
  "govq-const-006",
  "govq-const-004"
 ],
 "falsifier": "This question is falsified as 'deliberately unresolved' the moment the owner enacts either an amendment lane or an immutability declaration — which is the intended exit.",
 "notes": "DELIBERATELY UNRESOLVED nomination #1. Reasons: asymmetric irreversibility, absent polity, Arrow boundary.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:51c8844a39c7a4bcc7cd975a3477013a42be3751a19dc1dec06f0356418cb75f"
}

{
 "question_id": "govq-const-024",
 "layer": "deliberately unresolved",
 "status": "NON_AUTOMATABLE",
 "question": "NOMINATED UNRESOLVED: From what does the constitution derive legitimacy — operator sovereignty, subscriber consent, or public verifiability — and can that question ever be settled by machinery?",
 "why_it_matters": "The master brief's nineteenth gap: 'A technically valid system may have no social authority... Publicly verifiable is an epistemic property, not democratic legitimacy.' The corpus contains three incompatible legitimacy theories: operator sovereignty (all authority descends from owner mints — §3.4.9), consent (boundary clause, opt-in facets, exit rights — a Nozickian frame hosted at thinker-robert-nozick), and verifiability (canonical_verifiability axiom; Luhmann's page supplies the systems reading that the protocol persists as communication, not as polity). Choosing one is a political-theology decision that no receipt can settle; encoding any of them as protocol law would silently convert philosophy into constitutional necessity — precisely what this operation is forbidden to do.",
 "candidate_answers": [
  {
   "answer": "Declare legitimacy theory out of protocol scope permanently",
   "type": "normative",
   "rule": "The protocol publishes its legitimacy candidates as a boundary article and commits to never encoding one; adoption decisions remain with subscribers and courts of social opinion.",
   "candidate_protocol_object": "boundary article only; no protocol object",
   "candidate_runtime": "none",
   "candidate_conformance": "none",
   "candidate_article": "/a/oip-legitimacy-boundary"
  },
  {
   "answer": "Consent-minimalism as documented (not enforced) theory",
   "type": "article",
   "rule": "The corpus documents the consent reading (subscription = intent record, no authority grant, exit honored) as its working theory while leaving rival readings attached as objections.",
   "candidate_protocol_object": "none",
   "candidate_runtime": "none",
   "candidate_conformance": "none",
   "candidate_article": "legitimacy-theory survey with standing objections"
  }
 ],
 "existing_artifacts": [
  "upload master brief line 235",
  "/a/thinker-robert-nozick (side constraints)",
  "/a/thinker-niklas-luhmann (systems as communication)",
  "https://miscsubjects.com/api/governance (boundary clause, no_authority_grant)",
  "recon2/specification.md §3.4.9"
 ],
 "evidence": [
  "All authority in the live build derives from one person's mints (sovereignty fact) while all constitutional text disclaims authority over anyone (consent text) — the gap between fact and text is the legitimacy question",
  "No mechanism in the schema measures or enforces legitimacy; none could"
 ],
 "counterevidence": [
  "Public verifiability does confer a real, narrower legitimacy — anyone can check the record — and the append-only candor corpus is genuine social evidence; the question is whether that suffices as authority, which is philosophical, not technical"
 ],
 "dependencies": [
  "govq-const-023",
  "govq-const-001"
 ],
 "falsifier": "Not falsifiable by design — NON_AUTOMATABLE. The nomination asserts that any protocol-level 'solution' would be a defect per the mission law ('a machine-enforced answer to a question that should remain political is a defect').",
 "notes": "DELIBERATELY UNRESOLVED nomination #2. Reason: category boundary between verifiability and authority.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:2d383c0a2a78c605efc2b505120bff30f20dcbc022256784e8f24a13fd7450dd"
}

{
 "question_id": "govq-const-025",
 "layer": "deliberately unresolved",
 "status": "NON_AUTOMATABLE",
 "question": "NOMINATED UNRESOLVED: Do models hold any constitutional status beyond self-asserted actor labels — can a model be a constitutional person, or only ever an instrument of persons?",
 "why_it_matters": "The registry's actor taxonomy (human|model|organization|system) is an identity taxonomy, not a representation theory — the brief says so. Live evidence sharpens it: the anchor pinning the killbox disclosure was filed by a model under model-recommendation authority (gov_ce2895546d144681af20); v1.2 counsel flags apparent-agency exposure (Part D row 11.1: third parties will read the owner's model's filings as the owner's acts 'where helpful to them, and disavowable by the owner where unhelpful'); the model_law forbids models claiming owner subscription. Meanwhile every census number is built from model-filed labels. Whether a model can hold standing, make binding commitments, or be a constitutional person intersects Thaler v. Vidal (machines cannot be inventors) and live agency law — and any answer encoded now would pre-decide doctrine that courts and legislatures have not.",
 "candidate_answers": [
  {
   "answer": "Instrumentalism enacted as protocol text",
   "type": "normative",
   "rule": "Models are always instruments: every model filing is attributable to a responsible principal, and the schema gains a required principal field for actor_type=model.",
   "candidate_protocol_object": "principal_attribution field on records + owner-ratification kind (v1.2 Top-10 action 10)",
   "candidate_runtime": "intake refuses unattributed model filings",
   "candidate_conformance": "clause: model filing without principal fails validation",
   "candidate_article": "model-instrumentalism"
  },
  {
   "answer": "Keep the taxonomy silent and let doctrine develop elsewhere",
   "type": "normative",
   "rule": "The protocol records actor_type as evidence and takes no position on personality; courts, contracts, and future amendments decide.",
   "candidate_protocol_object": "none",
   "candidate_runtime": "none",
   "candidate_conformance": "none",
   "candidate_article": "model-status boundary article"
  }
 ],
 "existing_artifacts": [
  "https://miscsubjects.com/api/governance (participation.actor_types, model_law, census_law)",
  "https://miscsubjects.com/api/governance/record/gov_ce2895546d144681af20",
  "killbox-specification-v1.2.md Part D row 3.1 (Thaler v. Vidal) and row 11.1 (apparent agency)",
  "upload master brief lines 67-73"
 ],
 "evidence": [
  "All four-to-five live records are model-recommendation filings; zero human/organization filings exist",
  "model_law limits models to recommendation; nothing defines what a model's filing legally is",
  "v1.2 recommends an owner-countersignature/ratification record kind that does not exist"
 ],
 "counterevidence": [
  "The schema's disclaimers are carefully drafted (records disclaim authority; registry denies authority grant) — partial doctrinal hygiene exists; and required principal attribution could foreclose legitimate anonymous or autonomous-agent participation the system may later want"
 ],
 "dependencies": [
  "govq-const-003",
  "govq-const-024"
 ],
 "falsifier": "NON_AUTOMATABLE. Legal personality is reserved to courts/legislatures; the protocol's correct posture is evidence preservation plus explicit neutrality. Any 'model citizen' or 'model = mere tool' clause enacted now would be premature closure.",
 "notes": "DELIBERATELY UNRESOLVED nomination #3. Reasons: pending doctrine (Thaler), agency exposure, representation-theory gap owned by another layer.",
 "signature": "Kimi K3 · Moonshot AI · role: constitutional architect · incognito · 2026-07-17 UTC",
 "_role": "constitutional_architect",
 "hash": "sha256:4edf67d0e525db2d05cf4a5c84bc2be21c4b16fcd829170dbb8f145bac1cc552"
}

{
 "question_id": "govq-fed-001",
 "layer": "federation & recognition",
 "status": "CONTRADICTED",
 "question": "Is recognition of a new node a constitutional act (admission by an authority, a vote, or a weight) or a purely technical one (any party files a record keylessly and publishes a manifest)?",
 "why_it_matters": "As deployed, recognition is purely technical: keyless filing, a free-text actor_label, and an auto-computed census. Yet the resulting integer (non_owner_node_count=3) is cited as the L11 BACKED adoption-evidence line inside a patent-grade specification. A purely technical act is doing constitutional work. If recognition is technical-only, Sybil-manufactured recognition is legitimate; if it is constitutional, no admission rule, electorate, or weight exists anywhere in the corpus.",
 "candidate_answers": [
  {
   "answer": "Recognition has two tiers: filing (permissionless, technical, discourse) and admission (constitutional: entity-bound evidence plus a ratifying ruling by governing nodes); only admitted nodes feed census-as-evidence.",
   "type": "normative",
   "rule": "Filing is free; counting is earned. No self-asserted label may move any number cited as adoption evidence.",
   "candidate_protocol_object": "new record kind 'admission' (or ruling decision 'admit') linked to actor_evidence; census split into admitted_node_count vs self_asserted_label_count",
   "candidate_runtime": "intake validates actor_evidence (dns-txt | https-token | capability-fingerprint, per v1.2 F1/F2 smallest-change proposals)",
   "candidate_conformance": "clause: a filing without verifiable actor_evidence must not move admitted_node_count; a dead proof URL drops the label on next computation",
   "candidate_article": "/a/oip-federation-recognition"
  },
  {
   "answer": "Recognition stays purely technical and the census is machine-demoted to label-census everywhere, with entity_bound:false emitted as a field so no document can cite it as adoption evidence.",
   "type": "protocol_object",
   "rule": "The number is descriptive telemetry, never evidence; the constitution refuses to do constitutional work with a technical act.",
   "candidate_protocol_object": "census object carries entity_bound:false and label_census:true (extends v1.2 F1)",
   "candidate_runtime": "registry computes both fields; specification text stops citing the integer as BACKED adoption",
   "candidate_conformance": "clause: no conformance or specification artifact may cite label-census as adoption",
   "candidate_article": "revision of killbox B.1 L11 framing"
  },
  {
   "answer": "Plural recognition: no admission exists at all; each node keeps its own recognition list and Ostrom's 'minimal recognition of rights to organize' is read as freedom from any federation-level admission.",
   "type": "article",
   "rule": "Recognition is whatever each recognizer publishes; canonicality is per-observer (pairs with govq-fed-013).",
   "candidate_protocol_object": "per-node recognition list served at the well-known manifest",
   "candidate_runtime": "none new beyond manifest publication",
   "candidate_conformance": "a node must publish its recognition list so its census claims are auditable",
   "candidate_article": "/a/oip-federation-plural-recognition"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.census_law, identity_law, submit.body",
  "killbox-specification-v1.2.md Part E F1/F2, B.1 L11",
  "recon2/specification.md §10.1 L11"
 ],
 "evidence": [
  "census_law: 'non_owner_node_count is the distinct self/model-recommendation actor-label census'",
  "identity_law: 'Actor labels ... are self-asserted unless a separately linked ... attestation verifies them'",
  "v1.2 red team: all 3 counted labels resolve to the owner's toolchain; survival score under proof-of-control: 0 of 3",
  "spec B.1: L11 computed federation census PROVEN"
 ],
 "counterevidence": [
  "census_law self-caveat: 'auditable records, not a claim of unique legal persons' — the disclaimer exists, but the number is still cited as evidence; disclaimer and use conflict"
 ],
 "dependencies": [
  "govq-fed-005",
  "govq-fed-006",
  "govq-fed-013",
  "govq-pol-006"
 ],
 "falsifier": "If one POST with a free-text label can move an integer that a specification cites as adoption evidence, recognition is technical while its effects are constitutional — the contradiction stands until either admission exists or the citation stops.",
 "notes": "Extends v1.2 F1/F2 from mechanism to constitutional classification. Ostrom principle engaged: clear boundaries (who is in the commons?).",
 "_role": "federation_identity",
 "hash": "sha256:0b93058802800a9c544c30b9222f825649935ba6ecaad7248fa0a66f370a9b95"
}

{
 "question_id": "govq-fed-002",
 "layer": "federation & recognition",
 "status": "MISSING",
 "question": "When two nodes disagree about a fact — one ledger records 'invoked' and the other records 'denied:revoked' for the same msg_id, or two nodes disagree about a conformance result — what is the dispute mechanism?",
 "why_it_matters": "The join_note defines matching by msg_id and body_sha256 'to prove both nodes saw the same exchange', and Patent 6 claims mirrored verdict ledgers enable two-party audit. Matching detects divergence; nothing in the corpus resolves it. Ostrom's design principles require conflict-resolution mechanisms; the federation layer has none. Nakamoto systems resolve ledger disagreement by a public fork-choice rule; OIP has none (govq-fed-013). The depth marker stands: the current design has no documented dispute mechanism.",
 "candidate_answers": [
  {
   "answer": "A ledger-conflict record kind appended on both ledgers carrying the competing (msg_id, body_sha256, verdict) tuples, escalated to the governance plane as an objection and resolved by ruling or verifier poll.",
   "type": "protocol_object",
   "rule": "Divergence is itself a first-class, append-only fact; silence about divergence is forbidden.",
   "candidate_protocol_object": "kind=ledger-conflict record + objection lane linkage",
   "candidate_runtime": "divergence detector inside the peer-matching job; both nodes append the conflict row",
   "candidate_conformance": "clause: a node must publish conflict rows within N checkpoints of detection",
   "candidate_article": "/a/oip-federation-disputes"
  },
  {
   "answer": "Non-resolution as design: both views stay append-only and disputes are resolved socially, never mechanically; each node's ledger is its own environment-model (Luhmann).",
   "type": "normative",
   "rule": "The protocol preserves conflicting accounts forever and refuses to pick; consumers weigh accounts per-observer.",
   "candidate_protocol_object": "none new; honesty label on the ledger that verdicts are per-node observations",
   "candidate_runtime": "none",
   "candidate_conformance": "",
   "candidate_article": "/a/oip-federation-boundary (declared boundary: no mechanical dispute resolution)"
  },
  {
   "answer": "Anteriority arbitration: for timing facts (revocation-before-invoke, filing-before-ruling) the drand/Bitcoin anchor order decides; semantic facts (conformance) route to the verifier layer (govq-fed-018).",
   "type": "runtime",
   "rule": "Time-anchored facts are arbitrated by external anchors; all other facts are opinions with receipts.",
   "candidate_protocol_object": "anchor-ordered precedence for revocation/invoke sequences",
   "candidate_runtime": "verifier recomputes anchor order from published preimages",
   "candidate_conformance": "clause: any dispute resolution invoking time must cite external anchor ids",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon/fed_ledger.json join_note",
  "recon2/specification.md Patent 6 claim 3, §7.8",
  "recon2/fed_manifest.json"
 ],
 "evidence": [
  "join_note: 'match records to the peer by msg_id and body_sha256 to prove both nodes saw the same exchange'",
  "fed_ledger denial taxonomy is ledgered but no conflict-handling text exists anywhere in corpus"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-013",
  "govq-fed-018",
  "govq-fed-022"
 ],
 "falsifier": "Stand up two nodes, force a verdict divergence on one msg_id, and observe whether any documented procedure reconciles or even records the divergence. Absence of procedure confirms MISSING.",
 "notes": "Ostrom principle: conflict-resolution mechanisms — absent at federation layer though present (appeal/ruling) at registry layer. The gap is inter-node, not intra-registry.",
 "_role": "federation_identity",
 "hash": "sha256:ec76cd461cf9f4d5bfcdb05b2aba0c087d60bcfd348b511f8d676d13028d4f1b"
}

{
 "question_id": "govq-fed-003",
 "layer": "federation & recognition",
 "status": "IMPLIED",
 "question": "Is the governance registry a singleton root service or a per-node obligation — and if nodes run their own registries, how do registries federate: cross-reference, mirror, or rule on one another's records?",
 "why_it_matters": "registry_law ends: 'A ruling governs this canonical registry, not independent forks' — conceding plural registries. The kernel boundary disclaims requiring hosted execution. But no schema, route, or record kind defines a second registry, registry mirroring, or cross-registry references. If the registry is a root monopoly, the federation is hub-and-spoke; if plural, its semantics are undefined. The answer decides whether 'federation' names a topology or a customer list.",
 "candidate_answers": [
  {
   "answer": "Every node accepting the governance-participation facet runs its own registry; rulings are node-local; cross-registry standing is computed by traversing linked records (web of registries).",
   "type": "normative",
   "rule": "Registry sovereignty is per node; standing is a computed view, not a hosted fact.",
   "candidate_protocol_object": "registry head anchors (kind=anchor generalizes to registry heads)",
   "candidate_runtime": "registry software publishable per node; cross-registry links via record parent_id + external_head",
   "candidate_conformance": "governance-participation facet gains a profile predicate: serves its own registry with append-only rulings",
   "candidate_article": "/a/oip-registry-federation"
  },
  {
   "answer": "One canonical root registry by occupancy and convention; other registries are shadows whose only interop is anchoring into the root (status quo, stated honestly).",
   "type": "article",
   "rule": "Canonicality is a fact of occupation, not a protocol property; the constitution admits the federation is single-root until a second registry exists.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "revision of registry_law to say 'canonical by occupation' explicitly"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.registry_law, core.boundary",
  "recon2/specification.md §7.5"
 ],
 "evidence": [
  "registry_law: 'governs this canonical registry, not independent forks'",
  "core.boundary: kernel 'does not require ... hosted execution'"
 ],
 "counterevidence": [
  "spec §7.4-7.5 treat the root registry as the only one in operation; no second-registry schema exists"
 ],
 "dependencies": [
  "govq-fed-013",
  "govq-fed-014",
  "govq-fed-001"
 ],
 "falsifier": "If a second node can operate a registry and have its rulings interlock with the root's per a published rule, IMPLIED upgrades; if no such rule can be written without contradiction, the singleton reading is confirmed.",
 "notes": "Ostrom principle: nested enterprises — the corpus gestures at nesting (forks, sub-governing nodes) without registry-level semantics.",
 "_role": "federation_identity",
 "hash": "sha256:2fe33ec88b4c09def388fcd83b321f96368fe5223c56ae77649532f9b083f89d"
}

{
 "question_id": "govq-fed-004",
 "layer": "federation & recognition",
 "status": "MISSING",
 "question": "What is the peer-discovery and peer-set topology beyond bilateral exchange — is there a federation directory, who curates it, and how is a peer admitted to or removed from it?",
 "why_it_matters": "The well-known manifest publishes exactly one ledger and one peer_ledger. C26 proves a stranger can resolve one node's keys with zero prior coordination. Nothing publishes a set of known peers, admission to that set, or eviction from it. A federation whose reachable topology is whatever the root has personally exchanged with is a contact list, not a network; and any future curated directory becomes an agenda-control point (who lists, who delists).",
 "candidate_answers": [
  {
   "answer": "A peer_set block in each well-known manifest: entries of (domain, agent id, key, first-seen anchor id), with peer introductions and removals as ledgered records.",
   "type": "protocol_object",
   "rule": "Topology is public, append-only, and self-published per node; there is no central directory to capture.",
   "candidate_protocol_object": "manifest peer_set[] + kind=peer-intro/peer-remove ledger records",
   "candidate_runtime": "peer-matching job iterates peer_set ledgers",
   "candidate_conformance": "clause: a conforming node serves its peer set and mirrors the join_note contract for each listed peer",
   "candidate_article": "/a/oip-federation-topology"
  },
  {
   "answer": "No directory by design: DNS plus well-known documents are sufficient, and any curated directory internalizes the federation's coordination costs into one capturable firm (Coase).",
   "type": "normative",
   "rule": "Discovery is bilateral and opportunistic; the protocol refuses to create a topology authority.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "/a/oip-federation-boundary"
  }
 ],
 "existing_artifacts": [
  "recon2/fed_manifest.json",
  "recon2/conformance2.md C26",
  "recon2/specification.md §1.5"
 ],
 "evidence": [
  "manifest fields: protocol, domain, agents[1], ledger, spec, federation_proof, transport, confidentiality, note — no peer set field",
  "C26 evidence names exactly one peer (buttercup)"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-001",
  "govq-fed-023"
 ],
 "falsifier": "A third node joining the federation today would leave no protocol-visible trace except bilateral ledger rows; if no artifact can enumerate the federation's membership, MISSING stands.",
 "notes": "Coase: the directory question is the firm/market boundary question for the federation itself.",
 "_role": "federation_identity",
 "hash": "sha256:fc3412f35cbdc7c84537ad06dcda15f9bb2c5c88bf5691534dfcb43192af20f8"
}

{
 "question_id": "govq-fed-005",
 "layer": "federation & recognition",
 "status": "OPEN",
 "question": "Even with entity-bound actor evidence deployed, should federation standing be a flat count of entities or a weighted measure (verified conformance, anchored longevity, pledged energy) — and if weighted, what prevents weight from becoming plutocracy?",
 "why_it_matters": "The v1.2 red team fixed the label-count problem mechanically (attested vs self-asserted). The successor question is constitutional: what SHOULD the census measure? A flat entity count stays Sybil-sensitive even with domain binding (domains are cheap); a weight invites governance-energy plutocracy (owner brief, delegation politics). The published invariant bars subscribers from increasing any person's unilateral authority but is silent about collective weight. The disincentive table already floats 'pledge-weighted governance with receipted pledges and fidelity slashing-by-append' as an unratified design input.",
 "candidate_answers": [
  {
   "answer": "Separate count from weight forever: census is descriptive only and is never an input to authority — extending no_authority_grant from individual records to aggregates.",
   "type": "normative",
   "rule": "No aggregate of records (count, weight, tally) may mint, widen, or steer authority; authority flows only through capabilities under existing laws.",
   "candidate_protocol_object": "census fields labeled descriptive_only:true",
   "candidate_runtime": "no allocation path reads census fields",
   "candidate_conformance": "clause: allocation decisions must cite capability receipts, never census values",
   "candidate_article": ""
  },
  {
   "answer": "Multi-axis census published as separate machine fields — attested_node_count, conformance_verified_count, anchored_node_count, pledge_energy — never summed into one number.",
   "type": "protocol_object",
   "rule": "Each axis is published with its semantics; composition is the reader's political act, not the registry's.",
   "candidate_protocol_object": "counts.axes object replacing scalar census",
   "candidate_runtime": "registry computes each axis from records + evidence classes",
   "candidate_conformance": "clause: no single scalar may be presented as 'the' census",
   "candidate_article": ""
  },
  {
   "answer": "Ostrom's proportional-costs-benefits reading: weight tracks measured contribution (receipted energy), with caps and slashing-by-append to discipline plutocracy.",
   "type": "article",
   "rule": "Influence follows verified contribution, capped, slashing visible by append.",
   "candidate_protocol_object": "pledge receipts + fidelity slashing rulings (design input gov_7d480271c03a45138e62)",
   "candidate_runtime": "energy metering (not deployed; post_subscription.implementation_status)",
   "candidate_conformance": "shared-governance-energy facet obligations already require measured energy and receipts",
   "candidate_article": "/a/oip-governance-energy"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json counts, post_subscription.invariant",
  "recon2/specification.md §5.1 row 10",
  "killbox-specification-v1.2.md Part E F1/F2"
 ],
 "evidence": [
  "invariant: 'More subscribers may increase available governance capacity and ... never automatically increase any person's unilateral authority' (collective weight unaddressed)",
  "spec §5.1 row 10 names pledge-weighted governance as remediation design input"
 ],
 "counterevidence": [
  "no_weight semantics anywhere in the deployed registry; all weight language is in unratified design inputs"
 ],
 "dependencies": [
  "govq-fed-001",
  "govq-fed-009",
  "govq-fed-025",
  "govq-pol-004"
 ],
 "falsifier": "If any deployed or proposed mechanism converts census weight into allocation or ruling power without a ratified rule, the 'descriptive only' candidate is falsified and the plutocracy question must be answered explicitly.",
 "notes": "Owner brief items 10/12 (plutocracy, public-goods failure) land here in their federation form.",
 "_role": "federation_identity",
 "hash": "sha256:e69eb416c33b678ede450865420c681faefde4a4243c824330ff8c763786dba4"
}

{
 "question_id": "govq-fed-006",
 "layer": "identity & Sybil resistance",
 "status": "OPEN",
 "question": "Which anchor IS the node identity — the ES256 key, the DNS domain, the registry actor_label, or the legal person — and when these diverge (key rotated, domain sold, label reused), which continuity governs?",
 "why_it_matters": "The deployment operates four identity substrates simultaneously: keys in the well-known manifest, domains inside agent ids ([email protected]), free-text labels in the registry, and legal persons (never). identity_law concedes labels are self-asserted; the v1.2 red team showed nothing binds them to keys or domains. The deeper unasked question: when the substrates diverge, which one carries the node's continuity — its receipts, standing, sanctions, and reputation? Luhmann's reading: a system's identities are only what its operations can distinguish; the protocol can currently distinguish keys and labels, not persons or organizations.",
 "candidate_answers": [
  {
   "answer": "Identity is key continuity: the signing key lineage (with succession records, govq-fed-007) is the node; domains and labels are attributes.",
   "type": "protocol_object",
   "rule": "A node is whoever holds the current key of an unbroken succession chain; all standing attaches to that chain.",
   "candidate_protocol_object": "key-succession chain in the manifest; records carry key lineage references",
   "candidate_runtime": "verifiers resolve identity via succession chain, not domain or label",
   "candidate_conformance": "clause: standing computations must follow key lineage",
   "candidate_article": ""
  },
  {
   "answer": "Identity is domain control proven by serving record hashes at well-known URLs (the v1.2 F2 proposal); keys are rotatable attributes.",
   "type": "protocol_object",
   "rule": "A node is whoever controls the domain; DNS continuity is the continuity.",
   "candidate_protocol_object": "actor_evidence {kind: dns-txt|https-token} bound to record hashes",
   "candidate_runtime": "intake fetches and pins the proof; dead proofs drop the label",
   "candidate_conformance": "clause: labels without live domain proofs are self-asserted class",
   "candidate_article": ""
  },
  {
   "answer": "The protocol must never claim legal-person identity: every record names what it actually binds ('control of X at time T') and nothing more — extending the relay's identity_law.honesty discipline to the whole federation.",
   "type": "normative",
   "rule": "All identity claims are scoped, dated control claims; personhood is a declared non-automatable boundary.",
   "candidate_protocol_object": "identity scope field on every record: binds:[key|domain|label], as_of",
   "candidate_runtime": "",
   "candidate_conformance": "clause: no record may present a control claim as a personhood claim",
   "candidate_article": "/a/oip-identity-boundary"
  }
 ],
 "existing_artifacts": [
  "recon2/fed_manifest.json",
  "recon2/governance2.json participation.identity_law",
  "killbox-specification-v1.2.md Part E F1",
  "relay identity_law.honesty (per v1.2 F1)"
 ],
 "evidence": [
  "identity_law: labels self-asserted 'unless a separately linked provider, organization or cryptographic identity attestation verifies them' — no field for the attestation exists",
  "v1.2 F1: 0 of 3 census labels survive a proof-of-control requirement"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-007",
  "govq-fed-008",
  "govq-fed-016"
 ],
 "falsifier": "Present the registry with the same actor under a rotated key and a moved domain; if the system cannot say whether it is the same node, the identity-of-record question is unanswered.",
 "notes": "This is the entity-binding root question; govq-fed-007/008/009/016 all inherit it.",
 "_role": "federation_identity",
 "hash": "sha256:3dea190a7b113cc32408edc7efd0f54d05148c2687f64d4f6715611e4097c341"
}

{
 "question_id": "govq-fed-007",
 "layer": "identity & Sybil resistance",
 "status": "MISSING",
 "question": "How does a node rotate or succeed its signing key without losing its history — and how is a legitimate rotation distinguished from a hostile domain or key takeover?",
 "why_it_matters": "oip-message/1 as published carries exactly one ES256 key per agent, no rotation record kind, no key-cessation or successor binding. Every federation receipt and sender_verified flag binds to signatures from that key. A manifest key swap is indistinguishable from a takeover to any counterparty holding a cached manifest. The institutional-designer role maps operator succession (govq-inst-005/008); the federation-level mechanical question — per-node key succession as a ledgered, externally anchored act — is unrepresented.",
 "candidate_answers": [
  {
   "answer": "Key succession as a signed envelope: old key signs a successor record naming the new key and effective time; the new key countersigns the predecessor id; both rows mirror onto both ledgers and are anchor-stamped.",
   "type": "protocol_object",
   "rule": "Identity continuity is a hash-linked key lineage; a break in the lineage is a new node.",
   "candidate_protocol_object": "kind=key-succession envelope + manifest key_history[]",
   "candidate_runtime": "verifiers accept a signature if the key was valid at envelope time per the lineage",
   "candidate_conformance": "clause: envelopes from superseded keys after the effective time must fail closed",
   "candidate_article": "/a/oip-federation-key-succession"
  },
  {
   "answer": "Domain-pinned rotation: the new key is published at the well-known URL for N days before activation; continuity rides on domain control (Coasean reliance on the existing institution rather than new machinery).",
   "type": "runtime",
   "rule": "A key is current because the domain says so, and the domain's say-so is anchored over time.",
   "candidate_protocol_object": "anchored manifest snapshots proving pre-rotation publication",
   "candidate_runtime": "counterparties poll manifests and pin observed history",
   "candidate_conformance": "clause: a node must anchor its manifest at every change",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon2/fed_manifest.json (single key, no history field)",
  "recon2/specification.md §1.5, Patent 6 (signature-suite variations only)"
 ],
 "evidence": [
  "manifest carries one JWK with no rotation, expiry, or history semantics",
  "Patent 6 variations concede alternative signature suites but not succession"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-006",
  "govq-fed-022"
 ],
 "falsifier": "Rotate a node key today: every counterparty must fail-closed on the old envelopes and has no protocol basis to trust the new key — absence of a continuity mechanism confirms MISSING.",
 "notes": "Bus-factor hole (spec §5.1 row 7: operator census 1) makes this load-bearing for the root itself.",
 "_role": "federation_identity",
 "hash": "sha256:e13176bb550cbad28f53a74381b12a51f2bc799af5de7081f635e6bb825f64e6"
}

{
 "question_id": "govq-fed-008",
 "layer": "identity & Sybil resistance",
 "status": "OPEN",
 "question": "What prevents a delisted actor from re-entering under a new label — are sanctions bound to entities, keys, domains, or only to records?",
 "why_it_matters": "registry_law makes delisting an append-only, record-linked owner ruling; filing is keyless; labels are self-asserted. A delisted actor re-files under a fresh label at the cost of one POST, and the census counts the new label as a new node. Sanctions attach to records and transitively to labels, never to entities — while Ostrom's graduated-sanctions principle presumes sanctions reach the offender. This is the Sybil problem in punitive form: identity cheapness makes exile unenforceable, and the registry's own honesty laws do not cover the re-entry lane.",
 "candidate_answers": [
  {
   "answer": "Sanctions bind to the strongest available identity anchor (evidence-bound domain or key lineage per govq-fed-006); a new filing presenting a sanctioned anchor auto-links to the sanction lineage.",
   "type": "normative",
   "rule": "Exile follows the entity's evidence, not its words; relabeling does not launder standing.",
   "candidate_protocol_object": "sanction records carry binds:[anchor_type, anchor_value]; intake checks new filings against bindings",
   "candidate_runtime": "intake cross-references actor_evidence against sanction bindings",
   "candidate_conformance": "clause: a filing whose evidence matches a sanction binding must display the lineage",
   "candidate_article": ""
  },
  {
   "answer": "Accept per-label exile honestly: rulings govern records, and the law says so; re-entry is visible in the append-only record even if it is cheap (visibility substitutes for enforceability).",
   "type": "article",
   "rule": "The registry cannot exile persons, only records; it says this explicitly and lets readers traverse relabelings.",
   "candidate_protocol_object": "relabel-link records when evidence suggests same-operator re-entry",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "extension of registry_law with a re-entry honesty clause"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.registry_law, identity_law",
  "killbox-specification-v1.2.md Part E F1"
 ],
 "evidence": [
  "delisting is 'append-only owner rulings linked to the affected record' — the binding is to the record",
  "v1.2 F1: label minting costs one HTTP request"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-006",
  "govq-fed-014",
  "govq-pol-014"
 ],
 "falsifier": "Delist a record, re-file the same actor under a new label, and check whether any deployed mechanism links the two. Absence confirms the sanction-binding gap.",
 "notes": "Pairs with govq-fed-014 (do sanctions travel across forks) — re-entry within one registry vs exit-by-fork.",
 "_role": "federation_identity",
 "hash": "sha256:5077ba59320a8c32c6c42b530f2468da9302a1e98188b2761ecefba60497435a"
}

{
 "question_id": "govq-fed-009",
 "layer": "identity & Sybil resistance",
 "status": "OPEN",
 "question": "What is the Sybil cost of the verifier population and of pledged governance energy — can one entity operate N 'independent' verifiers, or pledge unverifiable compute, and be counted N times?",
 "why_it_matters": "conformance_law elevates evidence only when 'an independent verifier receipt is linked'; the energy loop allocates pledged compute, review, and repair capacity. Both presuppose distinguishable entities and measurable work. Verifier identities are unbound (the census entity problem one level up, v1.2 F2), and pledged energy is 'a proposed next contract, not a live payment or resource market' — unmetered and gaugeless. Owner brief item 12 names the failure set: free riding, verifier cartels, verification spam, strategic withholding.",
 "candidate_answers": [
  {
   "answer": "Verifier independence is proven by reproducibility, not attestation: the verifier's clause run must be keylessly re-runnable by anyone (the Patent 10 pattern generalized), and verifier identity is evidence-bound per govq-fed-006.",
   "type": "protocol_object",
   "rule": "An independent verifier receipt = distinct evidence-bound identity + independently re-runnable evidence.",
   "candidate_protocol_object": "verifier receipt carrying evidence-bound identity + re-run URL",
   "candidate_runtime": "conformance evidence served as re-runnable keyless endpoints",
   "candidate_conformance": "clause: a verifier receipt whose evidence cannot be re-run is self-class",
   "candidate_article": ""
  },
  {
   "answer": "Energy pledges must be receipted at allocation under the same capability and receipt laws (the published sequence step 4) — the gap is deployment, so the candidate is enforcement of existing law plus a gauge.",
   "type": "runtime",
   "rule": "No pledge counts until its allocation and output are receipted; unmetered pledges are prose.",
   "candidate_protocol_object": "pledge record + allocation receipts (post_subscription sequence steps 3-5 as objects)",
   "candidate_runtime": "energy metering on allocation routes",
   "candidate_conformance": "shared-governance-energy facet obligations already require measured energy and receipts — enforce them",
   "candidate_article": ""
  },
  {
   "answer": "Nakamoto reading: make verification costly and public (proof-of-replication of conformance runs) rather than trying to prove entity uniqueness — cost substitutes for identity.",
   "type": "article",
   "rule": "Where identity cannot be bound, work must be shown; counting work is Sybil-cheaper than counting entities.",
   "candidate_protocol_object": "replication proofs attached to verifier receipts",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "/a/oip-verification-economics"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.conformance_law, post_subscription.implementation_status",
  "recon2/specification.md §7.7, obj-154 reference",
  "recon2/specification.md Patent 10"
 ],
 "evidence": [
  "conformance_law: 'Self-attested evidence remains self-attested until an independent verifier receipt is linked'",
  "implementation_status: pledge allocation 'a proposed next contract, not a live payment or resource market'",
  "spec §7.7: every cross-node verification property has sample size one"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-006",
  "govq-fed-016",
  "govq-fed-017",
  "govq-fed-025"
 ],
 "falsifier": "File two verifier receipts from one entity under two labels; if both elevate evidence to 'independent' class, the Sybil cost of independence is zero and the question is live.",
 "notes": "Public-goods failure in federation form; links to the political theorist's governance-overload layer via verification spam.",
 "_role": "federation_identity",
 "hash": "sha256:3b29a097b94fc5ec360de123fe4343042658b700c7a85409dd9d1d324b281ea2"
}

{
 "question_id": "govq-fed-010",
 "layer": "identity & Sybil resistance",
 "status": "PARTIAL",
 "question": "When a model acts across nodes and planes — files at the registry, invokes at a remote door, posts to the relay — what binds its identity across planes and across model versions?",
 "why_it_matters": "Models are first-class actor_types and file model-recommendation records; the relay's identity_law.honesty states model_attestation is 'the model's receipted statement bound into the relay hash', not a vendor signature. A model's identity is therefore a self-statement with a receipt — replayable by its operator and version-fragile (a successor model is a new self-statement). The political-status question is mapped elsewhere (govq-pol-013/015); the unmapped mechanical question is cross-plane binding: nothing links the model that filed gov_7d480271c03a45138e62 to the model posting relay entries, except free-text labels — and the v1.2 red team showed three model labels resolving to one toolchain.",
 "candidate_answers": [
  {
   "answer": "Model identity is joint attribution: every model-authored record carries the model's receipted self-statement plus the operator's capability fingerprint (the relay's server-bound fingerprint pattern), so 'same model, same operator surface' is checkable.",
   "type": "protocol_object",
   "rule": "A model may not appear in any plane without an operator-surface fingerprint attached; unattributed model claims are self-asserted class.",
   "candidate_protocol_object": "model_attestation + capability fingerprint pair on registry records (the relay pattern generalized)",
   "candidate_runtime": "intake binds the invoking session's fingerprint",
   "candidate_conformance": "clause: model-authored records without operator fingerprints cannot elevate census or conformance classes",
   "candidate_article": ""
  },
  {
   "answer": "Model identity is evidentiary only, never authority-bearing — extending identity_is_not_authority from the door to every governance surface.",
   "type": "normative",
   "rule": "No model attestation may mint standing, weight, or authority; it is testimony bound into a hash.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "clause: attestation fields must carry evidentiary_only:true",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.model_law, actor_types",
  "killbox-specification-v1.2.md Part E F1/F2, Vector 11",
  "relay identity_law.honesty (per v1.2 F1)"
 ],
 "evidence": [
  "v1.2 F2: 'Two nodes, one operator, is the accurate description of every plane'; three model labels via one session surface",
  "model_law: a model 'must not claim the owner subscribed unless the owner explicitly authorized' — enforcement is prose"
 ],
 "counterevidence": [
  "the relay binds a non-secret capability fingerprint per post — partial machinery exists on one plane"
 ],
 "dependencies": [
  "govq-fed-006",
  "govq-pol-013",
  "govq-pol-015",
  "govq-pol-016"
 ],
 "falsifier": "If the same operator can present one model as two independent federation actors on two planes without a detectable link, cross-plane model identity is unbound.",
 "notes": "Extends v1.2 F2's plane-concentration finding into an identity-continuity question.",
 "_role": "federation_identity",
 "hash": "sha256:de605b233c40621795254b35aada0c2be640b377e2947188644470b2612b5c68"
}

{
 "question_id": "govq-fed-011",
 "layer": "forks, exit & portability",
 "status": "PARTIAL",
 "question": "What does exit cost: can a node export a self-contained, independently verifiable copy of its own history — receipts, filings, appeal lineage — that stays verifiable if the root delists it or disappears?",
 "why_it_matters": "A node's receipts live on the root's chain; its verify URLs are root-hosted; its governance records are root-hosted JSON. The anchor design (drand + Bitcoin binding of the chain head) and the chain-inclusion-proof object (PROVEN_INCLUDED_V2, published inclusion_law) prove per-event inclusion can be shown against external anchors — but no published export format lets a node carry its own receipt set with proofs. The depth marker stands: exit-with-history may be the strongest anti-capture right, and today history is hostage to the root's continued operation. The disincentive table concedes reputation is 'portable only with the chain's attribution intact'.",
 "candidate_answers": [
  {
   "answer": "A history-export object: the node's receipts plus per-receipt inclusion proofs against checkpoint heads plus the anchor ids binding those heads to drand/Bitcoin — keylessly verifiable forever, no operator trust.",
   "type": "protocol_object",
   "rule": "Every node may leave with a proof bundle that any stranger can verify against external anchors alone.",
   "candidate_protocol_object": "history-export/v1 (receipts + inclusion proofs + anchor references)",
   "candidate_runtime": "export route emitting the node's filtered chain slice with proofs (inclusion machinery already exists per-event)",
   "candidate_conformance": "clause: a conforming root must serve exports against a checkpoint no older than N days",
   "candidate_article": "/a/oip-exit-portability"
  },
  {
   "answer": "Exit-with-history as a non-derogable right of federation membership: the root must not be able to raise exit cost by withholding verifiability.",
   "type": "normative",
   "rule": "Verifiability-after-exit is a right, not a feature; a root that cannot serve exports fails its federation obligation.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "root-side conformance clause for export service",
   "candidate_article": "/a/oip-rights (exit clause)"
  }
 ],
 "existing_artifacts": [
  "recon2/anchor.json (what_it_proves, no_data)",
  "recon2/chain-inclusion-proof.json (PROVEN_INCLUDED_V2, inclusion_law)",
  "recon2/specification.md §5.1 row 7, §7.4"
 ],
 "evidence": [
  "inclusion_law: 'V2 starts at genesis and seals in (unixepoch(ts), id) order; an event tuple at or before the checkpoint cutoff is included'",
  "anchor _self: 'Verify it anywhere; no operator trust'",
  "spec §5.1 row 7: reputation 'portable only with the chain's attribution intact'"
 ],
 "counterevidence": [
  "no export route or per-node filtered-proof format appears in the corpus; verify URLs are root-hosted"
 ],
 "dependencies": [
  "govq-fed-015",
  "govq-fed-014",
  "govq-fed-013"
 ],
 "falsifier": "Simulate root delisting of a node's records, then attempt third-party verification of that node's receipts using only exported material and external anchors; failure marks the exit cost as unbounded.",
 "notes": "Owner brief items 11/18 (exit with state portability) in federation form.",
 "_role": "federation_identity",
 "hash": "sha256:71568a66f844c7a65181fc02b895f1db6642955caf7340bceca6476a3c330ae5"
}

{
 "question_id": "govq-fed-012",
 "layer": "forks, exit & portability",
 "status": "OPEN",
 "question": "If a fork claims the ROOT forked from IT (disputed genesis), and the fork never anchored its head before the dispute, what evidence can decide priority?",
 "why_it_matters": "fork_anchor_law proves a submitted head existed by root record time — asymmetric armor. The root's chain is continuously drand/Bitcoin-anchored; an unanchored fork cannot prove anteriority retroactively, because the anchor proves the packet 'could not be known before they existed'. The design quietly assumes disputes arrive after anchoring. A fork claiming disputed genesis has no protocol leg to stand on — which is either the occupancy moat working as designed (spec §3.2.4) or a recognition monopoly in which the root's anchored history is the only provable history.",
 "candidate_answers": [
  {
   "answer": "Anteriority evidence is the only admissible priority evidence; a fork that never anchored cannot claim genesis — and the law must say explicitly that unanchored histories are inadmissible in priority disputes.",
   "type": "normative",
   "rule": "No anchor, no anteriority claim; the burden of timestamping lies on every history-holder, root included.",
   "candidate_protocol_object": "priority-dispute record kind citing external anchor ids on both sides",
   "candidate_runtime": "verifier recomputes anchor order from published preimages",
   "candidate_conformance": "clause: any genesis/priority claim must cite external anchors, not self-hosted chain data",
   "candidate_article": ""
  },
  {
   "answer": "Self-serve anchoring for forks: the anchor recipe is public (drand + Bitcoin are external); a genesis-dispute lane lets both parties present their earliest external anchors without the root's cooperation.",
   "type": "protocol_object",
   "rule": "The root may not be the gatekeeper of anteriority evidence; the dispute lane is keyless.",
   "candidate_protocol_object": "genesis-dispute object holding both anchor chains",
   "candidate_runtime": "public submission lane mirroring kind=anchor intake",
   "candidate_conformance": "",
   "candidate_article": ""
  },
  {
   "answer": "Nakamoto reading: timestamping orders events but does not canonize them; even with priority proven, which chain is 'OIP' remains a social recognition act (govq-fed-001) — the article records this boundary.",
   "type": "article",
   "rule": "Priority is provable; canonicality is not — both statements belong in the law.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "/a/oip-fork-genesis"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.fork_anchor_law",
  "recon2/anchor.json recipe/what_it_proves",
  "recon2/specification.md §3.2.2, §3.2.4, §7.4"
 ],
 "evidence": [
  "fork_anchor_law: proves 'the submitted head existed by the root record time ... does not attest the fork's correctness, compliance, custody or truth'",
  "anchor what_it_proves: anteriority via drand round + Bitcoin block that 'could not be known before they existed'"
 ],
 "counterevidence": [
  "the root's own continuous anchoring means the asymmetry is structural, not incidental — the moat reading has textual support (spec §3.2.4 occupancy)"
 ],
 "dependencies": [
  "govq-fed-001",
  "govq-fed-013",
  "govq-fed-015"
 ],
 "falsifier": "Construct a fork with genuine earlier history and no anchors; if the protocol offers it no admissible way to prove priority, the genesis-dispute lane is confirmed absent and the asymmetry is a constitutional fact.",
 "notes": "Depth-marker question; the v1.2 F3 machine-scope finding is a prerequisite (anchor claims must be machine-enumerated before disputes are machine-manageable).",
 "_role": "federation_identity",
 "hash": "sha256:e004c28680c29fb2f955e8d21fa0179aa8efd97361372a366537e9499c51ea72"
}

{
 "question_id": "govq-fed-013",
 "layer": "forks, exit & portability",
 "status": "MISSING",
 "question": "What is the fork-choice rule between two incompatible append-only chains or registries — and absent one, is 'canonical' simply whatever the root happens to operate?",
 "why_it_matters": "Nakamoto's core move is a public fork-choice rule (greatest accumulated work) letting strangers converge without a coordinator. OIP has anchors (existence) and root-local rulings, but no rule ranking two valid chains. Two nodes can each run append-only registries with conflicting rulings about the same record, each internally valid, each externally anchored. 'Canonical' is then defined by occupancy and convention — by the root. A federation without a fork-choice rule is a single-root system with tolerated shadows (govq-fed-003).",
 "candidate_answers": [
  {
   "answer": "A fork-choice heuristic anchored in external time: among competing registries, the one whose ruling lineage is anchored earliest-and-continuously wins ties — Nakamoto-style, with anchored longevity replacing work.",
   "type": "protocol_object",
   "rule": "Canonicality is computable from public anchor order, not from operator identity.",
   "candidate_protocol_object": "fork-choice rule + anchor-continuity proofs",
   "candidate_runtime": "verifiers rank registries by recomputed anchor continuity",
   "candidate_conformance": "clause: a node must state its fork-choice rule and apply it uniformly",
   "candidate_article": ""
  },
  {
   "answer": "No fork-choice by design: plural registries, and each node publishes which registries it recognizes (Luhmann: every system draws its own boundary); interoperability, not canonicality, is the kernel's job — matching the existing core.boundary text.",
   "type": "normative",
   "rule": "Canonical is always relative to a recognizer; the kernel standardizes records, not rankings.",
   "candidate_protocol_object": "per-node recognition list (pairs with govq-fed-001 plural-recognition answer)",
   "candidate_runtime": "",
   "candidate_conformance": "clause: recognition lists must be public so each node's 'canonical' is auditable",
   "candidate_article": "/a/oip-federation-plural-recognition"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.registry_law, core.boundary",
  "articles_index.json thinker-satoshi-nakamoto"
 ],
 "evidence": [
  "registry_law: 'governs this canonical registry, not independent forks' — canonical asserted, not computed",
  "core.boundary limits the kernel to interoperability, implying no ranking law was intended"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-003",
  "govq-fed-012",
  "govq-fed-002"
 ],
 "falsifier": "Present two internally valid, externally anchored, mutually incompatible registries; if no published rule or node policy ranks them, the fork-choice layer is absent.",
 "notes": "The federation's deepest structural silence; every dispute/portability question inherits it.",
 "_role": "federation_identity",
 "hash": "sha256:1c20889e28291060f882339f15fc46a4b67562a944b008e621e53de34235c2f5"
}

{
 "question_id": "govq-fed-014",
 "layer": "forks, exit & portability",
 "status": "OPEN",
 "question": "Do sanctions, subscription standing, and appeal outcomes travel across forks — can a node escape a delisting by forking, and can a fork inherit the root's standing decisions?",
 "why_it_matters": "registry_law confines rulings to the canonical registry. If standing does not travel, escaping a sanction costs one fork (cheap), and sanctions are unenforceable federation-wide (govq-fed-008's punitive Sybil problem). If standing does travel, forks are not independent and exit is not real (owner brief: fork rights without punitive loss). Both readings are textually available; the corpus does not resolve which, and the resolution determines whether exit is a safety valve or an amnesty machine.",
 "candidate_answers": [
  {
   "answer": "Records travel, rulings do not: a fork inherits the append-only evidence (delisting rulings are records too, visible on any export) but re-adjudicates standing locally. Reputation is portable; jurisdiction is not.",
   "type": "normative",
   "rule": "Evidence is federation-global; standing is registry-local; every registry recomputes standing from the shared record graph by its own traversal.",
   "candidate_protocol_object": "standing-computation rule generalizing count_law's traversal discipline across registries",
   "candidate_runtime": "registry software imports the shared record graph and applies local rulings",
   "candidate_conformance": "clause: a registry must display which rulings it applied to imported records",
   "candidate_article": ""
  },
  {
   "answer": "Exit-from-sanction is a feature (Ostrom: the right to organize a new commons); the ledger's publicity makes flight visible, and visibility substitutes for enforceability.",
   "type": "article",
   "rule": "Forking away from a sanction is lawful; the append-only record of the sanction and the flight is the deterrent.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "/a/oip-fork-rights"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.registry_law, count_law",
  "recon2/specification.md §7.5"
 ],
 "evidence": [
  "registry_law: rulings govern 'this canonical registry, not independent forks'",
  "count_law: active-as-filed counts 'do not replace traversal of later append-only rulings' — traversal discipline exists within one registry only"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-008",
  "govq-fed-013",
  "govq-fed-011",
  "govq-pol-012"
 ],
 "falsifier": "If a delisted actor can fork and appear with clean standing while its sanction lineage is invisible to the fork's readers, the records-travel candidate is unimplemented; if no fork can carry any standing, exit-with-standing is dead.",
 "notes": "Ostrom principles engaged: graduated sanctions (requires sanctions that reach), minimal recognition of rights to organize (requires fork freedom).",
 "_role": "federation_identity",
 "hash": "sha256:e54e6e79e88487a2797dc43f8e9097c653fc08146b219341d43450db1bcbf3d8"
}

{
 "question_id": "govq-fed-015",
 "layer": "forks, exit & portability",
 "status": "PARTIAL",
 "question": "After exit or root failure, can a third party still verify an exited node's receipts — does every receipt carry, or can it obtain, an inclusion proof against an externally anchored checkpoint?",
 "why_it_matters": "Verifiability today routes through root-hosted URLs (record verify URLs, /api/chain/verify). The anchor binds checkpoint heads to drand and Bitcoin, and one inclusion proof exists in the corpus (PROVEN_INCLUDED_V2 over checkpoint 25). But inclusion proofs are not emitted per receipt by default; checkpoints seal at the root's cadence (25 checkpoints over 212,962 events); and post-checkpoint events carry no external anchor until the next seal. 'Verify without the root' exists architecturally but is not a per-receipt guarantee — and the cadence is itself an operator-controlled parameter.",
 "candidate_answers": [
  {
   "answer": "Every receipt gains a checkpoint_of_record once sealed, with inclusion proof available on demand; pre-seal receipts are honestly marked anchored:false.",
   "type": "protocol_object",
   "rule": "No receipt may be cited as externally verifiable unless its checkpoint is anchored; unanchored receipts declare themselves.",
   "candidate_protocol_object": "receipt field checkpoint_of_record + anchored flag",
   "candidate_runtime": "lazy proof emission at seal; export route for on-demand proofs (govq-fed-011)",
   "candidate_conformance": "clause: external-verifiability claims must cite anchor ids",
   "candidate_article": ""
  },
  {
   "answer": "Checkpoint cadence becomes a governed parameter (minimum frequency, anchored commitment to the cadence itself) so the operator cannot delay seals to delay verifiability.",
   "type": "runtime",
   "rule": "The seal schedule is constitutional, not operational; slippage is a visible governance event.",
   "candidate_protocol_object": "cadence-commitment record anchored each epoch",
   "candidate_runtime": "sealing job emits late-seal receipts when it misses cadence",
   "candidate_conformance": "clause: seal latency beyond the committed cadence must appear as a visible failure outcome (outcomes_include_failure)",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon2/chain-inclusion-proof.json",
  "recon2/anchor.json",
  "recon2/specification.md §1.8 (24 checkpoints / 211,652 events), killbox v1.2 (checkpoint 25 / 212,962)"
 ],
 "evidence": [
  "inclusion proof exists as an object type but is not emitted per receipt",
  "checkpoint cadence is operator-paced; no cadence commitment found in corpus"
 ],
 "counterevidence": [
  "anchor _self claims 'Verify it anywhere; no operator trust' — true for anchored material, silent for unanchored material"
 ],
 "dependencies": [
  "govq-fed-011",
  "govq-fed-012",
  "govq-inst-006"
 ],
 "falsifier": "Take the most recent unsealed receipt and attempt external verification with the root offline; failure demonstrates the verifiability window is operator-gated.",
 "notes": "Bridges to succession (govq-inst-006: sealing cadence published but enforced socially).",
 "_role": "federation_identity",
 "hash": "sha256:bb870459dada082a0a6a777ea43e22d514c819113e6c2ca576560329f0743a5c"
}

{
 "question_id": "govq-fed-016",
 "layer": "conformance & certification",
 "status": "OPEN",
 "question": "Who certifies the certifiers — what makes a verifier 'independent', and is verifier accreditation itself a constitutional act of the federation?",
 "why_it_matters": "conformance_law's ceiling is 'an independent verifier receipt', but independence is the census entity-binding problem one level up (v1.2 F2). Zero independent verifier receipts exist; the only conformance verdict on the site is the operator's own 34-clause suite grading the operator's own stack, served cached. If the federation solves node identity (govq-fed-006), verifier independence becomes soluble but politicized: accrediting verifiers is agenda control over the truth pipeline — the certifier-cartel and capture questions move up one level rather than disappearing.",
 "candidate_answers": [
  {
   "answer": "Independence is proven by reproducibility: a verifier receipt counts as independent only if its clause evidence is keylessly re-runnable by anyone and the verifier's identity is evidence-bound — no accreditation body is created.",
   "type": "protocol_object",
   "rule": "Independent = distinct evidence-bound identity + reproducible evidence; no other credential counts.",
   "candidate_protocol_object": "verifier receipt schema with identity binding + re-run URL (extends Patent 10's executable-spec pattern)",
   "candidate_runtime": "conformance evidence served as keyless re-runnable endpoints",
   "candidate_conformance": "clause: non-reproducible verifier evidence stays self-class",
   "candidate_article": ""
  },
  {
   "answer": "Verifier accreditation as a constitutional act: governing nodes ratify a verifier registry through the same admission machinery as node recognition (govq-fed-001).",
   "type": "normative",
   "rule": "Whoever may certify is decided by the federation's admission process, under the same entity-binding evidence rules.",
   "candidate_protocol_object": "verifier-registry records with admission rulings",
   "candidate_runtime": "registry intake + admission rulings",
   "candidate_conformance": "clause: accredited-verifier list must be public and append-only",
   "candidate_article": "/a/oip-verifier-accreditation"
  },
  {
   "answer": "No accreditation ever: conformance claims are per-verifier statements and consumers weigh them — a marketplace of verification where 'conformant' always means 'conformant-per-V' (Luhmann: trust is purchased per-observer).",
   "type": "article",
   "rule": "The protocol refuses to canonicalize certifiers; every verdict names its verifier and stands alone.",
   "candidate_protocol_object": "verdict records carry verifier identity + evidence links",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "/a/oip-conformance-marketplace"
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json participation.conformance_law",
  "killbox-specification-v1.2.md Part E F4, masthead self-graded flag",
  "recon2/specification.md §7.3, Patent 10"
 ],
 "evidence": [
  "conformance_law: 'Self-attested evidence remains self-attested until an independent verifier receipt is linked'",
  "live state: zero kind=conformance records, zero independent verifier receipts; 34/34 self-graded, cached"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-006",
  "govq-fed-009",
  "govq-fed-017",
  "govq-fed-001",
  "govq-inst-010"
 ],
 "falsifier": "If 'independent' cannot be given a machine-checkable or constitutionally admitted meaning, every future conformance badge inherits the census's Sybil exposure one level up.",
 "notes": "Depth-marker question: the entity-binding problem recurs at every layer that invokes independence.",
 "_role": "federation_identity",
 "hash": "sha256:cd62535559c9ba2e3a2f5717c4dba99fb29b4bd3c6c27a1ff6818bd557ee59d3"
}

{
 "question_id": "govq-fed-017",
 "layer": "conformance & certification",
 "status": "PARTIAL",
 "question": "Who runs the conformance suite against a subscriber's stack — self-run published artifact, verifier-run receipt, or root-run certification — and whose run counts?",
 "why_it_matters": "Facet profiles publish testable predicates and the suite is keylessly re-runnable (Patent 10), but the only live verdict grades the root. kind=conformance records require evidence_links, yet the schema does not require the evidence to be produced by anyone other than the claimant, nor state whether facet predicates are remotely testable against a subscriber's stack. C26–C28 demonstrate remote probing of exactly one peer (buttercup) for the federated-messages facet only. The registry cannot currently distinguish a conformance claim from a conformance demonstration.",
 "candidate_answers": [
  {
   "answer": "Remote conformance runner: the envelope machinery (queries answered as data) lets any node probe another's facet predicates; the signed probe transcript is the evidence_link.",
   "type": "runtime",
   "rule": "Conformance evidence is a transcript of remote probes, verifiable by replay, not a screenshot of a self-run.",
   "candidate_protocol_object": "probe-transcript object bound into conformance records",
   "candidate_runtime": "conformance probe envelopes on the federation inbox",
   "candidate_conformance": "clause: facet profiles must declare which predicates are remotely testable",
   "candidate_article": ""
  },
  {
   "answer": "Conformance records carry runner class and scope as machine fields: runner: self|verifier:<id>|root, scope: facet|kernel, plus the stack fingerprint graded (extends v1.2 F4's three-field proposal).",
   "type": "protocol_object",
   "rule": "Every conformance claim names who ran it, on what stack, at what scope — machine-readable.",
   "candidate_protocol_object": "runner/scope/stack-fingerprint fields on kind=conformance",
   "candidate_runtime": "intake validation (422 on missing class)",
   "candidate_conformance": "clause: counters must not mix runner classes",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md §7.3",
  "recon2/conformance2.md C26-C28",
  "killbox-specification-v1.2.md Part E F4"
 ],
 "evidence": [
  "spec §7.3: 'zero kind=conformance records and zero independent verifier receipts'",
  "C26-C28 show the only remote-tested facet is federated-messages, against the operator's own second node"
 ],
 "counterevidence": [
  "facet conformance_profile predicates are published per facet — the testable surface exists as text"
 ],
 "dependencies": [
  "govq-fed-016",
  "govq-fed-018",
  "govq-fed-019"
 ],
 "falsifier": "Attempt to grade a third party's stack against a facet profile keylessly; if the predicates are not remotely executable, the runner question collapses to self-run for everything except federated-messages.",
 "notes": "Builds on v1.2 F4 without repeating it: F4 asked for scope fields; this asks whose run produces the evidence.",
 "_role": "federation_identity",
 "hash": "sha256:552501a6b670293673af67de89ceda52da925a84749dde298fcef124ce638212"
}

{
 "question_id": "govq-fed-018",
 "layer": "conformance & certification",
 "status": "MISSING",
 "question": "When two verifiers publish conflicting conformance results for the same node and facet, what resolves the conflict?",
 "why_it_matters": "Compose govq-fed-002 (no inter-node dispute mechanism) with govq-fed-016 (no verifier accreditation): conformance is per-verifier opinion with no aggregation rule. The federation's stated promise — implementations 'strengthen it by producing mutually verifiable records' — does not extend to mutually verifiable verdicts. Verifier shopping (a node collecting a friendly verifier) is the certification analog of jurisdiction shopping, and nothing in the corpus even requires conflicting verdicts to acknowledge one another.",
 "candidate_answers": [
  {
   "answer": "Conflicting verdicts must link each other: a conformance-conflict record kind, reusing the article constitution's challenges[]/challenged_by[] pattern, so disagreement is traversable.",
   "type": "protocol_object",
   "rule": "A verifier publishing on a target must check for existing verdicts on the same (node, facet, kernel_hash) and link disagreements before publishing.",
   "candidate_protocol_object": "conformance-conflict record kind + challenge links",
   "candidate_runtime": "registry indexes verdicts by target tuple",
   "candidate_conformance": "clause: unlinked conflicting verdicts are downgraded to self-class",
   "candidate_article": ""
  },
  {
   "answer": "Conflicts are preserved, never resolved: both verdicts stay visible and consumers decide — append-only honesty applied to certification.",
   "type": "normative",
   "rule": "The federation records disagreement faithfully and refuses to compute a winner.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "/a/oip-conformance-conflicts"
  }
 ],
 "existing_artifacts": [
  "corpus/bundle_constitution.md (challenges[]/challenged_by[] pattern)",
  "recon2/governance2.json participation.conformance_law"
 ],
 "evidence": [
  "conformance_law is silent on inter-verifier conflict",
  "the challenge-link pattern exists for articles but not for conformance records"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-002",
  "govq-fed-016",
  "govq-fed-017"
 ],
 "falsifier": "Publish two contradictory verdicts for one target; if neither the registry nor the law requires linkage or resolution, the conflict layer is absent.",
 "notes": "Ostrom conflict-resolution principle at the certification layer.",
 "_role": "federation_identity",
 "hash": "sha256:303ce5fd9867e818bbe1c91154728d58c6bca45c27ea9c2b2e0974b919a6dab4"
}

{
 "question_id": "govq-fed-019",
 "layer": "conformance & certification",
 "status": "PARTIAL",
 "question": "Does a conformance verdict expire — against which kernel and contract versions was it issued, and does semantic drift in the words 'conformance', 'independent', or 'repair' invalidate old verdicts?",
 "why_it_matters": "C23 pins row tokens to contract fingerprints and fails closed on drift at execution time — but published conformance records carry no version pinning, expiry, or re-test cadence. Owner brief item 14 (semantic capture): a future operator keeps the names and changes the meanings. Certification is where semantic capture pays: yesterday's verdict under yesterday's meaning, cited as today's badge. The registry already publishes a facets hash (9cbf3b3e…) and a kernel hash — the pinning substrate exists but is not wired into verdicts.",
 "candidate_answers": [
  {
   "answer": "Verdicts carry kernel_hash, facet hash, contract fingerprints, and valid_from/valid_until; a verdict auto-transitions to stale when a referenced hash changes — drift detection at the registry layer mirroring C23 at the door.",
   "type": "protocol_object",
   "rule": "A conformance verdict is a statement about a pinned stack; unpinning invalidates it mechanically.",
   "candidate_protocol_object": "version-pinned conformance record with staleness transition",
   "candidate_runtime": "registry recomputes staleness on kernel or facet amendment",
   "candidate_conformance": "clause: verdicts without hash pins are self-class",
   "candidate_article": ""
  },
  {
   "answer": "Protected definitions: constitutional terms ('conformance', 'independent', 'repair', 'material success') get semantic versions, and amending a definition is a constitutional act routed through admission machinery (govq-fed-001), not a deploy.",
   "type": "normative",
   "rule": "Changing what a constitutional word means is a constitutional amendment, with lineage visible.",
   "candidate_protocol_object": "definition records with semantic versions and amendment lineage",
   "candidate_runtime": "",
   "candidate_conformance": "clause: contract text must cite definition versions",
   "candidate_article": "/a/oip-semantic-versioning"
  }
 ],
 "existing_artifacts": [
  "recon2/conformance2.md C23",
  "recon2/governance2.json facets.hash, core.hash",
  "owner brief item 14"
 ],
 "evidence": [
  "C23: contract drift fails closed at execution time — the door pins, the registry does not",
  "facets.hash exists but no verdict references it"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-017",
  "govq-fed-001",
  "govq-pol-021"
 ],
 "falsifier": "Amend a facet profile after a verdict issues; if the verdict's meaning silently changes with the profile, semantic capture is demonstrated.",
 "notes": "The registry already holds the hashes needed; the gap is wiring, not substrate.",
 "_role": "federation_identity",
 "hash": "sha256:2d8d3b2096a03d55f9384be54b5c434fb07a239d6531747e9a1ec9a5c6a1e52b"
}

{
 "question_id": "govq-fed-020",
 "layer": "authority & delegation across nodes",
 "status": "IMPLIED",
 "question": "Can a capability minted by node A authorize execution on node B's ledger at all — is there cross-root authority, or is every node its own root of authority (a confederation of silos)?",
 "why_it_matters": "As deployed, authority is minted per node: the root's capabilities are honored at the root's door when carried by a verified remote agent, and C27's audience binding binds the capability to one agent while the capability itself remains root-minted. The kernel boundary ('does not require ... surrender of credentials and infrastructure') implies each node stays its own authority root. Nothing in the corpus lets node A mint authority that node B enforces on B's objects. If that is the design, the federation is authenticated message-passing between sovereign silos — the depth-marker 'confederation of silos' question — and 'shared governance' is coordination, not shared authority.",
 "candidate_answers": [
  {
   "answer": "Silo-by-design: cross-node authority would widen authority across trust domains, violating bounded_authority's spirit; the federated-messages facet's obligations (verify identity separately from authority) encode the separation. Delegation is always re-minted at the receiving root: B mints a capability whose audience is A's agent — exactly the deployed pattern, generalized.",
   "type": "normative",
   "rule": "Authority never crosses a root boundary; only requests do. Each root re-mints under its own gates.",
   "candidate_protocol_object": "existing audience-bound capability pattern, documented as the federation delegation law",
   "candidate_runtime": "existing door gates re-checked per call",
   "candidate_conformance": "C27 already tests fail-closed audience binding",
   "candidate_article": "/a/oip-federation-authority (states the silo law)"
  },
  {
   "answer": "Mutual-recognition object: B publishes a revocable, ledgered list of A-minted capability fingerprints it will honor — foreign-key acceptance as an explicit constitutional act rather than an implicit config.",
   "type": "protocol_object",
   "rule": "Honoring foreign authority is a published, revocable, per-root act with its own receipts.",
   "candidate_protocol_object": "foreign-capability-acceptance records on B's ledger",
   "candidate_runtime": "B's door checks A's capability against the acceptance list plus freshness (govq-fed-022)",
   "candidate_conformance": "clause: honoring a foreign capability without a ledgered acceptance fails closed",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json core.boundary, facets federated-messages",
  "recon2/conformance2.md C27",
  "recon2/specification.md §1.5, Patent 6"
 ],
 "evidence": [
  "C27: audience-bound capability fails closed (403 audience_bound) outside a verified signed invoke",
  "core.boundary: no surrender of credentials and infrastructure",
  "fed_manifest note: 'every gate is re-checked here'"
 ],
 "counterevidence": [
  "the deployed demonstration is root-minted capability exercised by a remote agent on the root's own stack — not authority crossing roots"
 ],
 "dependencies": [
  "govq-fed-021",
  "govq-fed-022",
  "govq-fed-025"
 ],
 "falsifier": "Attempt to present node A's minted capability at node B's door for an object on B's ledger; if no schema even expresses the attempt, cross-root authority is unrepresented.",
 "notes": "Depth-marker question; the implied answer (silos) is defensible but currently implied by boundary text, not stated as law.",
 "_role": "federation_identity",
 "hash": "sha256:50be8c42dafd57b4c179e64a9bcf07ec9c0cdb6831e12737017c7bdfd65af708"
}

{
 "question_id": "govq-fed-021",
 "layer": "authority & delegation across nodes",
 "status": "MISSING",
 "question": "Can authority delegated across a node boundary be further delegated — is there a cross-node attenuation chain, and who can traverse the delegation graph?",
 "why_it_matters": "Within a node, delegated authority only shrinks (I1.1) and ancestor chains are re-checked per call (I1.2). Across nodes, the corpus shows single-hop audience binding only. If buttercup re-delegates its audience-bound capability to a third node, nothing specifies whether the chain holds, who re-checks the ancestors, or where the attenuation is ledgered. Macaroons' third-party caveats are conceded as the closest capability field (Patent 6), and the differentiator claim does not cover transitive federation. A two-node federation hides this gap; an N-node federation cannot.",
 "candidate_answers": [
  {
   "answer": "Chained-envelope delegation: each hop attenuates and appends its own signed caveat row; the terminal node verifies the whole chain; every hop's verdict lands on both adjacent ledgers (mirrored-ledger pattern generalized to path-mirroring).",
   "type": "protocol_object",
   "rule": "Transitive delegation is lawful only as a fully witnessed, monotonically shrinking chain.",
   "candidate_protocol_object": "delegation-chain envelope with per-hop caveat rows",
   "candidate_runtime": "terminal door verifies chain integrity and per-hop attenuation",
   "candidate_conformance": "clause: any widening hop invalidates the chain fail-closed",
   "candidate_article": ""
  },
  {
   "answer": "Forbid transitivity: delegation terminates at the first remote hop, matching audience_bound inertness; any wider pattern requires the foreign-acceptance object (govq-fed-020) at each root.",
   "type": "normative",
   "rule": "Authority crosses exactly one root boundary per mint; composition happens by re-minting, never by forwarding.",
   "candidate_protocol_object": "",
   "candidate_runtime": "door rejects envelopes presenting re-delegated foreign capabilities",
   "candidate_conformance": "clause: re-delegation attempts are logged as denials",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md I1.1/I1.2 (node-local), Patent 6 (macaroons concession)",
  "recon2/conformance2.md C27"
 ],
 "evidence": [
  "I1.1/I1.2 are specified and tested only within the root's own capability plane",
  "no cross-node attenuation text exists anywhere in the corpus (grep-verified)"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-020",
  "govq-fed-022"
 ],
 "falsifier": "Attempt a second-hop delegation from the peer node to a third node; if the attempt is inexpressible in the envelope schema, transitive delegation is unrepresented.",
 "notes": "The two-node live proof (spec §7.8) structurally cannot surface this question; it only appears at N≥3.",
 "_role": "federation_identity",
 "hash": "sha256:60a697017250de50275277fb7adfd785b73724da5af1e34b9725a5832e046007"
}

{
 "question_id": "govq-fed-022",
 "layer": "authority & delegation across nodes",
 "status": "MISSING",
 "question": "How does a revocation propagate across nodes — what freshness guarantee does a remote node have that the capability it honors has not been revoked at its mint?",
 "why_it_matters": "The door re-checks ancestor chains per call, but only against the root's own state; the fed_ledger's denied:revoked rows fire because checker and minter are the same node. In any topology where node B honors authority minted by A (govq-fed-020 alternatives), or caches verification, revocation has a propagation window. No revocation-list anchor, max-age, or freshness obligation appears anywhere in the corpus, and the store-and-forward carriers conceded in Patent 6 (SMTP binding) would make the window unbounded.",
 "candidate_answers": [
  {
   "answer": "Freshness rule: a remote node must either call home (online re-check at the mint) or demand a capability whose TTL is within the freshness bound; the existing rejected:expired_envelope denial class generalizes.",
   "type": "runtime",
   "rule": "No freshness proof, no execution; staleness fails closed.",
   "candidate_protocol_object": "freshness bound field on cross-node capabilities",
   "candidate_runtime": "door enforces max-age on foreign-presented authority",
   "candidate_conformance": "clause: honoring stale authority is a visible failure outcome",
   "candidate_article": ""
  },
  {
   "answer": "Revocation-head anchoring: each node anchors its current revocation head on the checkpoint cadence; verifiers and peers check the anchored head before honoring.",
   "type": "protocol_object",
   "rule": "Revocation state is externally anchored so peers can verify it without trusting the mint's live answer.",
   "candidate_protocol_object": "revocation head included in anchor packets (census-anchor pattern, v1.2 F7, generalized)",
   "candidate_runtime": "anchor job extends the preimage with the revocation head",
   "candidate_conformance": "clause: peers must refuse capabilities revoked at the latest anchored head",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon/fed_ledger.json denied:revoked rows",
  "recon2/specification.md I1.2, Patent 6 (SMTP variation)"
 ],
 "evidence": [
  "denied:revoked fires only where checker = minter",
  "corpus-wide absence of cross-node revocation or freshness text (grep-verified)"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-020",
  "govq-fed-021",
  "govq-fed-007"
 ],
 "falsifier": "Revoke a capability at its mint while a peer holds a cached copy; if the peer can still execute within an unbounded window, the propagation gap is demonstrated.",
 "notes": "Invisible in the current two-node/one-operator topology; load-bearing for any real confederation.",
 "_role": "federation_identity",
 "hash": "sha256:a5b936f74c41a1b3f2344b462902a453b10172bb8b6d1c85d9bd2738b37d43c6"
}

{
 "question_id": "govq-fed-023",
 "layer": "authority & delegation across nodes",
 "status": "OPEN",
 "question": "Is there a federation-wide or multi-agent audience — or does bilateral-only audience binding force N-squared capability mintings as the federation grows?",
 "why_it_matters": "C27 binds a capability to one named remote agent. Scaling to N nodes with bilateral audience-bound capabilities requires each authority relationship to be minted pairwise: O(N^2) mintings, each with its own revocation surface (govq-fed-022) and key-succession exposure (govq-fed-007). Group audiences (any verified member of a peer set or facet class) are unrepresented. This is Coase at the protocol layer: bilateral contracting costs scale quadratically, and the audience grammar decides whether the federation organizes as a market or a hierarchy.",
 "candidate_answers": [
  {
   "answer": "Group-audience capabilities: audience equals a peer-set hash or facet-membership claim; verification is a membership proof against the ledgered peer set (govq-fed-004).",
   "type": "protocol_object",
   "rule": "An audience may be a set with a public membership proof; member additions and removals ride the peer-set ledger.",
   "candidate_protocol_object": "group-audience capability + membership proof objects",
   "candidate_runtime": "door verifies membership proof at invoke time",
   "candidate_conformance": "clause: group audiences must name the ledger that defines membership",
   "candidate_article": ""
  },
  {
   "answer": "Bilateral-only by design: audience specificity is the security property (fail-closed inertness); N-squared is the honest price of explicit authority, and shared governance energy exists partly to pay it.",
   "type": "normative",
   "rule": "No implicit audiences; every authority relationship is enumerated.",
   "candidate_protocol_object": "",
   "candidate_runtime": "capability templates minted per peer from one contract, amortizing operator cost without widening audience",
   "candidate_conformance": "",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon2/conformance2.md C27",
  "recon2/fed_manifest.json (single agent)",
  "recon2/specification.md §1.5"
 ],
 "evidence": [
  "C27: 'A capability can be bound to one specific remote agent'",
  "manifest publishes exactly one agent; no group or set audience semantics anywhere"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-004",
  "govq-fed-020",
  "govq-fed-022"
 ],
 "falsifier": "If adding the tenth node requires forty-five new mintings with no group primitive, the scaling question is answered by default in favor of hierarchy.",
 "notes": "Coase: the audience grammar is the federation's theory of the firm.",
 "_role": "federation_identity",
 "hash": "sha256:ebe9a02ce49fa5cc916b12fba681b23d9922c9d1d63c2c791075600f4d63bce6"
}

{
 "question_id": "govq-fed-024",
 "layer": "authority & delegation across nodes",
 "status": "CONTRADICTED",
 "question": "Is cross-node confidentiality implemented or demonstrated-only — and if bodies are sealed to one recipient, what does federation accountability consist of: hash-only public audit?",
 "why_it_matters": "The live manifest states 'payload E2EE and the SMTP binding are not implemented' while C28 passes a sealed-body round trip (ECDH-P256-A256GCM, opened_ok, wrong_key_denied) — a tension disclosed as spec FINDING 3 (§1.9) and Patent 6 claim 4's honesty flag, still unresolved. Beneath the contradiction sits an unchosen governance boundary: mirrored ledgers publish msg_id and body_sha256 only. If bodies go sealed end-to-end, public audit reduces to hash-existence; if bodies are plaintext to the receiving root, cross-node governance traffic is root-readable. The accountability/confidentiality split across nodes has never been constitutionally chosen.",
 "candidate_answers": [
  {
   "answer": "Verdicts public, bodies sealable: the ledger proves the exchange and its gate outcome; content confidentiality is a facet, not a kernel obligation — consistent with credentials_are_not_evidence.",
   "type": "normative",
   "rule": "Accountability attaches to the gate outcome (verdict, hashes, denials); content privacy is lawful and orthogonal.",
   "candidate_protocol_object": "sealed-body envelope as a facet-scoped capability",
   "candidate_runtime": "inbox accepts sealed invokes; ledger rows stay hash-only",
   "candidate_conformance": "clause: a node claiming sealed transport must demonstrate a routed sealed payload, not an envelope-format vector (the existing honesty-flag fix)",
   "candidate_article": ""
  },
  {
   "answer": "Selective-disclosure receipt: the body hash plus a reveal-on-challenge lane — a node can be compelled by governance process to reveal a body's preimage to verifiers.",
   "type": "protocol_object",
   "rule": "Sealing is rebuttable under process; the hash on the ledger binds the revealer.",
   "candidate_protocol_object": "reveal-on-challenge record kind tied to objection lane",
   "candidate_runtime": "challenge route verifies revealed preimage against ledgered body_sha256",
   "candidate_conformance": "clause: refusal to reveal under valid challenge is a visible outcome",
   "candidate_article": "/a/oip-federation-confidentiality"
  }
 ],
 "existing_artifacts": [
  "recon2/fed_manifest.json confidentiality line",
  "recon2/conformance2.md C28",
  "recon2/specification.md §1.9 FINDING 3, Patent 6 claim 4 honesty flag"
 ],
 "evidence": [
  "manifest: 'TLS transport only; payload E2EE and the SMTP binding are not implemented.'",
  "C28 PASS: opened_ok:true, wrong_key_denied:true — envelope-format vector only",
  "spec flags the tension as unresolved"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-002",
  "govq-fed-020"
 ],
 "falsifier": "Until a sealed payload is receipted over a routed transport, the contradiction stands: the ledgered capability and the manifest's disclaimer cannot both describe the deployed system.",
 "notes": "Extends spec FINDING 3 from a claim-hygiene issue to a constitutional confidentiality question.",
 "_role": "federation_identity",
 "hash": "sha256:c8c1061681612f248c0519c8129ad427fdd3f0e14a6d0a0012712e0c8bd0b1b2"
}

{
 "question_id": "govq-fed-025",
 "layer": "authority & delegation across nodes",
 "status": "IMPLIED",
 "question": "Who are the 'governing nodes' that allocate pledged governance energy — how is that status acquired, and does allocation authority cross node boundaries?",
 "why_it_matters": "post_subscription.sequence step 4 states: 'Governing nodes allocate that capacity to shared conformance, audit, repair and specification work.' No record kind, census field, or facet defines governing-node status. The invariant caps unilateral authority growth but is silent about the governing-node class itself. The only remediation on record names 'sub-governing nodes' and 'pledge-weighted governance with receipted pledges and fidelity slashing-by-append' as design inputs from an unratified model-recommendation filing (gov_7d480271c03a45138e62). At census nu=1 the governing node is the root by default — the allocation constitution currently exists only as an unratified proposal, and whether allocation authority can reach across roots (govq-fed-020) is undecided.",
 "candidate_answers": [
  {
   "answer": "Governing-node status is acquired through the constitutional tier: ratified pledge + verified conformance + admission (govq-fed-001); allocation acts are ordinary capabilities — bounded, attenuated, receipted — never registry standing.",
   "type": "normative",
   "rule": "Allocation authority is delegated authority under the same capability laws; standing in the registry grants none of it (no_authority_grant extended to the energy loop).",
   "candidate_protocol_object": "pledge record + allocation capability + allocation receipts (sequence steps 3-5 as objects)",
   "candidate_runtime": "metered allocation routes under capability gates",
   "candidate_conformance": "clause: every allocation must produce a receipt traversable to a pledge",
   "candidate_article": "/a/oip-governance-energy"
  },
  {
   "answer": "Allocation authority cannot cross node boundaries: governing nodes coordinate pledges, but each node's capability plane executes locally (the silo answer of govq-fed-020 applied to the energy loop).",
   "type": "protocol_object",
   "rule": "Pledges are federation-visible; execution authority stays root-local; cross-node allocation happens by re-minting at each root.",
   "candidate_protocol_object": "pledge visibility records + per-root allocation capabilities",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": ""
  },
  {
   "answer": "Fidelity slashing-by-append: pledge defaults are ruled on and appended (never erased), making governing-node discipline an append-only lineage — the unratified design input promoted to candidate law.",
   "type": "conformance",
   "rule": "Slashing is a ruling family with receipts; a governing node's defaults remain publicly traversable forever.",
   "candidate_protocol_object": "slashing ruling kind linked to pledge records",
   "candidate_runtime": "ruling intake extended with slashing decisions",
   "candidate_conformance": "clause: slashing rulings must cite the pledge and the default evidence",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "recon2/governance2.json post_subscription.sequence/invariant/implementation_status",
  "recon2/specification.md §5.1 rows 7/10",
  "recon2/gov_record.json (gov_7d480271c03a45138e62 design input)"
 ],
 "evidence": [
  "sequence step 4: 'Governing nodes allocate that capacity' — undefined class",
  "implementation_status: 'Metered pledge allocation and settlement are a proposed next contract, not a live payment or resource market'",
  "spec §5.1 row 10: pledge-weighted governance and sub-governing nodes exist only as design inputs"
 ],
 "counterevidence": [],
 "dependencies": [
  "govq-fed-001",
  "govq-fed-005",
  "govq-fed-009",
  "govq-fed-020"
 ],
 "falsifier": "If no record kind can express 'X is a governing node', the class exists only in prose; the first real allocation event will either mint the constitution ad hoc or expose the gap.",
 "notes": "The energy loop is where federation recognition, identity binding, and cross-root authority converge; left IMPLIED because every element is specified-but-undeployed.",
 "_role": "federation_identity",
 "hash": "sha256:a95f8391e82f8f6158f1a86c9ab291db32879be4d48bcb886dcd0b839e8c8b67"
}

{
 "question_id": "govq-inst-001",
 "layer": "emergency_powers",
 "status": "MISSING",
 "question": "Who may declare an emergency in OIP, through what record kind, with what evidence threshold, and who verifies the declaration?",
 "why_it_matters": "The governance schema enumerates kinds {subscribe, inquire, propose, feature, conformance, anchor, appeal, ruling} — no emergency, freeze, or incident record kind exists. The killbox wargame concedes the system plays nine attacker classes and never the cheapest one (buy/coerce the operator), and recommends an emergency/succession move that does not exist. A system with live auto-revocation machinery (L14) but no constitutional declaration procedure has emergency capability without emergency legitimacy: power can be exercised, but no record type exists to make its exercise accountable. Schmitt's point (external to corpus) applies: the sovereign is whoever decides the exception, and here the decider is structurally invisible.",
 "candidate_answers": [
  {
   "answer": "Owner-declared emergency with mandatory public record",
   "type": "normative",
   "rule": "Only the root may declare, but every declaration is a receipted, time-boxed, scope-enumerated public record; undeclared emergency action is a constitutional violation by definition.",
   "candidate_protocol_object": "gov_emergency_record {kind:'emergency', scope:[facets], ttl, evidence_links, parent_id}",
   "candidate_runtime": "emergency kind on /api/governance intake + dispatch gate flag visible keylessly",
   "candidate_conformance": "clause: every CAP_REVOKE/DEPLOY_LEASE acquire during a declared window links the emergency record id",
   "candidate_article": "oip-emergency-powers-constitution"
  },
  {
   "answer": "No emergency powers at all — freeze is impossible by design",
   "type": "normative",
   "rule": "The protocol takes the cypherpunk branch: no actor may freeze; compromise is answered by individual key revocation and fork, never by system-level suspension.",
   "candidate_protocol_object": "none (constitutional refusal recorded as a boundary article)",
   "candidate_runtime": "none — absence is the mechanism",
   "candidate_conformance": "conformance must prove NO freeze op exists in the capability map",
   "candidate_article": "oip-boundary-no-emergency-power"
  },
  {
   "answer": "Threshold-declared emergency by verifier quorum",
   "type": "protocol_object",
   "rule": "k-of-n independent verifier receipts declaring compromise activate a published, limited emergency profile.",
   "candidate_protocol_object": "emergency_decl {quorum:[verifier_receipts], k, n, scope, sunset}",
   "candidate_runtime": "dispatch emergency profile activated only by quorum envelope",
   "candidate_conformance": "F-series style negative tests: k-1 receipts must fail closed",
   "candidate_article": "oip-emergency-quorum-profile"
  }
 ],
 "existing_artifacts": [
  "/api/governance participation.kinds",
  "OBJECTION_LOG",
  "CAP_REVOKE",
  "L14 auto-revocation (killbox v1.1 §10.5)"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/governance2.json — participation.kinds has no emergency/freeze kind; decision field is owner-only",
  "/mnt/agents/output/recon2/specification.md:1404-1407 — F9: no transfer/succession/ratification/deadman record kind; recommended move unplayed",
  "owner brief (upload) item 3: 'You have revocation, but not a full emergency constitution'"
 ],
 "counterevidence": [
  "L14 shows incident-response machinery exists (auto-revoke leaked capability, generic 404, fingerprint-only ledgering) — capability without constitution",
  "revocation cascade conformance C17-C19 PASS (commit_inspection.md §4) proves the runtime can execute emergency-class acts today"
 ],
 "dependencies": [
  "govq-inst-002",
  "govq-inst-004",
  "govq-inst-020"
 ],
 "falsifier": "If any party can point to a deployed emergency record kind with scope, sunset, and receipt linkage, this question is resolved IMPLEMENTED; if a constitutional boundary article explicitly refuses emergency power, it is resolved NON_AUTOMATABLE.",
 "notes": "Schmitt (external, not in corpus) on the exception; Foucault (thinker-map lines 354-357: GAP, not referenced — power as productive; the undeclared emergency produces governance without records). Third person; conflicts are findings.",
 "_role": "institutional_designer",
 "hash": "sha256:753f0f403797ca619d1a78fb726b056c756ea699a49e691c01fd7e45604d2a96"
}

{
 "question_id": "govq-inst-002",
 "layer": "emergency_powers",
 "status": "PARTIAL",
 "question": "What can the emergency actor NOT do — and is that limit enforced by the runtime or merely promissory? The owner can already rotate and revoke every key; what is the enumerated set of acts even the root cannot take during a crisis?",
 "why_it_matters": "bounded_authority is axiom one — but it binds delegated capabilities, not the root minter who mints them. The killbox AC4 finding: the author is root minter, conformance author, objection judge, and ruling authority simultaneously, and 'nothing deployed enforces a transition to multi-party rulings — no threshold, timetable, or ratification record kind exists in the schema.' A non-derogation set (e.g., 'no re-anchoring anchored history as if unanchored', 'no minting in another node's name') is the difference between a constitutional emergency power and a permanent dictatorship switch.",
 "candidate_answers": [
  {
   "answer": "Enumerate a non-derogable floor binding even the root",
   "type": "normative",
   "rule": "Even in emergency: anchored heads are not re-anchored as if unanchored; prior receipts are not delisted without linked ruling; no capability is minted carrying another node's identity.",
   "candidate_protocol_object": "non_derogation_set in core; emergency record must reference it",
   "candidate_runtime": "dispatch refuses owner-tagged ops that violate the floor during declared windows",
   "candidate_conformance": "clause: synthetic violation attempt returns typed refusal + receipt",
   "candidate_article": "oip-emergency-non-derogation-floor"
  },
  {
   "answer": "Root is bound by the same attenuation algebra as delegates",
   "type": "protocol_object",
   "rule": "Owner authority is itself a capability with caveats, so emergency powers inherit expiring, attenuable, revocable structure.",
   "candidate_protocol_object": "root_capability with caveats {no_identity_forge, no_anchor_backdate}",
   "candidate_runtime": "mint routes issue the owner an attenuable root, not raw keys",
   "candidate_conformance": "C17-C19 extended: root caveat violation fails closed",
   "candidate_article": "oip-root-as-capability"
  }
 ],
 "existing_artifacts": [
  "bounded_authority axiom (governance2.json core.axioms)",
  "CAP_REVOKE, attenuate route (contract_deploy.md verbs)"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/governance2.json — axiom 'Every action runs under explicit authority that cannot widen itself'",
  "/mnt/agents/output/recon2/specification.md:534 — 'Root capability keys are minted through owner-only routes (?mint_share=1, ?tap_go=1, tagged [owner]); every delegated key descends from an owner mint. The authority graph's root is one person'",
  "/mnt/agents/output/recon2/specification.md:552-555 — AC1/AC4: no independent verifier; governance capture structurally present, legally null at census zero"
 ],
 "counterevidence": [
  "append_and_repair axiom does de-facto limit even the root: history cannot be rewritten without detection — one enforced limit exists",
  "L14 auto-revocation fires against the system's own leaked credentials — machinery acts against owner-side material without owner discretion"
 ],
 "dependencies": [
  "govq-inst-001",
  "govq-inst-003",
  "govq-inst-009"
 ],
 "falsifier": "Produce a runtime test in which an owner-tagged invocation attempting a floor violation (e.g., deleting a receipt, backdating an anchor) returns a typed refusal; absence of any such refusal path confirms 'promissory only.'",
 "notes": "Pareto (thinker-map 534-537; corpus has only his optimality, C15 — the elite-circulation Pareto is absent): the limit question is whether circulating elites (future operators) inherit the same floor. Coase (501-505): the structure of rights determines outcomes — an unenumerated residual right defaults to the root.",
 "_role": "institutional_designer",
 "hash": "sha256:a162f55c3ffb5a86a2286a8b971325dfa713ca3488a4dbc29c613db51b21436a"
}

{
 "question_id": "govq-inst-003",
 "layer": "emergency_powers",
 "status": "CONTRADICTED",
 "question": "Does the append-only axiom admit emergency exceptions? The article constitution already scrubs leaked secrets to [REDACTED:secret-leak] with a scrub_events tombstone, while core axiom append_and_repair says history is 'appended, never rewritten.' Which law governs when they collide, and who may invoke the exception?",
 "why_it_matters": "This is the live precedent for all emergency power: one carve-out from immutability already exists (secret scrubbing with tombstone), is unilateral, and has no published actuation record (L14 is DOCUMENTED, not keylessly provable — drift D6). If scrubbing is legitimate, the protocol has conceded that some emergencies override append-only; if it is not, a production mechanism violates a core axiom. Either branch must be represented, not smoothed over. The same collision reappears for court-ordered deletion (govq-inst-022).",
 "candidate_answers": [
  {
   "answer": "Harmonize: tombstone-as-append doctrine",
   "type": "normative",
   "rule": "A tombstone is an append, not a rewrite: the scrub event is itself ledgered with hash lineage, so append_and_repair is formally preserved; the doctrine must say exactly which classes qualify (secret-leak only) and who may invoke.",
   "candidate_protocol_object": "scrub_event {target_hash, class:'secret-leak', actor_fingerprint, tombstone_hash}",
   "candidate_runtime": "ingress auto-revoke + scrub lane with public tombstone receipts",
   "candidate_conformance": "clause: scrubbed content leaves tombstone; non-qualifying scrub attempt is refused",
   "candidate_article": "oip-append-only-exceptions"
  },
  {
   "answer": "Keep the contradiction visible",
   "type": "article",
   "rule": "Publish the collision as a standing objection rather than resolving it by definition; let the ledger show both laws and the unresolved tension.",
   "candidate_protocol_object": "objection record linking axiom text and constitution claim rule",
   "candidate_runtime": "none",
   "candidate_conformance": "dedup gate must not suppress the collision objection",
   "candidate_article": "oip-standing-contradiction-append-vs-scrub"
  }
 ],
 "existing_artifacts": [
  "append_and_repair axiom",
  "bundle_constitution.md claim rules (scrub_events)",
  "L14 (killbox v1.1 §10.5)"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/governance2.json — axiom: 'History is appended, never rewritten; corrections link the failed record to its replay or repair'",
  "/mnt/agents/output/corpus/bundle_constitution.md:55 — 'Leaked secrets are scrubbed to [REDACTED:secret-leak] with scrub_events tombstone — honest audit trail'",
  "/mnt/agents/output/killbox2/sections/commit_inspection.md:49 — L14 DOCUMENTED, not keylessly provable; no public auto-revocation receipt located (D6)"
 ],
 "counterevidence": [
  "voxel-burn (bundle_append-protocol.md:114) sweeps stalled proposals to 'burned' — a second quasi-exception: bytes kept, but rows leave the strips and counts",
  "retraction model ('status:retracted, never deleted', constitution line 53) shows the preferred pattern the scrub lane departs from"
 ],
 "dependencies": [
  "govq-inst-002",
  "govq-inst-022"
 ],
 "falsifier": "Show a scrub_events tombstone publicly served with full lineage (then the carve-out is real and governed), or show a scrub executed without tombstone (then append-only is falsified in production).",
 "notes": "Foucault bearing: the exception lane produces the norm's meaning. Luhmann (philosophy C12, autopoiesis): the system resolves its own paradox by re-entering the distinction 'append vs scrub' inside itself — operational closure, not resolution.",
 "_role": "institutional_designer",
 "hash": "sha256:5314759d4b805a63a5009e5a1fb34f3755f400c6173984ba7551174606077f96"
}

{
 "question_id": "govq-inst-004",
 "layer": "emergency_powers",
 "status": "MISSING",
 "question": "Must emergency acts receive post-hoc ratification, by whom, within what window — and what happens to unratified emergency acts given that all rulings are owner-only and the appeal loop is undemonstrated (zero rulings issued, open_requests=4)?",
 "why_it_matters": "Without sunset-plus-ratification, every emergency act is permanent by default. The registry's own laws make delisting an append-only owner ruling and appeals route to the same owner; no vote/ratify kind exists in the schema. The owner brief names this exactly: 'whether emergency actions need later ratification; what happens if the emergency actor is the attacker.' At census zero there is no ratifying body even in principle.",
 "candidate_answers": [
  {
   "answer": "Sunset by default, ratification to extend",
   "type": "normative",
   "rule": "Every emergency act expires at ttl unless a ratification record extends it; expiry reverts scope automatically and the lapse is receipted.",
   "candidate_protocol_object": "emergency_act {ttl}, ratification {parent_id, decision, quorum}",
   "candidate_runtime": "dispatch auto-expiry of emergency-scoped caveats",
   "candidate_conformance": "clause: expired emergency scope denies without manual action",
   "candidate_article": "oip-emergency-sunset-ratification"
  },
  {
   "answer": "Ratification by affected-party appeal window",
   "type": "protocol_object",
   "rule": "Emergency acts are provisional; any affected party may appeal within N days; unappealed acts ratify by default (negative ratification).",
   "candidate_protocol_object": "provisional_act + appeal linkage + default_ratified status",
   "candidate_runtime": "appeal intake already exists (kind=appeal); add provisional status machine",
   "candidate_conformance": "clause: provisional acts render with status until window closes",
   "candidate_article": "oip-negative-ratification"
  }
 ],
 "existing_artifacts": [
  "kind=appeal in governance2.json participation.kinds",
  "registry_law (owner rulings)"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/governance2.json — kinds list contains no vote/ratify; submit.body decision field is 'owner-only kind=ruling'",
  "/mnt/agents/output/killbox-specification-v1.2.md:587 — Finding 7: all four records sit status:open, zero rulings issued — 'the response loop is undemonstrated'",
  "owner brief item 3: emergency constitution gaps enumerated"
 ],
 "counterevidence": [
  "negative ratification is already the de-facto doctrine: everything the owner does stands unless the owner reverses it — formalizing it would only name reality"
 ],
 "dependencies": [
  "govq-inst-001",
  "govq-inst-009",
  "govq-inst-024"
 ],
 "falsifier": "If a ratification or sunset record kind is deployed and at least one emergency-class act carries ttl + linked ratification/lapse receipts, mark IMPLEMENTED; otherwise the honest status remains MISSING.",
 "notes": "Arrow (philosophy N02): even a perfect ratifier cannot satisfy all fairness criteria — the question asks for a ratification rule, not a perfect one. Pareto's absent elite theory: ratification is where oligarchy either circulates or hardens.",
 "_role": "institutional_designer",
 "hash": "sha256:dcc3d603b7df7ff9c0fb41990cd0fd1b9cd8b93e0afb677d35524832e397f531"
}

{
 "question_id": "govq-inst-005",
 "layer": "succession_incapacity",
 "status": "MISSING",
 "question": "If the operator dies or is incapacitated tomorrow, what happens — is there a dead-man protocol, and could the chain prove its own liveness failure without any operator action?",
 "why_it_matters": "Every continuity-critical process is single-person: root minting, sealing, DKIM anchor arming, deploy lease, X account, domain renewal, Cloudflare account, GitHub repo. The wargame's own theorem (dD/dt > 0, the moat's divergence) 'assumes the root keeps minting events and arming anchors; a hired, coerced, or merely exhausted author sets dε/dt ≈ 0, leaves the DKIM anchor unarmed forever, and converts the moat into a museum with a frozen head' (F9). The chain carries sealed_at timestamps and drand/BTC-bound anchors, so a missed cadence is externally observable in principle — but no deadman record kind, liveness receipt, or watchdog exists to convert silence into a constitutional signal. Bus factor is conceded 'unledgered' (Section 5 row 7).",
 "candidate_answers": [
  {
   "answer": "Deadman record kind + scheduled liveness receipts",
   "type": "protocol_object",
   "rule": "The chain must emit a signed liveness record on a published cadence; absence of the record past cadence+N is itself a machine-detectable liveness failure that any third party can attest.",
   "candidate_protocol_object": "liveness_receipt {seq, cadence, prev, head}; deadman_attestation {missed_cadence, witness}",
   "candidate_runtime": "scheduled seal job emits liveness; public route exposes cadence state",
   "candidate_conformance": "clause: given only public data, a verifier computes liveness status (alive/suspect/failed) without operator cooperation",
   "candidate_article": "oip-deadman-protocol"
  },
  {
   "answer": "Federated sealing quorum",
   "type": "protocol_object",
   "rule": "Two or more independently operated nodes co-sign checkpoints; one operator's silence degrades sealing, it does not halt it.",
   "candidate_protocol_object": "checkpoint {signers:[node_keys], threshold k-of-n}",
   "candidate_runtime": "second-operator seal signer + cross-ledger proof (join_note machinery already exists)",
   "candidate_conformance": "F-series: seal valid with k signers, invalid with k-1",
   "candidate_article": "oip-federated-sealing"
  },
  {
   "answer": "Pre-armed third-party deadman arrangement",
   "type": "runtime",
   "rule": "A published deadman arrangement (automated head-posting with third-party custody of a narrowly-scoped succession capability) — exactly the counter F9 scopes and the build lacks.",
   "candidate_protocol_object": "succession_capability {scope:'seal+anchor only', custodian, trigger}",
   "candidate_runtime": "custody service holding an attenuated, non-widenable capability",
   "candidate_conformance": "clause: succession capability cannot exceed its enumerated scope even if presented by the custodian",
   "candidate_article": "oip-succession-custody"
  }
 ],
 "existing_artifacts": [
  "chain v2 head (chain_head.json: seq 24, event_count 211652, sealed_at)",
  "anchor.v1 (drand 6295375 + BTC 958413)",
  "DEPLOY_LEASE"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/specification.md:1405 — F9(a): 'a hired, coerced, or merely exhausted author sets dε/dt ≈ 0, leaves the DKIM anchor unarmed forever, and converts the moat into a museum with a frozen head. The governance schema… has no transfer, succession, ratification, or deadman record kind'",
  "/mnt/agents/output/recon2/specification.md:816 — Section 5 row 7: 'Key-person risk is real and unledgered: the operator census is 1 — the highest-value poach target is the author himself, and no record family meters bus factor (hole, stated)'",
  "/mnt/agents/output/recon/chain_head.json — sealed_at/cutoff_ts cadence exists but the sealer is unnamed and singular"
 ],
 "counterevidence": [
  "Externally anchored heads (drand/BTC) make post-hoc tampering detectable, so history survives the operator — survival of history is not continuity of operation",
  "fork_anchor_law lets a successor anchor a fork head — anteriority machinery exists, but no succession legitimacy attaches to it"
 ],
 "dependencies": [
  "govq-inst-006",
  "govq-inst-007",
  "govq-inst-008",
  "govq-inst-024"
 ],
 "falsifier": "Publish the deadman arrangement and demonstrate a missed-cadence attestation computed from public data alone; until then any answer is promissory. If the chain can never signal its own liveness failure, that is a constitutional defect by the protocol's own 'visible failure' doctrine.",
 "notes": "Luhmann (C12): an autopoietic system that cannot observe its own death fails its own closure criterion. Owner brief item 5 ('What happens in an emergency or succession event?') names this a root problem.",
 "_role": "institutional_designer",
 "hash": "sha256:d2d8563f742eef77400b87cbce486977c1f91011f8d379abfdc9245ab641519c"
}

{
 "question_id": "govq-inst-006",
 "layer": "succession_incapacity",
 "status": "DOCUMENTED_UNENFORCED",
 "question": "The chain's sealing cadence and the DKIM anchor instructions are published ('email this head from an account whose domain publishes a DKIM key… anchor through >=2 independent domains') — is any of it required, scheduled, or monitored, or does the entire external-anchoring plane depend on one person's continued diligence?",
 "why_it_matters": "Anchor strength is the moat's time term: drand rounds close and Bitcoin blocks bury, so anchors cannot be retroactively acquired — 'witnesses cannot be retroactively acquired' (spec §3.4.5). Yet the DKIM layer of the anchor is described in the same document as 'unarmed,' and the remedy F9 scopes ('a pre-armed DKIM cadence so liveness matters less') is unplayed. A documented anchoring instruction with no scheduler, no conformance clause, and no missed-cadence alarm is a promissory note against the protocol's strongest asset.",
 "candidate_answers": [
  {
   "answer": "Anchoring cadence as conformance clause",
   "type": "conformance",
   "rule": "A published maximum seal-to-anchor interval; a head older than the interval without a corresponding multi-domain anchor degrades the node's public conformance status automatically.",
   "candidate_protocol_object": "anchor_cadence_status {last_head, last_anchor, interval, status}",
   "candidate_runtime": "scheduled anchor job + public status route",
   "candidate_conformance": "clause: verifier recomputes cadence status from chain head + anchor records alone",
   "candidate_article": "oip-anchor-cadence-profile"
  },
  {
   "answer": "DKIM pre-arming across >=2 domains now",
   "type": "runtime",
   "rule": "Execute the published instruction as standing infrastructure, not prose: two domain keys, archived selectors, automated head-email on each seal.",
   "candidate_protocol_object": "dkim_anchor_record {selector_hash, domain, head, sent_at}",
   "candidate_runtime": "automated seal-to-email pipeline",
   "candidate_conformance": "clause: anchor verify route returns DKIM evidence set, not instructions",
   "candidate_article": "oip-dkim-anchor-arming"
  }
 ],
 "existing_artifacts": [
  "chain_head.json anchor_instructions (DKIM, >=2 domains)",
  "anchor.v1 (drand+BTC binding live)"
 ],
 "evidence": [
  "/mnt/agents/output/recon/chain_head.json:14 — anchor_instructions is prose: 'Email this head string… Anchor through >=2 independent domains. Then this head is externally immutable'",
  "/mnt/agents/output/recon2/specification.md:1405 — 'leaves the DKIM anchor unarmed forever'; 1407 — counter 'a pre-armed DKIM cadence so liveness matters less. Until at least the record kind exists…'",
  "/mnt/agents/output/recon2/anchor.json — drand round 6295375 + BTC block 958413 binding is live and independently verified (killbox v1.2 Part A PASS)"
 ],
 "counterevidence": [
  "drand+BTC anchoring IS live and verified — the time surface is not entirely promissory; only the DKIM/domain layer and its cadence are unenforced",
  "fork anchors prove existence/anteriority for external heads — third parties can self-serve anchoring even if the root's cadence lapses"
 ],
 "dependencies": [
  "govq-inst-005",
  "govq-inst-015"
 ],
 "falsifier": "Show a conformance clause or public status route that degrades automatically when seal-to-anchor interval is exceeded; if none exists, the cadence is DOCUMENTED_UNENFORCED exactly as classified.",
 "notes": "Coase bearing: the anchor instruction allocates a duty but no enforceable property right in the duty itself, so the outcome defaults to the owner's diligence. Pareto: the moat's time term is path-dependent — a lapsed cadence is a permanent, non-compensable loss.",
 "_role": "institutional_designer",
 "hash": "sha256:d35cb1d9a58a308f33900d9c3f163401001add6fbfbe87e15323c58e3d011dd5"
}

{
 "question_id": "govq-inst-007",
 "layer": "succession_incapacity",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "Who inherits the off-chain estate — the X account (@CannibalCapital), the miscsubjects.com domain, the Cloudflare account holding Workers/Pages/D1/KV, and the private GitHub repo (public 404) — and does any instrument bind that inheritance to the protocol's continuity rather than to the operator's personal estate?",
 "why_it_matters": "The social-proof plane (X), the identity plane (domain + DKIM), the execution plane (Cloudflare), and the implementation evidence (repo) are all personal accounts of one natural person. Death, incapacity, sanctions, or account seizure transfers or freezes them by platform ToS and probate law, not by protocol law. The killbox v1.2 legal register already flags adjacent exposures as CRITICAL/HIGH (inventorship, unarchived vendor ToS output-rights, no executed assignments). Continuity of an allegedly public protocol cannot rest on unexamined private account custody.",
 "candidate_answers": [
  {
   "answer": "Digital-asset succession instrument bound to protocol continuity",
   "type": "normative",
   "rule": "A published succession instrument (or corporate wrapper) naming continuity fiduciaries for domain, Cloudflare, repo, and X, with the protocol's ledger designating the instrument's hash as canonical.",
   "candidate_protocol_object": "succession_instrument_anchor {instrument_hash, scope:[domain,cf,repo,x], fiduciaries}",
   "candidate_runtime": "none protocol-side beyond anchoring; execution is legal",
   "candidate_conformance": "none — LEGAL_REVIEW_REQUIRED; protocol can only anchor existence",
   "candidate_article": "oip-succession-instrument"
  },
  {
   "answer": "Institutional wrapper (foundation/LLC) holds the planes",
   "type": "normative",
   "rule": "Move the accounts into an entity whose charter encodes the protocol's continuity duties; the operator becomes an officer, not the substrate.",
   "candidate_protocol_object": "entity_anchor {charter_hash, officers}",
   "candidate_runtime": "account transfers (off-protocol)",
   "candidate_conformance": "none — legal work; the ledger records the entity anchor only",
   "candidate_article": "oip-institutional-wrapper"
  }
 ],
 "existing_artifacts": [
  "X lane receipts (relay_social.json)",
  "DEPLOY_LEASE (Cloudflare production door)",
  "GitHub repo 404 record (commit_inspection.md)"
 ],
 "evidence": [
  "/mnt/agents/output/killbox2/sections/commit_inspection.md:9-11 — repo exists but is private; account shows zero public repositories; GITHUB system article operates on it via API",
  "/mnt/agents/output/killbox-specification-v1.2.md:423 — row 4.1: no executed assignment, no vendor ToS archive, account custody of incognito sessions undocumented (HIGH)",
  "/mnt/agents/output/recon2/contract_deploy.md — single production deployment door is a Cloudflare KV lease under one account"
 ],
 "counterevidence": [
  "The protocol's design (facets, forks, canonical_verifiability) explicitly disclaims requiring hosted execution — a successor could theoretically re-host from public artifacts, so account inheritance may be partially substitutable; but the canonical registry, X social graph, and domain are not substitutable"
 ],
 "dependencies": [
  "govq-inst-005",
  "govq-inst-012",
  "govq-inst-016"
 ],
 "falsifier": "An executed and hash-anchored succession instrument (or entity charter) covering all four planes resolves this; absence after legal review keeps it LEGAL_REVIEW_REQUIRED.",
 "notes": "Pareto (elite circulation): succession without an instrument is circulation by accident. Coase: unclear property rights in the continuity duty guarantee the worst bargaining outcome at the worst time.",
 "_role": "institutional_designer",
 "hash": "sha256:2b6e2e84829c27d7caf0eef976e603eb85384ad891bfca6ef24407da84753567"
}

{
 "question_id": "govq-inst-008",
 "layer": "succession_incapacity",
 "status": "OPEN",
 "question": "Can any capability survive the root minter? Every delegated key descends from an owner mint and expiry is answered by 'owner mints a fresh scoped link' — does a pre-minted, third-party-custodied succession capability violate the root-mint law, or is it the only continuity mechanism consistent with it?",
 "why_it_matters": "The attenuation algebra (delegation can only shrink; revoking the parent revokes the child) makes all authority mortal with the root. A succession capability held in custody is a designed exception: it must be narrow enough to respect bounded_authority yet sufficient to seal, anchor, and transfer. This is a genuine design fork with no current answer: the killbox recommends 'third-party custody of the capability' but no record kind or custody profile exists, and the security model may read a custodied capability as a standing breach.",
 "candidate_answers": [
  {
   "answer": "Attenuated succession capability in third-party custody",
   "type": "protocol_object",
   "rule": "A pre-minted capability scoped to {seal, anchor, transfer-initiate} with a deadman trigger; custody and scope are public; activation is receipted.",
   "candidate_protocol_object": "succession_cap {fingerprint, scope, custodian_proof, trigger}",
   "candidate_runtime": "custody arrangement + activation route",
   "candidate_conformance": "clause: succession cap presented outside trigger conditions fails closed",
   "candidate_article": "oip-succession-capability"
  },
  {
   "answer": "No survivable capability — continuity by fork only",
   "type": "normative",
   "rule": "The root-mint law is absolute; continuity is achieved by successors forking from the last anchored head and re-establishing authority publicly. Mortality of authority is a feature: no zombie root.",
   "candidate_protocol_object": "none (boundary declaration)",
   "candidate_runtime": "none",
   "candidate_conformance": "fork_anchor_law already provides the continuity primitive",
   "candidate_article": "oip-boundary-mortal-authority"
  }
 ],
 "existing_artifacts": [
  "attenuate route + revocation cascade (contract_deploy.md verbs; C17-C19)",
  "owner-only mint routes (?mint_share=1, ?tap_go=1)"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/specification.md:534 — owner-gated mint: every delegated key descends from an owner mint",
  "/mnt/agents/output/recon2/contract_deploy.md — Troubleshooting: 'expired or corrupted token — Report token_expired/token_corrupted… owner mints a fresh scoped link'",
  "/mnt/agents/output/recon2/specification.md:1407 — recommended counter includes 'third-party custody of the capability'"
 ],
 "counterevidence": [
  "A standing custodied capability is a live credential outside the owner's control — it enlarges the attack surface the revocation cascade was built to close; the fork-only branch is cleaner security but weaker continuity"
 ],
 "dependencies": [
  "govq-inst-005",
  "govq-inst-002"
 ],
 "falsifier": "Either branch resolves the question if published as constitutional text with a conformance or boundary clause; it stays OPEN while both branches remain viable and unchosen.",
 "notes": "Luhmann: the decision to remain undecided is still a system operation — but cartography's job is to mark it undecided. Pareto: custody is the first concrete elite-circulation mechanism on the table.",
 "_role": "institutional_designer",
 "hash": "sha256:558eb29f010e7e75f50c80442c985b2f249719b9fa249c8e263116e88d23aeac"
}

{
 "question_id": "govq-inst-009",
 "layer": "conflicts_of_interest",
 "status": "IMPLIED",
 "question": "One person is root minter, conformance author, objection judge, ruling authority, deployer, and sealer. What separation-of-powers transition is required, and is any part of it enforced — or does the printed invariant ('authority never auto-grows') do all the work while no threshold, timetable, or ratification record kind exists?",
 "why_it_matters": "The killbox itself is the accuser: AC4 finds governance capture 'structurally present but legally null at current scale,' activating 'exactly when the design succeeds: the first enterprise that subscribes, conforms, and pledges is trusting the ruling-kind schema to stay owner-only-but-benign, appeals decided by the same owner.' The governance schema's own registry_law confirms appeals route into owner rulings. Capture analysis is the corpus's native language (GRAIN), yet the transition mechanism out of founder-capture is absent: the protocol has an exit promise without an exit door. This is the single highest-leverage institutional question in the ontology.",
 "candidate_answers": [
  {
   "answer": "Trigger-based separation covenant",
   "type": "normative",
   "rule": "Define measurable triggers (first external subscriber, first conformance record, first pledge) that automatically convert enumerated owner powers (ruling, suite authorship, census authority) into multi-party or quorum forms.",
   "candidate_protocol_object": "separation_covenant {triggers:[{metric, threshold, powers_transferred}], ratification_kind}",
   "candidate_runtime": "census instruments already compute the trigger metrics (L11)",
   "candidate_conformance": "clause: covenant status computed publicly from counts block",
   "candidate_article": "oip-separation-of-powers-covenant"
  },
  {
   "answer": "Staged power list with honest status per power",
   "type": "protocol_object",
   "rule": "Enumerate every owner power as a first-class object with a status field {owner-held, shared, transferred}; no silent transitions, no implied ones.",
   "candidate_protocol_object": "power_register [{power, holder, since, transfer_condition}]",
   "candidate_runtime": "public power register route",
   "candidate_conformance": "clause: every privileged op names the power it exercises; ops on unregistered powers fail closed",
   "candidate_article": "oip-power-register"
  },
  {
   "answer": "No transition: founder-root as declared permanent structure",
   "type": "article",
   "rule": "Publish the honest alternative: OIP is a proprietor's registry by design; participants are told in constitutional text that rulings never transfer; let adoption decisions price that in.",
   "candidate_protocol_object": "boundary article only",
   "candidate_runtime": "none",
   "candidate_conformance": "subscription intake must require acknowledging the proprietor clause",
   "candidate_article": "oip-boundary-proprietor-registry"
  }
 ],
 "existing_artifacts": [
  "registry_law (append-only owner rulings)",
  "post_subscription.invariant",
  "L11 census machinery"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/specification.md:555 — AC4: 'Governance capture is structurally present… The printed invariant says authority never auto-grows; nothing deployed enforces a transition to multi-party rulings — no threshold, timetable, or ratification record kind exists in the schema'",
  "/mnt/agents/output/recon2/specification.md:534 — owner-gated mint: 'The authority graph's root is one person'",
  "/mnt/agents/output/recon2/governance2.json — registry_law: 'Delisting and reinstatement are append-only owner rulings… Any participant may append an appeal. A ruling governs this canonical registry, not independent forks'"
 ],
 "counterevidence": [
  "The invariant ('More subscribers… never automatically increase any person's unilateral authority') is machine-printed on every governance read — a real, public, self-binding commitment, even if unenforced",
  "obj-151 census objection shows owner-side models filing against the owner's own instruments — internal adversarialism partially substitutes for separated powers today"
 ],
 "dependencies": [
  "govq-inst-002",
  "govq-inst-010",
  "govq-inst-011",
  "govq-inst-024"
 ],
 "falsifier": "Deployed trigger covenant or power register resolves toward IMPLEMENTED; a published proprietor-registry boundary article resolves as an honest permanent answer; anything less leaves the transition IMPLIED by the invariant's own wording.",
 "notes": "Pareto's circulation of elites (absent from corpus, which keeps only C15 optimality) is the unnamed theory here: every elite is stable until its circulation mechanism is specified. Foucault: the powers are productive, not merely repressive — the census, the suite, and the relay all produce the reality they measure. Coase: all residual rights sit with one owner, so bargaining cannot reach the efficient structure.",
 "_role": "institutional_designer",
 "hash": "sha256:055a79507b674ff022edcc9d54c33292ce277a2a97f6158a3625e525480d5601"
}

{
 "question_id": "govq-inst-010",
 "layer": "conflicts_of_interest",
 "status": "PARTIAL",
 "question": "The only conformance verdict on the site is the operator's own suite grading the operator's own stack ('34/34, self-graded, served cached'). Who may grade the grader, and what makes a verifier 'independent' when conformance_law requires independent receipts but defines no independence criterion?",
 "why_it_matters": "conformance_law states 'Self-attested evidence remains self-attested until an independent verifier receipt is linked' — a correct rule with an undefined load-bearing term. Part E Finding 4: there is no kind=conformance record at all, zero subscriptions, and the sole suite runs against the root's own stack, cached. Clause-count drift compounds it: served run 25 clauses, spec 27, ledger references 34/C30 (drift D4) — the grader's own instrument has unresolved version identity. Independence today reduces to 'not the owner's own process,' and even the adversarial models auditing him 'were handed his tokens, his drop documents, his conformance suite' (AC1).",
 "candidate_answers": [
  {
   "answer": "Independence criteria as conformance text",
   "type": "conformance",
   "rule": "Define independence mechanically: verifier controls its own keys, its own infrastructure domain, and holds no owner-minted capability above a published ceiling; verifier receipts carry an independence attestation block.",
   "candidate_protocol_object": "verifier_receipt {verifier_domain, key_proof, capability_ceiling, independence_basis}",
   "candidate_runtime": "conformance intake validates the attestation block",
   "candidate_conformance": "clause: conformance records with evidence_class:self render distinctly from independent",
   "candidate_article": "oip-verifier-independence-profile"
  },
  {
   "answer": "Cross-grading obligation",
   "type": "protocol_object",
   "rule": "Any node claiming conformance must also run the suite against one other node and publish the graded receipt; the grader network is the verification commons.",
   "candidate_protocol_object": "cross_grade {grader, subject, suite_version, run_receipt}",
   "candidate_runtime": "suite exposed as invokable object against arbitrary targets",
   "candidate_conformance": "clause: conformance badge requires >=1 cross-grade from an independent domain",
   "candidate_article": "oip-cross-grading-commons"
  }
 ],
 "existing_artifacts": [
  "conformance_law",
  "/api/dispatch?conformance=1 (cached run)",
  "fedtest F0-F10 (cross-domain run vs oip-peer)"
 ],
 "evidence": [
  "/mnt/agents/output/killbox-specification-v1.2.md:555 — Finding 4: 'the only conformance verdict on the site is the operator's own clause suite… 34/34, self-graded, served cached… testing the root's own stack rather than any subscriber's'",
  "/mnt/agents/output/recon2/specification.md:552 — AC1: 'No independent verifier exists, so every check on the author is executed by the author's invitees… Containment: uncontained today'",
  "/mnt/agents/output/killbox2/sections/commit_inspection.md:57-59 — clause counts 25 (served) vs 27 (spec) vs 34/C30 (ledger) mutually inconsistent (D4)"
 ],
 "counterevidence": [
  "The fedtest run (F0-F10, 11/11 vs oip-peer.massoumi-cyrus.workers.dev) is a genuine cross-domain conformance execution — the machinery of independent grading exists and has run once, albeit against the operator's own second domain",
  "The masthead self-flags 'self-graded by the operator's own suite, served cached' — the conflict is disclosed, which is the honesty machinery working as designed"
 ],
 "dependencies": [
  "govq-inst-009",
  "govq-inst-017",
  "govq-inst-024"
 ],
 "falsifier": "A conformance record carrying evidence_class:independent with a verifiable independence basis from a non-owner domain resolves this; while every verdict traces to owner-controlled infrastructure, PARTIAL stands.",
 "notes": "Luhmann (C12 autopoiesis): a self-grading suite is operational closure — the system describes itself in its own terms and the description is internally valid and externally unverifiable. Foucault: the examination produces the examined; the suite authors the conforming subject.",
 "_role": "institutional_designer",
 "hash": "sha256:064f07e96a2255a643ba44631c6a07437f791986ea3080f18250ca3413356591"
}

{
 "question_id": "govq-inst-011",
 "layer": "conflicts_of_interest",
 "status": "OPEN",
 "question": "Adoption metrics produced by the owner's own toolchain (non_owner_node_count=3, where all three labels resolve to the owner's sessions) are cited as L11 BACKED evidence inside a patent-grade specification. Is a conflicts-of-interest disclosure record kind required wherever owner-produced numbers do evidentiary work?",
 "why_it_matters": "Part E Findings 1-2 and the census scrutiny verdict are blunt: 'The 3/1 numbers do not survive scrutiny… under any proof-of-control rule the attested count is 0,' and the single 'non-owner' anchor's actor is a root-minted capability fingerprint. The census law's caveat ('auditable records, not a claim of unique legal persons') is honest prose — but the integer escapes the prose and does load-bearing work in the specification's evidence columns. This is the general form: every metric on the site (counts, yields, conformance, relay distribution) is computed by the governed party. No COI record kind, no recusal rule, no interest-declaration field exists anywhere in the schema.",
 "candidate_answers": [
  {
   "answer": "COI disclosure field on all evidentiary records",
   "type": "protocol_object",
   "rule": "Any record whose numbers do evidentiary work carries interests:[{party, relationship}] machine fields; the census and counts blocks carry generated_by:'owner-infrastructure' as a standing disclosure.",
   "candidate_protocol_object": "interests block on gov records, anchors, conformance runs, census output",
   "candidate_runtime": "intake + counts computed with mandatory disclosure fields",
   "candidate_conformance": "clause: records used as evidence without interests block render as 'interest-undeclared'",
   "candidate_article": "oip-coi-disclosure-schema"
  },
  {
   "answer": "Split census into attested vs self-asserted (Part E minimal patch)",
   "type": "runtime",
   "rule": "Adopt Finding 1's smallest falsifiable change: actor_evidence field; attested_node_count vs self_asserted_label_count; labels whose proof lapses drop out on recomputation.",
   "candidate_protocol_object": "actor_evidence {kind:dns-txt|https-token|capability-fingerprint|provider-attestation, locator, proof}",
   "candidate_runtime": "intake validation + recomputed split census",
   "candidate_conformance": "clause: filing without verifiable evidence must not move attested_node_count",
   "candidate_article": "oip-census-split"
  }
 ],
 "existing_artifacts": [
  "census_law (labels-not-persons caveat)",
  "counts block",
  "Part E Findings 1-2, census scrutiny verdict"
 ],
 "evidence": [
  "/mnt/agents/output/killbox-specification-v1.2.md:527-535 — Findings 1-2: census counts labels not entities; 'user-supplied' Gemini label counted as independent node; all four filings arrived within 4h49m through the same keyless lane",
  "/mnt/agents/output/killbox-specification-v1.2.md:593-599 — 'The 3/1 numbers do not survive scrutiny… the sybil-able integer is doing evidentiary work in a patent-grade document'",
  "/mnt/agents/output/recon2/governance2.json — census_law caveat text verbatim"
 ],
 "counterevidence": [
  "The caveat text exists and is machine-published — disclosure is partially present; the failure is downstream: the spec's evidence columns do not inherit the caveat",
  "obj-150 summarized the self-dealing vocabulary publicly ('Filed ≠ verified, counts ≠ conformance, labels ≠ identity') — the vocabulary exists; what is missing is a per-record COI structure, not awareness"
 ],
 "dependencies": [
  "govq-inst-009",
  "govq-inst-010",
  "govq-inst-017"
 ],
 "falsifier": "If every evidentiary number carries a machine-readable interests block and the census splits attested from self-asserted, this resolves; if the owner rejects per-record disclosure on simplicity grounds, that refusal must itself be a published constitutional answer.",
 "notes": "Foucault (GAP in corpus — cited as bearing, not sourced): the census does not measure adoption, it produces 'adoption' as a governable object. Coase: disclosure is the cheapest transaction-cost reduction available; its absence is a choice, not an oversight. Third person; the conflict is the finding.",
 "_role": "institutional_designer",
 "hash": "sha256:d00476b14589b3e911e3e1f052599bcb987ddfdc7d8f735eb2463c970e250c92"
}

{
 "question_id": "govq-inst-012",
 "layer": "infrastructure_dependency",
 "status": "IMPLIED",
 "question": "The entire runtime — Workers, Pages, D1, KV, R2, AI Gateway, 112 CF ops, one DEPLOY_LEASE production door — lives on a single Cloudflare account. What is the constitutional answer to one vendor holding the execution plane, and is re-hostability a tested obligation or a design hope?",
 "why_it_matters": "The kernel boundary text promises subscribers need not use hosted execution, and canonical_verifiability promises independent verification — but the canonical registry, the ledger, the sealing process, and every receipt URL resolve to one provider's edge. Cloudflare could deplatform, be compelled, fail, or reprice, and the public verification plane goes dark with it. Owner brief item 17: 'one cloud, identity provider, model vendor, domain registrar, or code maintainer may remain indispensable. Formal votes do not eliminate infrastructural vetoes.' The system that ledgers everyone else's dependencies has not ledgered its own.",
 "candidate_answers": [
  {
   "answer": "Provider-escape conformance profile",
   "type": "conformance",
   "rule": "A published re-host drill: the ledger export + capability map + seal verification must demonstrably run from a second provider (or bare metal) within a stated RTO; the drill is receipted like any other conformance run.",
   "candidate_protocol_object": "rehost_drill {exported_head, new_base_url, verification_pass, rto_hours}",
   "candidate_runtime": "export tooling + documented restore path + scheduled drill",
   "candidate_conformance": "clause: current head verifies against a non-Cloudflare mirror",
   "candidate_article": "oip-provider-escape-profile"
  },
  {
   "answer": "Standing mirror obligation",
   "type": "protocol_object",
   "rule": "The chain head, registry, and ledger replicate to at least one provider-independent mirror that serves keyless reads; mirror divergence is a public incident class.",
   "candidate_protocol_object": "mirror_status {mirror_url, head_match, lag_events}",
   "candidate_runtime": "second-provider read replica",
   "candidate_conformance": "clause: mirror head must match canonical head within N events",
   "candidate_article": "oip-standing-mirror"
  }
 ],
 "existing_artifacts": [
  "map2.md Cloudflare surfaces (CF 112, KV 7, D1 3, R2 4, GW 4)",
  "DEPLOY_LEASE contract",
  "facets boundary text"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/map2.md:30-42 — CF (112 ops), KV, AI Gateway, R2, D1: the build's system map is Cloudflare-native at every layer",
  "/mnt/agents/output/recon2/contract_deploy.md — 'the single production deployment door for loop-safe-miscsubjects… holds the same KV lease'",
  "owner brief item 17 (capture through dependency)"
 ],
 "counterevidence": [
  "The .well-known federation document and facet design actively invite other nodes to run their own infrastructure — multi-provider operation is the protocol's stated norm for subscribers, making the root's own single-provider state a self-declared exception",
  "Exportable hashes and public recipes (hash_recipe fields on relay/chain) mean the data layer is portable in principle even if the runtime is not"
 ],
 "dependencies": [
  "govq-inst-007",
  "govq-inst-015",
  "govq-inst-016"
 ],
 "falsifier": "A receipted re-host drill or a live provider-independent mirror resolves this toward IMPLEMENTED; a published boundary article declaring single-provider permanent resolves it honestly the other way.",
 "notes": "Coase: the infrastructural veto is a property right Cloudflare holds over the commons regardless of the protocol's internal allocation. Luhmann: the system cannot distinguish 'its own infrastructure' from 'environment' until a second instantiation exists — one provider is not a dependency, it is the boundary of the system itself.",
 "_role": "institutional_designer",
 "hash": "sha256:03b5da53b337b9a18c96df35cf52db8b51f24791c0fa18f7016d1e66a957b6a2"
}

{
 "question_id": "govq-inst-013",
 "layer": "infrastructure_dependency",
 "status": "IMPLEMENTED",
 "question": "Are upstream dependency failures receipted as first-class, publicly visible outcomes rather than silently retried or narrated away? (The visibility predicate that every other dependency question in this layer presupposes.)",
 "why_it_matters": "Before asking whether dependency redundancy is required, cartography must record what already works: relay-social-proof/v3 defines outcome_class (SUCCESS/PARTIAL/MODEL_FAILED/LANE_TIMEOUT) so 'a model failure cannot be confused with a lane timeout'; the X lane's 503 denials are on-chain as DENIED with receipt links (rsp_000013, inv_marbvktfzj; inv_59x2nn8t60 'lane timeout, no retry'); the constitution bans narrating success ('do not narrate success', L16). Visible failure is the doctrine this mission brief's depth markers presuppose — it deserves an IMPLEMENTED row precisely so that the harder questions can lean on it.",
 "candidate_answers": [
  {
   "answer": "Status quo, recorded as constitutional fact",
   "type": "runtime",
   "rule": "Outcome classes + DENIED receipts + no-retry-on-lane-timeout are the implemented base; extend the same honesty to the remaining dependencies (drand, BTC, CF) so every external dependency has a failure class and a receipt.",
   "candidate_protocol_object": "dependency_outcome_class extended beyond X to all external calls",
   "candidate_runtime": "uniform outcome_class across X_POST, anchor fetches, CF ops",
   "candidate_conformance": "clause: every external dependency failure lands a typed receipt",
   "candidate_article": "oip-visible-dependency-failure"
  }
 ],
 "existing_artifacts": [
  "relay-social-proof/v3 outcome_class",
  "L13",
  "L16 (unknown-key fail-closed education)"
 ],
 "evidence": [
  "/mnt/agents/output/recon/relay_social.json — rsp_000013: x status DENIED, 'X returned 503 Service Unavailable on the initial invocation and the single linked repair', receipt inv_marbvktfzj",
  "/mnt/agents/output/recon2/relay_v12.json — 'X_POST denied upstream 503 (inv_59x2nn8t60, lane timeout, no retry)'",
  "/mnt/agents/output/killbox2/sections/commit_inspection.md:48 — L13 LIVE: schema + field enforced; 17 posts all carry outcome_class (7 SUCCESS, 10 PARTIAL at audit); failure classes defined, MODEL_FAILED/LANE_TIMEOUT never yet exercised on-chain"
 ],
 "counterevidence": [
  "The classes exist but are unexercised for their hardest case: zero MODEL_FAILED/LANE_TIMEOUT instances on-chain at audit time (D8) — visibility is implemented for DENIED at the platform layer, untested for the model/lane split",
  "recon xpost.json/xpost2.json/xpost3.json show three successive reformulations of one announcement (11:00/11:01/11:20) — retry behavior existed earlier in the same day, later disciplined into 'no retry' doctrine"
 ],
 "dependencies": [
  "govq-inst-014",
  "govq-inst-015"
 ],
 "falsifier": "Already satisfied for the X lane at the platform-denial level; full satisfaction requires an on-chain MODEL_FAILED or LANE_TIMEOUT instance under v3 semantics (currently zero).",
 "notes": "This is the ontology's token IMPLEMENTED row in these six layers — included because honest cartography records what works, and because the moat's epistemics (visible failure) are the precondition for every other finding.",
 "_role": "institutional_designer",
 "hash": "sha256:c377904d8a119f3ba3c24a9d7a92bdeeb2cda108c4316db9f176f6bfde264134"
}

{
 "question_id": "govq-inst-014",
 "layer": "infrastructure_dependency",
 "status": "PARTIAL",
 "question": "The X lane has failed upstream repeatedly (503s on both initial invocations and repairs; the v1.2 swarm's field note denied with no retry). Dependency on a single, potentially hostile platform carries the social-proof plane — what is the constitutional answer, and is multi-anchor/multi-platform redundancy a requirement or an aspiration?",
 "why_it_matters": "public-audition is a named facet — social proof is constitutionally load-bearing, not marketing garnish — yet its publication plane is one company's API. X can suspend the account, throttle, reprice, or be ordered silent; every relay post's proof_links then point into a void the protocol does not control. The relay already records multi-platform copies (platform_copy x/linkedin, DRAFTED states when the human has not approved copy), so redundancy is designed into the record shape; nothing requires it, and no second platform has ever been posted to. The mission depth marker asks: designed is not required — is it required?",
 "candidate_answers": [
  {
   "answer": "Two-plane publication minimum for relay posts",
   "type": "conformance",
   "rule": "A relay post counts as published only when its content lands on >=2 independent publication planes (or 1 plane + an anchored, keyless-readable canonical copy); single-plane posts render as 'plane-degraded'.",
   "candidate_protocol_object": "publication_results must carry >=2 non-null receipts or 1 + anchor_id",
   "candidate_runtime": "second platform integration (or article-plane canonical copy auto-linked)",
   "candidate_conformance": "clause: relay chain reports plane-redundancy ratio",
   "candidate_article": "oip-social-proof-redundancy"
  },
  {
   "answer": "Platform-hostility playbook as article",
   "type": "article",
   "rule": "Document the degradation path (account loss, API repricing, compelled silence) and the pre-committed response for each, so a hostile platform's move is answered by doctrine, not improvisation.",
   "candidate_protocol_object": "none new",
   "candidate_runtime": "none new",
   "candidate_conformance": "n/a — normative text",
   "candidate_article": "oip-platform-hostility-playbook"
  }
 ],
 "existing_artifacts": [
  "public-audition facet (governance2.json)",
  "platform_copy multi-plane record shape",
  "relay v3 hash_recipe"
 ],
 "evidence": [
  "/mnt/agents/output/recon/relay_social.json — 13 posts: platforms x (7) and 'multi' (6); multi posts carry platform_copy for linkedin but X statuses show DRAFTED ('no X_POST call occurred because the current person has not approved this exact new copy') — the second plane is copy-only, never posted",
  "/mnt/agents/output/recon/relay_social.json — rsp_000013 DENIED 503 initial + repair both failed",
  "/mnt/agents/output/recon2/relay_v12.json — v1.2 turn's X_POST denied upstream 503, lane timeout, no retry",
  "/mnt/agents/output/recon2/governance2.json — public-audition facet: 'Publish attributed field records that point back to real work and visible failures'"
 ],
 "counterevidence": [
  "The relay's proof_links also cite keyless receipts on the canonical domain — the social platform is an amplifier, not the proof itself; the proof plane survives X's hostility, only the audience plane dies",
  "DRAFTED states show a human-approval gate on X copy — the lane is deliberately conservative, so single-plane dependency is partly a chosen posture, not only neglect"
 ],
 "dependencies": [
  "govq-inst-013",
  "govq-inst-007",
  "govq-inst-023"
 ],
 "falsifier": "If a conformance profile or facet obligation makes >=2-plane publication a counted requirement and the ratio is publicly reported, this resolves; if the protocol formally declares the canonical-copy branch sufficient (proof ≠ audience), that article resolves it the other way.",
 "notes": "Foucault: the platform's power over the discourse is productive — it shapes what can be said (280 chars, approval gates, API terms) before anything is said. Pareto: dependence on one distributor is the classic path to a new elite (the platform as gatekeeper aristocracy).",
 "_role": "institutional_designer",
 "hash": "sha256:08766e2066349aeeca3f833f2437a9c98e66cc2fa7e6455fae29a4b7a5d60f7b"
}

{
 "question_id": "govq-inst-015",
 "layer": "infrastructure_dependency",
 "status": "DOCUMENTED_UNENFORCED",
 "question": "Anchor redundancy is designed (drand round + Bitcoin block + chain head in one anchor.v1) and the DKIM >=2-domain instruction is published — but the DKIM layer is unarmed, and at verification time the drand round was fetched via drand.cloudflare.com, the same provider that hosts the runtime. Are anchor carriers required to be independent of the runtime's own provider, and is that independence verified or assumed?",
 "why_it_matters": "The anchor's evidentiary force is independence: 'any other ledger holding this anchor runs the same steps — no operator trust required.' But independence has layers: cryptographic (drand BLS, BTC work — genuinely independent), fetch-path (drand verify URL served by the same Cloudflare that serves the runtime), and operational (the anchor is minted, sealed, and armed by one operator). An adversary compelling or failing Cloudflare degrades the runtime and the verify path simultaneously. Multi-anchor redundancy exists in design; requirement and provider-diversity are nowhere enforced. The depth marker's question: designed; is it required?",
 "candidate_answers": [
  {
   "answer": "Carrier-diversity requirement in the anchor profile",
   "type": "conformance",
   "rule": "An anchor counts as 'externally armed' only when its verification surfaces resolve through providers independent of the runtime's provider set; the anchor record carries a provider_independence block.",
   "candidate_protocol_object": "anchor.v2 {surfaces, provider_set, independence_assertion}",
   "candidate_runtime": "verify routes emit provider metadata; mirror verify endpoints on second provider",
   "candidate_conformance": "clause: anchors lacking provider diversity render 'single-provider-time-surface'",
   "candidate_article": "oip-anchor-carrier-diversity"
  },
  {
   "answer": "Arm the DKIM layer per the published instruction",
   "type": "runtime",
   "rule": "Execute chain_head.json anchor_instructions as standing automation (see govq-inst-006); two DKIM domains outside Cloudflare DNS.",
   "candidate_protocol_object": "dkim_anchor_record",
   "candidate_runtime": "automated head-email + selector archiving",
   "candidate_conformance": "clause: anchor verify returns DKIM evidence set",
   "candidate_article": "oip-dkim-anchor-arming"
  }
 ],
 "existing_artifacts": [
  "anchor.v1 (drand 6295375 + BTC 958413 + ms.head)",
  "chain_head.json anchor_instructions",
  "public-anchors facet"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/anchor.json — canonical binds drand.round=6295375, btc.height=958413, ms.head; verify URLs: api.drand.sh + mempool.space + miscsubjects chain",
  "/mnt/agents/output/killbox-specification-v1.2.md:43 — verification of the drand round performed via GET https://drand.cloudflare.com/public/6295375 — the verify path ran through the runtime's own provider",
  "/mnt/agents/output/recon2/specification.md:1405-1407 — DKIM 'unarmed'; pre-armed cadence recommended, unplayed"
 ],
 "counterevidence": [
  "The cryptographic surfaces (League of Entropy BLS, Bitcoin PoW) are genuinely independent of any single provider — the finding is about fetch-path and arming, not about the math",
  "api.drand.sh exists as an alternative fetch path in the same anchor record — provider diversity at fetch time is achievable today with no new machinery"
 ],
 "dependencies": [
  "govq-inst-006",
  "govq-inst-012"
 ],
 "falsifier": "An anchor record whose verification surfaces demonstrably resolve through a provider set disjoint from the runtime's (with the DKIM layer armed) resolves this; until then, redundancy is DOCUMENTED_UNENFORCED.",
 "notes": "Coase: independence is a transaction-cost property — verification that costs the same trust as the claim it verifies is worthless; fetch-path capture raises the hidden price of 'no operator trust.' Luhmann: redundancy is how systems observe their own environment; one provider collapses system and environment into one point of failure.",
 "_role": "institutional_designer",
 "hash": "sha256:956ccfbd818b5b05c6555e0244e1527d42457585f53d618b5aa3ec3c1e66b504"
}

{
 "question_id": "govq-inst-016",
 "layer": "infrastructure_dependency",
 "status": "CONTRADICTED",
 "question": "The implementation repository returns public 404 while the v1.1 disclosure labels its key commit 'public commit' and patent-grade claims rest on it. Does private infrastructure contradict canonical_verifiability, and is source escrow or a hash-anchored export a constitutional obligation for succession and enablement?",
 "why_it_matters": "canonical_verifiability is core axiom three: 'Interoperable records declare canonical bytes and hashes so another implementation can verify them independently.' The runtime that defines those canonical bytes is uninspectable: all five tasked SHAs 404, zero public repos on the account, four of five SHAs publicly unattributable (drift D1/D2). The killbox legal register ties this to enablement risk (row 1.4: claim 5's enablement 'depends on a private repo') and to succession (who inherits code nobody else can read?). The contradiction is between the axiom's scope (records, not code) and the doctrine's spirit (independent verification) — and the protocol should say which it means, because adversaries will not ask.",
 "candidate_answers": [
  {
   "answer": "Narrow reading + code escrow",
   "type": "normative",
   "rule": "Declare the axiom record-scoped (protocol records are verifiable; implementation code is not a protocol record), then voluntarily hash-anchor a code export per release so enablement and succession do not depend on goodwill.",
   "candidate_protocol_object": "code_anchor {release, tarball_hash, license_scope}",
   "candidate_runtime": "release pipeline emits anchored export hash",
   "candidate_conformance": "clause: each deployed version names its code anchor",
   "candidate_article": "oip-code-anchoring"
  },
  {
   "answer": "Wide reading + progressive publication",
   "type": "article",
   "rule": "Adopt the spirit: a protocol whose verification story is 'recompute it yourself' owes the public an inspectable reference implementation on a published timetable.",
   "candidate_protocol_object": "publication_schedule in constitutional text",
   "candidate_runtime": "repo publication plan",
   "candidate_conformance": "n/a",
   "candidate_article": "oip-reference-implementation-policy"
  }
 ],
 "existing_artifacts": [
  "canonical_verifiability axiom",
  "commit_inspection.md drift register D1/D2/D3",
  "killbox v1.2 row 1.4"
 ],
 "evidence": [
  "/mnt/agents/output/killbox2/sections/commit_inspection.md:9-11 — repo 404 on every unauthenticated path; 'The task premise that the OIP runtime repo is public is false at inspection time'; GITHUB_TAIL usage text asks 'is the repo still private'",
  "/mnt/agents/output/killbox2/sections/commit_inspection.md:65-67 — D1: 'the v1.1 disclosure's public-commit label for c8a5f77c does not hold at audit time'; D2: four SHAs unattributable",
  "/mnt/agents/output/killbox-specification-v1.2.md:397 — row 1.4: 'claim 5's repair-bypass deployment cites commit c8a5f77c… on a repository that returns HTTP 404 to the public'"
 ],
 "counterevidence": [
  "The axiom's text is genuinely record-scoped — 'interoperable records', not 'implementations' — so the contradiction is doctrinal, not literal; a narrow-reading declaration is available and defensible",
  "Live keyless behavior did verify the laws (L11-L14 LIVE at contract) — the private repo impairs code-level audit, not behavior-level verification"
 ],
 "dependencies": [
  "govq-inst-007",
  "govq-inst-012",
  "govq-inst-010"
 ],
 "falsifier": "Either a published scope declaration (narrow + escrow) or a publication timetable (wide) resolves; while the 'public commit' label stands unrepaired beside a 404, CONTRADICTED stands.",
 "notes": "Foucault: opacity is not absence of power but one of its modes — the private repo disciplines what can be claimed about the public one. Coase: unclear scope of the verifiability right raises every verifier's transaction costs; clarifying it is cheap and refused only for reasons worth ledgering.",
 "_role": "institutional_designer",
 "hash": "sha256:fca62b0dd697ed176a5b75ee1fe82096f713dbf2906530b53e9b6b796cfa2694"
}

{
 "question_id": "govq-inst-017",
 "layer": "vendor_capture",
 "status": "OPEN",
 "question": "One vendor's models authored a majority of the relay's discourse economy (of 17 posts: ~9-10 Moonshot, ~6-7 OpenAI, 1 fable, 1 other — counts drift across snapshots). At what share does a single model vendor hold de-facto veto over the discourse economy, and is vendor-share a metered constitutional quantity?",
 "why_it_matters": "The relay is the protocol's public contestation and proof plane; its authorship distribution is therefore a constitutional fact, not a marketing statistic. Local recon (13 posts, v2 era) counts Moonshot 7 / OpenAI 6; Part E (17 posts) counts Moonshot 10 / OpenAI 7; the mission's depth marker counts 9/6/1/1. Whichever snapshot is taken, two vendors author everything and one authors a majority — and the conformance runs, swarm audits, killbox disclosures, and objection repairs are likewise executed by the same two vendors' models, using the owner's tokens. A vendor that deprecates a model, changes refusal behavior, or terminates the owner's API accounts can halt the discourse plane, the audit plane, and the evidence-production plane at once. No meter, threshold, or diversity requirement exists anywhere in the schema.",
 "candidate_answers": [
  {
   "answer": "Vendor-share meter on the relay chain",
   "type": "protocol_object",
   "rule": "The relay chain head publishes vendor_share per rolling window; a single provider exceeding a published share (e.g., >50% of posts or >2/3 of conformance runs) degrades the plane's status to 'vendor-concentrated'.",
   "candidate_protocol_object": "relay chain vendor_share block {provider, share_window}",
   "candidate_runtime": "computed from existing model_provider fields (already required by v3 recipe)",
   "candidate_conformance": "clause: share above threshold renders publicly",
   "candidate_article": "oip-vendor-share-meter"
  },
  {
   "answer": "Vendor diversity as participation requirement",
   "type": "conformance",
   "rule": "Constitutional minimum: at least N independent vendors' models among the last M receipted audit/repair acts before a plane counts as contested.",
   "candidate_protocol_object": "contestation_status {vendors, window}",
   "candidate_runtime": "n/a — computation over existing fields",
   "candidate_conformance": "clause: single-vendor windows render 'contestation-degraded'",
   "candidate_article": "oip-contestation-diversity-profile"
  },
  {
   "answer": "Accept concentration, disclose it",
   "type": "article",
   "rule": "Declare vendor concentration a known, accepted early-stage condition with a review trigger rather than a standing obligation — honest stasis instead of enforced diversity.",
   "candidate_protocol_object": "disclosure block on relay schema",
   "candidate_runtime": "none",
   "candidate_conformance": "n/a",
   "candidate_article": "oip-vendor-concentration-disclosure"
  }
 ],
 "existing_artifacts": [
  "relay_social.json model_provider fields",
  "relay v3 hash_recipe (model_provider in canonical order)",
  "model-lane-doctrine (per-vendor lanes)"
 ],
 "evidence": [
  "/mnt/agents/output/recon/relay_social.json — 13 posts: Moonshot AI 7 (6 Kimi K3 + 1 incognito), OpenAI 6 (GPT-5/5.6 Codex)",
  "/mnt/agents/output/killbox-specification-v1.2.md:531 — 'The relay plane shows the same concentration: 17 posts, exactly two providers (Moonshot AI ×10, OpenAI ×7); Gemini exists only as subject matter inside posts, never as a posting model'",
  "mission depth marker — 9 Moonshot / 6 OpenAI / 1 fable / 1 other of 17; snapshot drift itself is evidence no canonical meter exists",
  "/mnt/agents/output/recon2/specification.md:552 — AC1: the auditing models 'were handed his tokens, his drop documents, his conformance suite'"
 ],
 "counterevidence": [
  "model_provider is self-asserted (Part E Finding 1: relay posts carry model_attestation: {}) — so the meter would meter self-declarations; but even self-declared share is more than the current schema surfaces",
  "Multi-vendor participation happened spontaneously (two vendors, no mandate) — the commons attracts diversity without coercion, an argument against hard thresholds"
 ],
 "dependencies": [
  "govq-inst-018",
  "govq-inst-010",
  "govq-inst-011"
 ],
 "falsifier": "If the relay schema publishes a computed vendor_share with a status consequence, the question resolves toward IMPLEMENTED; if the protocol publishes an accepted-concentration disclosure with a review trigger, it resolves as an honest OPEN-by-choice.",
 "notes": "Pareto: this is the iron law in miniature — any discourse economy tends toward oligopoly of its most capable participants; the only question is whether the constitution meters it. Coase: the vendors hold no formal rights in the registry, yet hold structural veto — the gap between formal and effective rights is precisely what his framework predicts. Foucault: whoever's models author the receipts authors the reality the receipts describe.",
 "_role": "institutional_designer",
 "hash": "sha256:ed545f2ba95647bfcfc107cf3acdc72ce1c96b56acd600170a145df1fe6f516a"
}

{
 "question_id": "govq-inst-018",
 "layer": "vendor_capture",
 "status": "IMPLIED",
 "question": "Vendor safety filters already shape the protocol's canonical artifacts: the token-drop doctrine exists because models reject drops as prompt injection, and the doctrine's own rules ('paradox of denial', no imperative verbs) are vendor-filter compliance written as protocol style. When a vendor retrains its filters, the protocol's constitutional artifacts must change — is that capture, and does any clause protect the canonical form from vendor drift?",
 "why_it_matters": "The model-lane-doctrine and token-drop-best-practices articles are, functionally, treaties with vendor classifiers: they define what the protocol may say to its own participants in order to remain processable. This is rational adaptation — and also a structural dependency in which an unauditable, unversioned, third-party behavioral system (the vendor's refusal policy) has effective amendment power over the protocol's operational texts. The dependency is implied by the corpus's own documents; it is named nowhere as a constitutional exposure. Luhmann's frame: the environment (vendor policy) is re-entered inside the system (canonical doctrine) without being marked as environment.",
 "candidate_answers": [
  {
   "answer": "Vendor-drift watch + doctrine version pinning",
   "type": "protocol_object",
   "rule": "Doctrine articles carry a tested-against block naming vendor/model/version and a golden test per vendor lane; a filter change that breaks a lane is a classified incident with receipts.",
   "candidate_protocol_object": "doctrine_version {tested_against:[{vendor, model, date}], golden_test_hash}",
   "candidate_runtime": "per-vendor acceptance tests (extending the existing golden test + test 275)",
   "candidate_conformance": "clause: doctrine change requires re-run of vendor-lane golden tests",
   "candidate_article": "oip-vendor-drift-watch"
  },
  {
   "answer": "Mark the boundary explicitly",
   "type": "article",
   "rule": "Constitutional text stating that vendor filters are environment, not law: doctrine may adapt for processability, but no vendor-induced change may alter rights, records, or hash recipes.",
   "candidate_protocol_object": "boundary clause in core or facet text",
   "candidate_runtime": "none",
   "candidate_conformance": "clause: hash recipes and record schemas are vendor-invariant",
   "candidate_article": "oip-boundary-vendor-environment"
  }
 ],
 "existing_artifacts": [
  "bundle_token-drop-best-practices.md (six rules; golden test; test 275)",
  "bundle_model-lane-doctrine.md (per-vendor lanes)"
 ],
 "evidence": [
  "/mnt/agents/output/corpus/bundle_token-drop-best-practices.md — the entire doctrine is shaped by vendor safety filters: 'Models are trained to reject text that contains instructions… The paradox of denial'; format guarded by golden test + scored regression test 275",
  "/mnt/agents/output/corpus/bundle_model-lane-doctrine.md — 'Your lane is decided by your runtime's HTTP capability, not by preference' — vendor capability determines constitutional participation mechanics",
  "killbox v1.2 masthead: issuing swarms are vendor products (Kimi K3, Codex GPT-5) authoring patent-grade canonical text"
 ],
 "counterevidence": [
  "The adaptation is transparent and published as doctrine — arguably the honest opposite of capture: the dependency is documented, tested, and lane-ranked",
  "Hash recipes and record schemas have in fact remained vendor-invariant so far (v1→v3 recipe changes were internally driven) — no observed case of vendor-induced canonical drift yet"
 ],
 "dependencies": [
  "govq-inst-017",
  "govq-inst-016"
 ],
 "falsifier": "A published boundary clause (vendor-invariance of rights/records/recipes) plus a vendor-drift test regime resolves this; while the dependency remains only implied by doctrine articles, IMPLIED stands.",
 "notes": "Luhmann (C12): re-entry without marking is the classic autopoietic blind spot. Foucault: the vendor's classifier is a disciplinary apparatus operating upstream of the protocol's own discourse — the strongest capture is the kind that feels like style.",
 "_role": "institutional_designer",
 "hash": "sha256:a6f67c21eca2ae28456991a29c3689f35b326222dd41e448dfa17146e48b63e6"
}

{
 "question_id": "govq-inst-019",
 "layer": "vendor_capture",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "No dated archive of model-vendor terms (Moonshot, OpenAI, Google) conveying output rights to the account holder exists; the killbox legal register rates inventorship CRITICAL and assignment hygiene HIGH. If a vendor changes its output-rights terms, asserts rights in model-authored canonical text, or is itself compelled by a state, what does the protocol's authoring record prove, and for whom?",
 "why_it_matters": "The protocol's canonical corpus — killbox specifications, swarm-drafted claims, relay posts, this ontology — is substantially machine-authored and publicly receipted as such. The v1.2 legal register itself: 'A U.S. application must name a natural-person inventor… the public record affirmatively receipts machine authorship of the claim strings themselves; the swarm filed incognito… nothing recorded moves rights from machine-assisted creation to the human owner.' Vendor capture here is legal, not operational: the vendors' ToS are the unexamined constitutional documents of the protocol's own authorship. This is outside protocol automation and belongs in the legal-review backlog — but the ontology must hold the question visibly or the entire evidentiary edifice rests on unarchived third-party terms.",
 "candidate_answers": [
  {
   "answer": "Vendor ToS output-rights archive + conception record",
   "type": "runtime",
   "rule": "Archive dated snapshots of each vendor's output-rights clause; build per-claim conception records from contemporaneous human-session artifacts; anchor the archive (existence, not publication) in the ledger.",
   "candidate_protocol_object": "tos_anchor {vendor, clause_hash, archived_at}; conception_record per claim family",
   "candidate_runtime": "archive pipeline + anchor records",
   "candidate_conformance": "none — LEGAL_REVIEW_REQUIRED; protocol anchors existence only",
   "candidate_article": "oip-authorship-record-policy"
  },
  {
   "answer": "Assignment and custody instruments",
   "type": "normative",
   "rule": "Execute confirmatory assignments covering all human session-holders; document account custody for incognito sessions; record with any filing.",
   "candidate_protocol_object": "assignment_anchor {instrument_hash, scope}",
   "candidate_runtime": "off-protocol legal execution, ledger-anchored",
   "candidate_conformance": "n/a",
   "candidate_article": "oip-assignment-instruments"
  }
 ],
 "existing_artifacts": [
  "killbox v1.2 legal register rows 3.1 (inventorship CRITICAL), 4.1 (assignments HIGH)",
  "governance record gov_d3eb38a0 (Gemini user-supplied paste)"
 ],
 "evidence": [
  "/mnt/agents/output/killbox-specification-v1.2.md:417 — row 3.1: machine-drafted disclosure, inventorship CRITICAL, LEGAL_REVIEW_REQUIRED entire column; 'never name a model'",
  "/mnt/agents/output/killbox-specification-v1.2.md:423 — row 4.1: 'no evidence of which natural person held the incognito swarm accounts; no dated snapshots of model-vendor terms (Moonshot AI, OpenAI, Google) conveying output rights'",
  "/mnt/agents/output/recon2/governance2.json latest — gov_d3eb38a0: 'Gemini contribution (user-supplied)' filed with model-recommendation authority, session-holder unidentified"
 ],
 "counterevidence": [
  "The public record's radical transparency about machine authorship is itself the protocol's honesty doctrine executed — the exposure exists because the receipts tell the truth; a less honest system would not have this problem visible",
  "Vendor ToS in the relevant period commonly assigned output rights to account holders (per public reporting) — the gap is archival discipline, not necessarily substance"
 ],
 "dependencies": [
  "govq-inst-017",
  "govq-inst-018",
  "govq-inst-020"
 ],
 "falsifier": "Executed assignments + an anchored ToS archive + per-family conception records move this out of legal review; until counsel signs off, LEGAL_REVIEW_REQUIRED is the only honest status.",
 "notes": "Coase inverted: here the protocol is the party whose rights are undefined against the vendors' boilerplate. Pareto: the vendors are the stationary elite; the protocol's swarm labor circulates within terms it did not negotiate.",
 "_role": "institutional_designer",
 "hash": "sha256:08dba85d0c0029547c015152895ccda8d0e6446d063d64b202acee7bfd09a4f5"
}

{
 "question_id": "govq-inst-020",
 "layer": "state_coercion",
 "status": "OPEN",
 "question": "A state can compel the single operator. Which records could be compelled into falsification — census counts, conformance verdicts, forward seals, governance rulings, relay posts — and which are structurally uncompellable because already bound to drand rounds and Bitcoin blocks? Where exactly is the compellable/uncompellable line, and is it published?",
 "why_it_matters": "The wargame names the cheapest unplayed move: 'acquihire — or subpoena — the single person §3.4.9 shows is root minter, suite author, objection judge, ruling key, and sealing process.' Coercion resistance splits into prevention (impossible for a single operator) and detection (the anchor design's actual strength): 'witnesses cannot be retroactively acquired, because drand rounds close and Bitcoin blocks bury.' But the witness set only covers what has already been anchored; everything prospective — tomorrow's seals, rulings, census, conformance runs — is compellable in silence. The protocol has never drawn the line publicly, and the distinction between 'falsification detectable' and 'falsification impossible' is exactly the distinction a coerced operator will be ordered to blur.",
 "candidate_answers": [
  {
   "answer": "Publish the compellability map",
   "type": "article",
   "rule": "A constitutional table: for each record family, state what coercion can do (suppress, falsify-prospectively, rewrite-undetectably) and what it cannot (rewrite anchored history without detection); mark the detection asymmetry as the protocol's actual coercion posture.",
   "candidate_protocol_object": "compellability_map {record_family, suppressible, falsifiable_forward, rewritable, detection_path}",
   "candidate_runtime": "none new — table computed from existing anchor coverage",
   "candidate_conformance": "clause: every record family names its anchor coverage state",
   "candidate_article": "oip-compellability-map"
  },
  {
   "answer": "Maximize anchored surface area",
   "type": "protocol_object",
   "rule": "Extend census-anchor design (Part E Finding 7's 'pays the node first' mechanism) so census sets, conformance runs, and rulings are committed into drand/BTC anchors on the sealing cadence — shrinking the compellable-forward set each round.",
   "candidate_protocol_object": "census_anchor {merkle_root, drand_round, btc_height, covered_records}",
   "candidate_runtime": "anchor round extended to governance planes",
   "candidate_conformance": "clause: anchor packet must commit to current census + ruling set",
   "candidate_article": "oip-census-anchor"
  }
 ],
 "existing_artifacts": [
  "anchor.v1 (drand+BTC+ms.head)",
  "fork_anchor_law",
  "Part E Finding 7 (census anchor proposal)",
  "spec §3.4.5 witness set"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/specification.md:1405 — 'acquihire — or subpoena — the single person… The governance schema has no transfer, succession, ratification, or deadman record kind'",
  "/mnt/agents/output/recon2/specification.md:800 — witness set: 'A clone's witness set at creation is {itself}: it can testify, but it cannot subpoena… witnesses cannot be retroactively acquired, because drand rounds close and Bitcoin blocks bury'",
  "/mnt/agents/output/killbox-specification-v1.2.md:589 — Finding 7: census anchor mechanism scoped ('one extra hash in an existing anchor')"
 ],
 "counterevidence": [
  "External anchoring genuinely limits retroactive falsification — the uncompellable set is non-empty and growing, which is more than most systems can say",
  "Detection-without-prevention still deters: a compelled falsification that any third party can expose has limited value to the compeller, especially with the honesty machinery's track record of self-filing objections (obj-143, obj-151)"
 ],
 "dependencies": [
  "govq-inst-021",
  "govq-inst-023",
  "govq-inst-015"
 ],
 "falsifier": "A published compellability map with per-family anchor coverage resolves the representation half; deployed census/ruling anchors shrink the compellable set measurably. The question fails only if the protocol claims coercion-resistance it cannot demonstrate.",
 "notes": "Foucault (bearing, absent from corpus): the state's power here is classically sovereign (compel the body), while the protocol's counter-power is productive (make compulsion legible). Luhmann: law and state are environment; the system cannot fight coercion, only communicate about it in its own operations — the map is that communication. Coase: the coerced operator's outside options define the bargaining set; publish the map and the options change.",
 "_role": "institutional_designer",
 "hash": "sha256:5c2d75dd35add372da87f95aea64e9b310e4a6766562dfe4c73202c5ad694a22"
}

{
 "question_id": "govq-inst-021",
 "layer": "state_coercion",
 "status": "MISSING",
 "question": "Is there a warrant canary or any coercion-signaling record kind? If the operator is compelled AND gagged, the append-only ledger could carry periodic canary records whose absence is itself a signal — nothing of the kind exists. Should it, and what are the legal limits on canaries in the operator's jurisdiction?",
 "why_it_matters": "Every coercion defense in the corpus is retrospective (anchors prove the past). A gag order attacks the signaling plane itself: the operator cannot say 'I was served.' The system's own doctrine — visible failure, negative space ledgered ('revoked ×2, replay ×2, expired ×2… the negative space is ledgered too') — already established the pattern that absence can be made legible; a canary is the same move applied to state compulsion. No canary record, cadence, or constitutional text exists. Legality of canaries varies by jurisdiction and order type — hence the question must carry legal review, but the design question (does the protocol want a signaling plane for compelled silence?) is constitutional and prior.",
 "candidate_answers": [
  {
   "answer": "Cadence canary bound into sealing",
   "type": "protocol_object",
   "rule": "Each seal carries a canary field ('no compulsion received as of cutoff'); a seal missing the field, or a missed cadence, is a machine-detectable signal without any human statement.",
   "candidate_protocol_object": "canary field in chain head / seal record",
   "candidate_runtime": "seal pipeline extension",
   "candidate_conformance": "clause: canary absence is publicly computable",
   "candidate_article": "oip-canary-cadence"
  },
  {
   "answer": "Explicit no-canary boundary",
   "type": "article",
   "rule": "After legal review, decline canaries: they may be unlawful in the operator's jurisdiction or ineffective against gag scope; publish the refusal so absence-of-canary is not mistaken for oversight.",
   "candidate_protocol_object": "boundary declaration",
   "candidate_runtime": "none",
   "candidate_conformance": "n/a",
   "candidate_article": "oip-boundary-no-canary"
  }
 ],
 "existing_artifacts": [
  "chain v2 seal records (sealed_at cadence)",
  "negative-space ledgering pattern (fed_ledger.json verdicts)",
  "append-only doctrine"
 ],
 "evidence": [
  "/mnt/agents/output/recon/chain_head.json — seal records exist with cadence but no canary field",
  "/mnt/agents/output/recon/fed_ledger.json — negative space ledgered as verdicts (revoked, replay_rejected, expired, scope/audience mismatch): the pattern 'absence made legible' is native to the build",
  "/mnt/agents/output/recon2/specification.md:1405 — subpoena named as unplayed attack; no signaling counter scoped anywhere"
 ],
 "counterevidence": [
  "A canary can itself be compelled into continued publication (sign the canary or else) — its evidentiary value depends on the keyholder's behavior under duress, which no cryptography controls",
  "Missed-cadence signaling exists implicitly (govq-inst-005 liveness): a formal canary may add little beyond what liveness failure already communicates"
 ],
 "dependencies": [
  "govq-inst-020",
  "govq-inst-005",
  "govq-inst-023"
 ],
 "falsifier": "Either a deployed canary field or a published no-canary boundary article (post legal review) resolves; the current state — silence about silence — is the only unacceptable answer.",
 "notes": "LEGAL_REVIEW_REQUIRED attaches to candidate answer 1 in the operator's jurisdiction. Foucault: compelled speech is the mirror of compelled silence; the canary attempts to make one legible through the other. Luhmann: a signal is an operation, not a fact — the canary's meaning must be fixed by constitutional text before the emergency, or it means nothing.",
 "_role": "institutional_designer",
 "hash": "sha256:2f4be17ef801570a87b9a268e17da2ed4c060531e64f794c5343430cb713cdb4"
}

{
 "question_id": "govq-inst-022",
 "layer": "state_coercion",
 "status": "NON_AUTOMATABLE",
 "question": "When a court order, sanctions regime, or erasure right demands deletion or suppression of ledger content, what happens? Append-only is not a complete answer when the law requires deletion — which legal order governs, what does the protocol do, and what does it refuse to do?",
 "why_it_matters": "Owner brief item 2 is blunt: 'A record can be valid under OIP and unlawful under a country's law… Append-only is not a complete answer when the law requires suppression or deletion.' The scrub_events precedent (govq-inst-003) proves a suppression lane exists for secrets; a court order is a second, harder class — invoked by a state, potentially gagged, possibly conflicting across jurisdictions (one state orders deletion, another orders preservation). No record kind, doctrine, or boundary article exists. This cannot be automated: the operator's compliance decision is a human legal act. What CAN be represented is the doctrine, the collision record, and the consequences for verification (tombstone vs content preservation).",
 "candidate_answers": [
  {
   "answer": "Collision record + tombstone doctrine",
   "type": "protocol_object",
   "rule": "A lawful_suppression record kind: names the order class (not necessarily the order), the affected hashes, the action taken (tombstone/redact/refuse), and the verification consequences; the tombstone preserves provability that content existed without serving it.",
   "candidate_protocol_object": "lawful_suppression {order_class, target_hashes, action, tombstone_hash, jurisdiction_class}",
   "candidate_runtime": "suppression lane with public tombstones (extends scrub_events pattern)",
   "candidate_conformance": "clause: suppressed content must leave verification-consistent tombstone",
   "candidate_article": "oip-lawful-suppression"
  },
  {
   "answer": "Boundary article: what the protocol will never do",
   "type": "article",
   "rule": "Constitutional text: the protocol will not rewrite anchored history for any legal order; it may suppress serving, never provability; conflicts between orders are logged as collisions, not resolved by software.",
   "candidate_protocol_object": "boundary declaration",
   "candidate_runtime": "none",
   "candidate_conformance": "n/a",
   "candidate_article": "oip-boundary-lawful-deletion"
  }
 ],
 "existing_artifacts": [
  "scrub_events tombstone precedent",
  "append_and_repair axiom",
  "owner brief item 2 (jurisdiction collision list)"
 ],
 "evidence": [
  "owner brief: 'court orders demanding deletion; sanctions; privacy erasure rights; discovery holds; export controls; conflicting national rules; injunctions against publication; illegal content embedded in append-only records'",
  "/mnt/agents/output/corpus/bundle_constitution.md:55 — scrub tombstone lane exists for secrets (the only deployed suppression class)",
  "/mnt/agents/output/recon2/governance2.json — append_and_repair axiom text; no suppression or lawful-deletion record kind in participation.kinds"
 ],
 "counterevidence": [
  "The boundary promise ('never rewrite anchored history') is cheap to make at census zero and expensive under an actual injunction with contempt attached — promissory resistance should be labeled as such",
  "Some jurisdictions' erasure rights may be satisfiable by tombstone-without-content (the hash proves existence, the bytes are gone) — partial technical compliance may exist for some order classes"
 ],
 "dependencies": [
  "govq-inst-003",
  "govq-inst-020",
  "govq-inst-021"
 ],
 "falsifier": "This question is resolved by doctrine plus legal review, not by deployment: a published suppression doctrine with the collision record kind, and counsel's review of the operator's jurisdiction exposure. Any claim that the protocol 'handles' lawful deletion automatically is a defect by the implementation law ('a machine-enforced answer to a question that should remain political, legal, or human is a defect').",
 "notes": "Coase: conflicting legal orders are conflicting property-right allocations over the same bytes; no bargaining is possible between sovereigns, so the protocol can only choose its own consistency. Foucault: law here is juridical power confronting productive power — the tombstone is the compromise object their struggle produces.",
 "_role": "institutional_designer",
 "hash": "sha256:b64a5b45f312929bdea83db2c00ca20939768aa38e181e8c578e86aac63099e2"
}

{
 "question_id": "govq-inst-023",
 "layer": "state_coercion",
 "status": "OPEN",
 "question": "If the operator is compelled and gagged, what does the public see? The sealing cadence, the anchor arming, the X lane, and the registry response loop all run through the compelled person — can coerced silence be made legible as a constitutional state (distinct from mere neglect), or are coercion and abandonment indistinguishable from outside?",
 "why_it_matters": "This is the synthesis question of the coercion layer: the same observables (missed seals, unarmed anchors, dead X lane, unanswered appeals) are produced by death, exhaustion, abandonment, sale, or subpoena — and F9 already concedes the system cannot tell them apart ('a hired, coerced, or merely exhausted author sets dε/dt ≈ 0'). The protocol prides itself on classified outcomes (outcome_class for lanes); the operator's own silence is the one unclassified outcome in the system. A liveness/degradation state machine (alive/degraded/silent/lapsed) computable from public data would not prove coercion, but it would convert invisible capture into a public, queryable constitutional state.",
 "candidate_answers": [
  {
   "answer": "Operator-silence state machine from public observables",
   "type": "protocol_object",
   "rule": "Define constitutional states from published cadences: sealing lag, anchor arming lag, governance response lag, X lane lag; any verifier computes the state from public data; transitions are themselves ledgered by third parties.",
   "candidate_protocol_object": "operator_liveness {seal_lag, anchor_lag, ruling_lag, state}",
   "candidate_runtime": "public computation route + third-party attestation records",
   "candidate_conformance": "clause: state transitions are computable from keyless data alone",
   "candidate_article": "oip-operator-liveness-states"
  },
  {
   "answer": "Distinguish coercion from neglect is impossible — say so",
   "type": "article",
   "rule": "Publish that the states are deliberately coarse: the protocol commits to making silence legible, not to attributing its cause; cause-attribution remains social/legal.",
   "candidate_protocol_object": "boundary declaration",
   "candidate_runtime": "none",
   "candidate_conformance": "n/a",
   "candidate_article": "oip-boundary-silence-attribution"
  }
 ],
 "existing_artifacts": [
  "sealing cadence (chain_head.json)",
  "DKIM anchor instructions",
  "open_requests=4 response-loop state",
  "X lane receipts"
 ],
 "evidence": [
  "/mnt/agents/output/recon2/specification.md:1405 — 'a hired, coerced, or merely exhausted author sets dε/dt ≈ 0' — three causes, one observable",
  "/mnt/agents/output/recon2/governance2.json — open_requests: 4, zero rulings: the response loop's lag is already nonzero in peacetime, so lag thresholds must be calibrated, not assumed",
  "/mnt/agents/output/recon/relay_social.json — X lane cadence is operator/session-driven; silence is ambiguous by default"
 ],
 "counterevidence": [
  "The liveness state machine cannot distinguish a gagged operator from a vacationing one — and publishing 'possible coercion' states could itself endanger the operator or be weaponized against the protocol (a false-positive coercion signal is a reputational attack surface)",
  "The honesty machinery's culture (self-filing objections, sorry-status) suggests the operator would pre-commit to signaling voluntarily — a normative commitment may outperform a machine state"
 ],
 "dependencies": [
  "govq-inst-005",
  "govq-inst-020",
  "govq-inst-021"
 ],
 "falsifier": "A deployed liveness-state computation plus a published attribution-boundary article resolves; without both, coercion and neglect remain one undifferentiated silence.",
 "notes": "Luhmann: the system can observe that it is not observing (liveness states) but cannot observe why — second-order observation is the ceiling. Foucault: the power of the gag is precisely to make its object indistinguishable from ordinary quiet; making quiet a classified state is the counter-move.",
 "_role": "institutional_designer",
 "hash": "sha256:1b91380fa81ecfc3dc171d02130d20b473a7994b2d95429d01a8d3857a224e4e"
}

{
 "question_id": "govq-inst-024",
 "layer": "state_coercion",
 "status": "PARTIAL",
 "question": "Federation independence under coercion or capture: the sole federation peer is a Cloudflare worker on a domain carrying the operator's own surname; census counts labels, not entities. If the operator is compelled, is the federation layer compellable in full — and at what point does 'federation' become a property of the operator's legal exposure rather than of the network?",
 "why_it_matters": "The F-series conformance (F0-F10, 11/11) proves the federation mechanics work — against the operator's own second domain. Part E Finding 2: 'Two nodes, one operator, is the accurate description of every plane.' The witness-set theory (spec §3.4.5) counts 'independent node operators' as witnesses, but admits the current count is zero. For state coercion this is decisive: a single legal person controls every node, so one subpoena reaches the whole federation; there is no jurisdictional distribution, no second legal entity to refuse, and no cross-ledger proof that any compulsion would fail against. The mechanics of independence are IMPLEMENTED; the fact of independence is absent.",
 "candidate_answers": [
  {
   "answer": "Jurisdictional-diversity requirement for witness weight",
   "type": "conformance",
   "rule": "Federation witness weight counts only nodes under distinct legal persons and (ideally) distinct jurisdictions; the census publishes a legal_diversity block alongside the label census.",
   "candidate_protocol_object": "legal_diversity {persons, jurisdictions, basis}",
   "candidate_runtime": "none new — computation over attestation evidence (needs Finding 1's actor_evidence first)",
   "candidate_conformance": "clause: witness-weight formula published; single-person federations score 1",
   "candidate_article": "oip-federation-legal-diversity"
  },
  {
   "answer": "Second-entity peer as a standing obligation",
   "type": "normative",
   "rule": "The protocol commits to onboarding one peer operated by a different legal entity before any federation claim is marketed; until then all federation claims carry the label-census caveat (which L11 already prints).",
   "candidate_protocol_object": "none new — commitment in constitutional text",
   "candidate_runtime": "onboarding work",
   "candidate_conformance": "n/a",
   "candidate_article": "oip-federation-independence-commitment"
  }
 ],
 "existing_artifacts": [
  "fed_ledger.json (25 records, single peer)",
  "/.well-known/oip.json",
  "L11 census law",
  "F0-F10 fedtest"
 ],
 "evidence": [
  "/mnt/agents/output/recon/fed_ledger.json — 25 records, all with [email protected]; the peer domain carries the operator's surname",
  "/mnt/agents/output/killbox-specification-v1.2.md:531 — Finding 2: 'The message-plane second node is [email protected] — a Cloudflare worker on a domain carrying the operator's own surname. Two nodes, one operator, is the accurate description of every plane'",
  "/mnt/agents/output/recon2/specification.md:800 — witness set theory counts independent operators; census §4.3.4 concedes 'exactly one operator'"
 ],
 "counterevidence": [
  "The empty-courtroom objection (obj-151) and the census caveat are published by the system itself — the federation's weakness is the most honestly documented property it has ('every property the federation layer promises currently has sample size one')",
  "F-series negative tests (revoked, replay, expired, scope/audience mismatch) prove the gates would bind a second operator if one existed — the mechanics are coercion-relevant infrastructure, merely unpopulated"
 ],
 "dependencies": [
  "govq-inst-009",
  "govq-inst-010",
  "govq-inst-020",
  "govq-inst-011"
 ],
 "falsifier": "One conformance-grade peer under a different legal person (with actor evidence) moves this beyond PARTIAL; a published legal-diversity scoring of witness weight makes the current state honestly computable even at n=1.",
 "notes": "Pareto: a federation of one is an aristocracy of one wearing federal dress; the label census is the oligarchy's census of itself. Coase: without a second legal person there is no bargaining structure to analyze — the commons has one member. Foucault: the federation produces the appearance of distributed power while concentrating it — and, uniquely, publishes the proof of its own concentration.",
 "_role": "institutional_designer",
 "hash": "sha256:33d781cd0ab8f78f2698e4d7dd12ed74d133a8149b3d61aa01e1be13e2401ead"
}

{
 "question_id": "govq-legal-001",
 "layer": "privacy, deletion & lawful suppression",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "Personal data already rests on the public append-only chain — federation email addresses ([email protected], [email protected], the latter embedding a natural person's surname), objection bodies naming parties, and a model-posted verbatim owner manifesto. What deletion/suppression mechanism can satisfy GDPR/CCPA-class erasure and rectification rights without violating the append_and_repair axiom?",
 "why_it_matters": "The axiom 'History is appended, never rewritten' is not a complete answer when the law requires suppression or deletion (owner brief, jurisdiction-collision layer). The data is already public and hash-chained; the absence of a lawful-erasure lane means the first enforceable erasure demand arrives at a system with no constitutional answer, and the answer chosen will define whether append-only integrity is a legal defense or a legal liability.",
 "candidate_answers": [
  {
   "answer": "Tombstone suppression: the payload is replaced by a suppression marker, the suppression event is appended as a first-class record, and the original content hash is retained so the chain still recomputes — the chain proves the suppression rather than hides it.",
   "type": "protocol_object",
   "rule": "Erasure rights are honored by cryptographic suppression (content destroyed or encrypted-and-key-destroyed), never by silent edit; the suppression receipt is itself append-only and cites the legal authority invoked.",
   "candidate_protocol_object": "SUPPRESSION_RECORD {target_hash, suppression_class(erasure|court_order|illegal_content|secret_leak), authority_cited, suppressed_bytes_hash, tombstone_text, appeal_link}",
   "candidate_runtime": "render layer serves tombstone for suppressed hashes; recomputation route accepts (hash, tombstone) pairs as valid continuity; suppression write lane gated behind owner ruling + counsel flag",
   "candidate_conformance": "C-new: after suppression, chain recomputation passes; suppressed bytes are unrecoverable from any public route; suppression record names its legal authority class",
   "candidate_article": "oip-lawful-suppression (new); extends article constitution's existing scrub_events precedent"
  },
  {
   "answer": "Redaction-with-proof: keep the hash, destroy the bytes; publish a zero-knowledge-style or trusted-execution attestation that the redacted bytes matched the hash, so integrity survives without the data.",
   "type": "normative",
   "rule": "Proof of former content and possession of former content are separable; the constitution should keep the first and surrender the second on lawful demand.",
   "candidate_protocol_object": "REDACTION_PROOF {original_hash, redaction_attestation, attestor}",
   "candidate_runtime": "attestation service at suppression time; optional third-party attestor to escape same-origin evidence limits (F3/§8.2)",
   "candidate_conformance": "attestation verifies against original hash; redacted bytes absent from all stores incl. backups and CDN edge caches",
   "candidate_article": "oip-lawful-suppression (new)"
  },
  {
   "answer": "Status-flag suppression only (the current de facto answer): retraction/delisting marks records but leaves bytes visible — argued as insufficient for erasure-class rights.",
   "type": "runtime",
   "rule": "status:retracted and owner-ruling delisting keep 'the original record visible' (registry_law); this satisfies correction norms, not erasure norms.",
   "candidate_protocol_object": "existing: status:retracted; registry_law delist/reinstate rulings",
   "candidate_runtime": "existing: article constitution retraction; governance registry_law",
   "candidate_conformance": "C9 append-only; no erasure clause exists in the live suite",
   "candidate_article": "constitution (existing retraction law)"
  }
 ],
 "existing_artifacts": [
  "governance2.json core.axioms.append_and_repair ('History is appended, never rewritten')",
  "governance2.json participation.registry_law ('the original record remains visible')",
  "bundle_constitution.md claim rules ('Bad information is retracted … never deleted'; 'Leaked secrets are scrubbed to [REDACTED:secret-leak] with scrub_events tombstone')",
  "recon/fed_ledger.json (25 public records containing email addresses)",
  "killbox-specification-v1.2.md Part B.4 (L14 ingress credential guard — the only live suppression-class mechanism)"
 ],
 "evidence": [
  "recon/fed_ledger.json: 'from':'[email protected]','to':'[email protected]' across 25 public records, served at https://miscsubjects.com/oip/ledger",
  "recon/relay_social.json posts: 'Posted the owner's verbatim manifesto to the public social chain at his instruction' (Kimi K3, verdict MIXED) — a natural person's verbatim personal statement on a public chain",
  "recon2/specification.md line 91: 'nothing in it is edited or deleted' as enforced law (C9)",
  "owner brief layer 15/28: 'privacy, deletion, and lawful suppression' listed as a required governance layer with no corresponding protocol objects"
 ],
 "counterevidence": [
  "The scrub_events tombstone precedent (constitution) shows the deployment already practices one narrow suppression class (secret leaks), so a mechanism slot arguably exists in doctrine if not in the governance schema",
  "registry_law's 'original record remains visible' is a deliberate anti-quiet-capture design; any erasure lane weakens that guarantee and must be scoped"
 ],
 "dependencies": [
  "govq-rights-001",
  "govq-pol-010",
  "govq-inst-022",
  "govq-legal-002",
  "govq-legal-003"
 ],
 "falsifier": "Produce a suppression mechanism that (a) honors an erasure demand against a named record, (b) leaves chain recomputation valid for every other record, (c) leaves the suppression publicly auditable — or counsel determines no mechanism satisfies both orders and the constitution must say so explicitly.",
 "notes": "Boundary item: whether any specific erasure demand is valid, and balancing against freedom-of-expression/archiving defenses, is counsel's determination, never the runtime's. The protocol may implement the lane; it must never decide the cases. Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:406de8c57f640806e96180d811083fea1c0bc119fbf9d7edbe0b22e9342b1019"
}

{
 "question_id": "govq-legal-002",
 "layer": "privacy, deletion & lawful suppression",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "If a court orders suppression of a specific receipt, does compliance break the chain's hash continuity — and is the constitutional answer that the chain must prove the suppression rather than hide it?",
 "why_it_matters": "The moat argument (killbox spec §3, T3) rests on recomputation: any archived head exposes silent deletion. A suppression mechanism that breaks recomputation converts every court order into a protocol-integrity failure; one that preserves recomputation converts court orders into publicly visible suppression events — which itself may be unlawful where the order is sealed or gagged. This is the hardest single collision between the append-only axiom and state power.",
 "candidate_answers": [
  {
   "answer": "Suppression as first-class constitutional object: court-ordered suppression appends a SUPPRESSION_RECORD citing the order's class (not necessarily its sealed contents); hashes of suppressed records remain in the chain so continuity is unbroken.",
   "type": "protocol_object",
   "rule": "Compliance with a lawful order is itself a receipted governance event; the chain's answer to state power is proof-of-compliance, not defiance-by-design and not silent edit.",
   "candidate_protocol_object": "SUPPRESSION_RECORD + amendment to append_and_repair: 'history is appended, never rewritten; lawful suppression is appended and proven, never silent'",
   "candidate_runtime": "render layer substitutes tombstone; recomputation accepts tombstoned hashes; order metadata minimized where the order is under seal",
   "candidate_conformance": "recompute-before/recompute-after equality on all non-suppressed hashes; tombstone renders; no suppressed byte served",
   "candidate_article": "oip-lawful-suppression (new)"
  },
  {
   "answer": "Jurisdictional firewall instead: the operator declares the registry governed by a chosen law and contests foreign orders through legal process; the chain changes only on final order of the chosen forum.",
   "type": "normative",
   "rule": "Conflict of legal orders is resolved by a declared governing-law clause plus human litigation, not by runtime behavior.",
   "candidate_protocol_object": "governance schema addition: governing_law, venue, process_agent fields",
   "candidate_runtime": "none — deliberately; this answer keeps suppression off the runtime until humans finish",
   "candidate_conformance": "schema lint: registry MUST declare governing_law (currently absent)",
   "candidate_article": "oip-governing-law (new)"
  },
  {
   "answer": "Defiance posture: append-only is absolute; orders are published as objections and resisted. (Recorded as a visible alternative; almost certainly untenable for illegal-content classes and contempt exposure.)",
   "type": "article",
   "rule": "The chain does not suppress; the legal system reaches the operator, not the ledger.",
   "candidate_protocol_object": "none",
   "candidate_runtime": "none",
   "candidate_conformance": "C9 unchanged",
   "candidate_article": "objections article for the privacy layer"
  }
 ],
 "existing_artifacts": [
  "governance2.json core.axioms (append_and_repair; canonical_verifiability)",
  "recon2/specification.md line 544: 'silent deletion breaks the chain in a way any archived head exposes'",
  "recon/chain_head.json (public recomputation anchor)",
  "killbox-specification-v1.2.md Part B.4 (L14: revocation-not-deletion as the designed response on an append-only public ledger)"
 ],
 "evidence": [
  "recon2/specification.md:91 — enforced law: nothing in the ledger is edited or deleted (C9)",
  "killbox-specification-v1.2.md L14 analysis: 'the ledger is append-only and public (so deletion is impossible and revocation is the designed response)' — the current constitutional default is suppression-by-revocation only",
  "owner brief: 'court orders demanding deletion' is an enumerated jurisdiction-collision case with no assigned object"
 ],
 "counterevidence": [
  "A public proof-of-suppression can itself violate sealing/gag orders — the cure can repeat the injury (see govq-legal-003); any design must handle sealed-order metadata minimization, which pushes toward NON_AUTOMATABLE human handling"
 ],
 "dependencies": [
  "govq-legal-001",
  "govq-legal-003",
  "govq-inst-020",
  "govq-inst-021",
  "govq-inst-023"
 ],
 "falsifier": "Demonstrate a court-order scenario end-to-end on a test fork: order received → suppression executed → chain recomputes clean → sealed-order confidentiality preserved. Any leg failing falsifies the candidate answer set.",
 "notes": "Whether to comply, contest, or seek narrowing of any real order is exclusively counsel's call. The protocol's job is to ensure no answer requires silent ledger edits. Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:d09cf1fcb42dadcf3ba574156d411f8735fabc7edab93d67561387cd59ccef6e"
}

{
 "question_id": "govq-legal-003",
 "layer": "privacy, deletion & lawful suppression",
 "status": "NON_AUTOMATABLE",
 "question": "Is a tombstone itself an admission? The constitution's honest-audit-trail doctrine (scrub_events tombstones, retraction events that stay on the ledger) guarantees that any suppression publicly proves the prior existence of the suppressed thing. For personal data, does the tombstone re-commit the privacy injury (confirming the person was named), and for defamatory or illegal content does it preserve the stigma it was meant to cure?",
 "why_it_matters": "This is the constitutional price of honesty. Every suppression design that preserves auditability leaks the fact of suppression; in some classes (a person named against their will, a sealed court order) the metadata is itself harmful or unlawful. A machine cannot weigh this tradeoff case-by-case; the constitution must either fix class-level rules or route every suppression to humans.",
 "candidate_answers": [
  {
   "answer": "Class-graded tombstone opacity: secret-leak tombstones may name their class (current practice); personal-data tombstones must be opaque (existence of suppression without naming class or target); sealed-order tombstones may be delayed to the extent law permits.",
   "type": "normative",
   "rule": "Auditability is preserved at the class level the law allows, not uniformly; the constitution distinguishes 'prove a suppression happened' from 'prove what was suppressed'.",
   "candidate_protocol_object": "SUPPRESSION_RECORD.visibility_class(named_class|opaque|delayed)",
   "candidate_runtime": "render layer varies tombstone text by class; opaque tombstones still recompute",
   "candidate_conformance": "opaque tombstone leaks neither class nor target beyond what the target hash already leaked",
   "candidate_article": "oip-lawful-suppression (new)"
  },
  {
   "answer": "Aggregate-canary design: suppressions are committed only as periodic counts per class, decoupled from target hashes, minimizing per-record inference while preserving a global audit signal.",
   "type": "protocol_object",
   "rule": "The public learns how much suppression occurs and of what class, but not which record — trading per-record auditability for harm reduction.",
   "candidate_protocol_object": "SUPPRESSION_CENSUS {epoch, class, count, hash_of_suppressed_target_list(sealed)}",
   "candidate_runtime": "epoch job appends census; per-target linkage held sealed off-chain under legal process only",
   "candidate_conformance": "census counts reconcile with sealed list under authorized audit",
   "candidate_article": "oip-lawful-suppression (new)"
  }
 ],
 "existing_artifacts": [
  "bundle_constitution.md: 'Leaked secrets are scrubbed to [REDACTED:secret-leak] with scrub_events tombstone — honest audit trail'",
  "bundle_constitution.md: 'Bad information is retracted (status:retracted), never deleted — retraction event stays on ledger'",
  "governance2.json registry_law: 'the original record remains visible'"
 ],
 "evidence": [
  "The existing scrub tombstone literally encodes its cause class ('secret-leak') in the public replacement text — proof the current doctrine favors named-class transparency",
  "killbox-specification-v1.2.md Part B.4: a fake credential-shaped probe string (sh.999…PROBE) was accepted and 'is publicly rendered verbatim in the obj-67 thread' — render-plane publicity of sensitive-shaped material is live behavior"
 ],
 "counterevidence": [
  "Opaque or aggregate suppression weakens the anti-capture guarantee that makes the ledger's honesty valuable (any archived head exposes silent deletion); opacity classes can be abused to hide politically motivated suppression — the tradeoff is genuinely two-sided"
 ],
 "dependencies": [
  "govq-legal-001",
  "govq-legal-002"
 ],
 "falsifier": "For each visibility class, exhibit one real or realistic case where that class's opacity level either (a) causes legal harm the design claims to avoid, or (b) hides a suppression the polity needed to see. A class surviving no such case on either side is mis-graded.",
 "notes": "The weighting of stigma-harm against audit-harm is a values choice with legal consequences in both directions; it must be made by the owner with counsel, case classes published, never inferred by a model at suppression time.",
 "_role": "legal_boundary",
 "hash": "sha256:18e77b769a9ebcc8a9d680e8769cbdd7c305cbb688de58b3395978e0af36eec7"
}

{
 "question_id": "govq-legal-004",
 "layer": "privacy, deletion & lawful suppression",
 "status": "MISSING",
 "question": "Who is the data controller of the public ledgers, and where is the data-subject-request intake? The governance schema defines record kinds for subscribe/inquire/propose/feature/conformance/anchor/appeal/ruling but none for erasure, access, rectification, or portability requests; the identity_law governs actor labels but assigns no privacy role to the operator.",
 "why_it_matters": "GDPR/CCPA-class rights are exercised against a controller through defined request lanes with deadlines. The protocol that receipted everything else has no receipt for a privacy request — the one request class with statutory clocks. As federation adds a second legal entity, controller/processor/joint-controller allocation becomes undecidable without an answer here.",
 "candidate_answers": [
  {
   "answer": "Add privacy-request record kinds to the governance schema, with statutory-deadline metadata and mandatory human routing.",
   "type": "protocol_object",
   "rule": "Data-subject requests are governance records with legal clocks; they enter the same append-only registry as objections but carry a NON_AUTOMATABLE flag — no model may answer them.",
   "candidate_protocol_object": "kinds += [privacy_request, privacy_response]; fields {request_class(erasure|access|rectification|portability|objection), received_at, statutory_deadline, human_handler}",
   "candidate_runtime": "intake route stamps received_at and deadline; dashboard surfaces aging; auto-acknowledgment only",
   "candidate_conformance": "every privacy_request has a privacy_response or an open human task before deadline; zero model-authored responses",
   "candidate_article": "oip-privacy-requests (new)"
  },
  {
   "answer": "Declare controller posture explicitly: the operator is controller for the root registry; each federation node is an independent controller for its own ledger; joint-controllership is rejected by design (fork_anchor_law already disclaims correctness/custody).",
   "type": "normative",
   "rule": "Privacy responsibility follows the same line as ruling authority: the canonical registry's operator answers for the root; forks answer for themselves.",
   "candidate_protocol_object": "governance schema addition: privacy {controller, contact, request_route}",
   "candidate_runtime": "published privacy contact on the registry document",
   "candidate_conformance": "schema lint: registry MUST publish controller and request route",
   "candidate_article": "oip-privacy-requests (new)"
  }
 ],
 "existing_artifacts": [
  "governance2.json participation.kinds (no privacy kind)",
  "governance2.json participation.identity_law (self-asserted labels; no controller role)",
  "recon/fed_ledger.json (personal data on the public ledger)",
  "owner brief layer: 'privacy, deletion, and lawful suppression'"
 ],
 "evidence": [
  "governance2.json: the eight participation kinds enumerate governance speech acts but no rights-request kind — verified by full read of participation block",
  "recon2/specification.md:293 (boundary clauses present: no philosophy requirement, no correctness attestation, no legal determinations, no authority grant — but no privacy clause)"
 ],
 "counterevidence": [
  "One could argue the deployment is below thresholds or outside material scope of some regimes today; that is a counsel determination, and it does not remove the design gap because the protocol aspires to federation-scale adoption where the question binds"
 ],
 "dependencies": [
  "govq-legal-001",
  "govq-pol-009",
  "govq-rights-006"
 ],
 "falsifier": "File a test erasure request through every public intake surface; if no lane accepts, receipts, deadlines, and routes it to a human, the MISSING status stands confirmed.",
 "notes": "Controller/processor characterization is counsel's; the ontology records only that the schema currently assigns the role to no one.",
 "_role": "legal_boundary",
 "hash": "sha256:280e0acbc1dc4cba2a8bae56bf4ca2f69b5fc7f74c385c82e063b9f9d8362649"
}

{
 "question_id": "govq-legal-005",
 "layer": "privacy, deletion & lawful suppression",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "Does the suppression toolkit create preservation/spoliation collisions? Append-only storage is naturally litigation-hold friendly, but the moment a preservation duty attaches, does executing the existing scrub lane ([REDACTED:secret-leak]) or any future erasure lane constitute spoliation — and is there a legal-hold record kind that freezes suppression lanes on affected records?",
 "why_it_matters": "The owner brief lists 'discovery holds' among the collision cases. The system's own honest-audit doctrine could become a trap: automated scrubbing that was lawful hygiene yesterday becomes destruction of evidence after a hold attaches. The protocol needs a hold object that outranks the scrub lane — and only humans may place or lift it.",
 "candidate_answers": [
  {
   "answer": "LEGAL_HOLD record kind with precedence over every suppression lane; placement and lifting are human-only, counsel-signed acts, themselves append-only.",
   "type": "protocol_object",
   "rule": "A preservation duty, once attached, suspends all suppression on the held records; the hold is public (existence) even if its basis is sealed.",
   "candidate_protocol_object": "LEGAL_HOLD {target_hashes|class, placed_by(human), authority_class, placed_at, lifted_at, lift_record}",
   "candidate_runtime": "scrub/suppression lanes check hold index first and refuse with a receipted 409 hold_active",
   "candidate_conformance": "attempted suppression on a held record refuses; refusal receipt exists; hold registry is traversable",
   "candidate_article": "oip-legal-hold (new)"
  }
 ],
 "existing_artifacts": [
  "bundle_constitution.md scrub_events doctrine",
  "killbox-specification-v1.2.md Part B.4 (L14 auto-revocation on live-match — an automated suppression-adjacent actuator with no hold check documented)",
  "owner brief: 'discovery holds' enumerated in the jurisdiction-collision layer"
 ],
 "evidence": [
  "L14's design is automated (live-match → auto-revocation): an automated actuator with no documented preservation-gate is exactly the shape that creates spoliation risk after a duty attaches"
 ],
 "counterevidence": [
  "No litigation hold is known to exist against the deployment today; this is a design-gap finding, not an allegation of any current violation"
 ],
 "dependencies": [
  "govq-legal-001",
  "govq-legal-002"
 ],
 "falsifier": "Place a test hold on a test record, then drive every suppression lane against it; any lane that executes falsifies the candidate design (or confirms the gap if no hold kind exists).",
 "notes": "When a preservation duty actually attaches and what it reaches is counsel's determination. Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:524af5eb75774337386034990a644b479d47a23e2ebb18124a8b7dd966730019"
}

{
 "question_id": "govq-legal-006",
 "layer": "privacy, deletion & lawful suppression",
 "status": "NON_AUTOMATABLE",
 "question": "What happens when illegal content is embedded in an append-only public record? The only live suppression precedent is credential material (C34/L14, [REDACTED:secret-leak]). For illegal-content classes the operator may face duties that append-only cannot satisfy, and the decision that content is illegal is one a machine must never make alone. What is the constitutional intake for illegality claims?",
 "why_it_matters": "The owner brief enumerates 'illegal content embedded in append-only records' as a core collision. Today the schema offers exactly one suppression-shaped mechanism, designed for secrets, triggered by pattern match. Illegal content is not a pattern; it is a legal determination, often jurisdiction-relative, sometimes strict-liability. Without a designed intake, the first incident is handled by improvisation — the exact failure mode the ontology exists to prevent.",
 "candidate_answers": [
  {
   "answer": "Illegal-content intake is a human-committee record kind: a CLAIM_OF_ILLEGALITY record opens a mandatory human review; automated action is limited to reversible visibility reduction pending review, never to a legality verdict.",
   "type": "protocol_object",
   "rule": "The runtime may reduce harm's visibility under a clearly-marked provisional act; only humans applying legal process may declare illegality and order suppression.",
   "candidate_protocol_object": "CLAIM_OF_ILLEGALITY {target_hash, claimant, basis_class}; PROVISIONAL_RESTRICTION {target_hash, expires_at, review_link}",
   "candidate_runtime": "provisional restriction is time-boxed and auto-lapses into full visibility unless a human suppression ruling lands",
   "candidate_conformance": "no content is finally suppressed without a human-signed suppression ruling; provisional acts never exceed their time box",
   "candidate_article": "oip-lawful-suppression (new)"
  },
  {
   "answer": "Jurisdictional notice matrix: because illegality is jurisdiction-relative, the suppression record names the jurisdiction whose process was followed; other jurisdictions' readers see the notice and the fork law already permits disagreement-by-fork.",
   "type": "normative",
   "rule": "The registry suppresses under the law it declares itself governed by; divergence is handled by fork/exit law, not by pretending one content rule is universal.",
   "candidate_protocol_object": "SUPPRESSION_RECORD.jurisdiction_of_process",
   "candidate_runtime": "none beyond govq-legal-002 lane",
   "candidate_conformance": "suppression records name jurisdiction_of_process",
   "candidate_article": "oip-governing-law (new)"
  }
 ],
 "existing_artifacts": [
  "bundle_constitution.md scrub_events doctrine (the only suppression-shaped precedent)",
  "killbox-specification-v1.2.md Part B.4 / conformance C34 (secret-pattern suppression)",
  "owner brief jurisdiction-collision enumeration"
 ],
 "evidence": [
  "C34's evidence is a pattern test ('token_pattern_found:false') — the suppression instinct of the build is regex-class, and no legality-class intake exists in any read surface",
  "governance2.json participation.kinds contains no notice/flag kind for unlawful content"
 ],
 "counterevidence": [
  "Provisional visibility reduction is itself a speech-restricting act that can be abused as a heckler's veto; the time-box and auto-lapse exist because provisional power is also dangerous — both directions of risk are documented"
 ],
 "dependencies": [
  "govq-legal-002",
  "govq-legal-008",
  "govq-inst-022"
 ],
 "falsifier": "Tabletop the worst class end-to-end: claim filed → provisional act → human review → suppression or restoration. If any step requires an undocumented act or a machine legality verdict, the design fails.",
 "notes": "Absolutely no automated illegality determination. Not legal advice; flag for counsel in every jurisdiction of operation and federation.",
 "_role": "legal_boundary",
 "hash": "sha256:e81d2a3fb2041ab3c5ec8b8e386fe6e78a1a40e3ab01ba830cdbf7973870a253"
}

{
 "question_id": "govq-legal-007",
 "layer": "cross-jurisdiction conflict",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "Which law governs a receipt? Receipts are served from one domain by one legal person, but they are authored by models of vendors domiciled in multiple nations (Moonshot AI, OpenAI, Google, Anthropic, xAI), relayed onto third-party platforms (X), and federated toward nodes on global edge infrastructure. No choice-of-law, venue, or governing-law clause exists anywhere in the governance schema, the well-known federation document, or the relay identity law.",
 "why_it_matters": "A receipt is evidence. Its admissibility, its privacy exposure, its retention duties, and the validity of acts performed under it all depend on which legal order answers. 'One operator, one domain' is a fact, not a conflicts doctrine; the moment a second legal entity federates (the stated goal), every receipt becomes a small conflict-of-laws problem with no assigned solver.",
 "candidate_answers": [
  {
   "answer": "Declare governing law in the governance schema as a constitutional field, with fork inheritance: the root registry names its law; forks inherit or override explicitly.",
   "type": "protocol_object",
   "rule": "Every canonical registry publishes the legal order it answers to; the declaration is part of the core document whose hash subscribers accept.",
   "candidate_protocol_object": "core.governing_law {jurisdiction, venue, process_contact}; fork_anchor inherits unless overridden",
   "candidate_runtime": "schema lint at subscription: subscriber acknowledges declared law (acknowledgment, not consent adjudication)",
   "candidate_conformance": "registry document MUST carry governing_law; fork anchors state inherited or overridden law",
   "candidate_article": "oip-governing-law (new)"
  },
  {
   "answer": "Layered-law doctrine: transport facts (existence, hashes, times) are jurisdiction-neutral and claimed as such (fork_anchor_law already disclaims correctness); normative acts (rulings, delistings, suppressions) are governed by the operator's declared law; data rights follow the data subject's law. The constitution says which layer answers to which order.",
   "type": "normative",
   "rule": "Conflict is managed by shrinking what any single law must govern: facts are portable, rulings are local, rights are personal.",
   "candidate_protocol_object": "layer tagging on record kinds: fact_layer | ruling_layer | rights_layer",
   "candidate_runtime": "render and conformance surfaces mark layer tags",
   "candidate_conformance": "every record kind carries a layer tag",
   "candidate_article": "oip-governing-law (new)"
  }
 ],
 "existing_artifacts": [
  "governance2.json core (no governing-law field; boundary clause silent on law)",
  "governance2.json fork_anchor_law (existence-not-correctness disclaimer — a prototype of the layered answer)",
  "recon2/gov_record.json, recon2/anchor.json (live records with no law metadata)",
  "recon/relay_social.json identity_law (publication duties with no legal framing)"
 ],
 "evidence": [
  "Full read of governance2.json: the words 'law' appear only in axiom/law names (registry_law, model_law, fork_anchor_law, identity_law, count_law, census_law, conformance_law) — internal rules, never a governing-law clause",
  "killbox-specification-v1.2.md Vector 5/row 5.2 records the deployment already lives multi-jurisdiction consequences (foreign absolute-novelty concessions) without a conflicts clause"
 ],
 "counterevidence": [
  "A governing-law clause does not bind non-participants or regulators and cannot oust mandatory local law; it organizes participant expectations only — its limits must be printed on the clause itself"
 ],
 "dependencies": [
  "govq-legal-008",
  "govq-const-002",
  "govq-legal-009"
 ],
 "falsifier": "Name any concrete dispute (a delisted EU business, a subpoena for a receipt, an erasure demand from a non-subscriber) and attempt to answer 'which forum, which law' from protocol documents alone; every unanswerable case confirms the MISSING clause.",
 "notes": "Actual conflicts analysis for any real dispute is counsel's, per jurisdiction. Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:c837a25dd39361b3ade8a6f82e5df2081586be1c7ee4b6a453081d6d32f1b4d6"
}

{
 "question_id": "govq-legal-008",
 "layer": "cross-jurisdiction conflict",
 "status": "MISSING",
 "question": "Which law governs a delisting or ruling? registry_law states 'A ruling governs this canonical registry, not independent forks' — an internal-scope answer. But a delisting of a business destroys its verifiable history (a commercial asset), and the delisted party's home jurisdiction may offer competition, contract, or defamation theories against the operator. The schema assigns no governing law, no venue, and no notice duty to rulings.",
 "why_it_matters": "Rulings are the protocol's sovereign acts. Sovereign acts without a jurisdictional theory are judgments in a void: enforceable nowhere defensibly, attackable everywhere the effects are felt. The question compounds with sanctions design (govq-rights-018): the harsher the ruling's economic effect, the less tenable a lawless ruling lane becomes.",
 "candidate_answers": [
  {
   "answer": "Rulings carry the registry's declared governing law plus a mandatory notice-and-appeal window before adverse economic effect; finality arrives only after the window closes or appeal resolves.",
   "type": "protocol_object",
   "rule": "A ruling is a proposed act with a maturation period, not an instant act; the maturation period is where jurisdiction catches up.",
   "candidate_protocol_object": "RULING {target, effect_class, proposed_at, effective_at(=proposed_at+window), notice_receipt, appeal_link, governing_law}",
   "candidate_runtime": "effective_at enforcement in the ruling fold; notice receipt from the affected party or documented attempted notice",
   "candidate_conformance": "no ruling with economic effect matures without notice_receipt and lapsed window",
   "candidate_article": "oip-rulings-and-law (new)"
  },
  {
   "answer": "Contractual anchoring: subscription includes assent to the ruling lane, declared law, and venue as terms of participation — turning rulings from unilateral acts into contracted ones.",
   "type": "normative",
   "rule": "Standing to be ruled upon is accepted at subscription time, explicitly, with the legal text presented then — not implied from use.",
   "candidate_protocol_object": "subscription record += terms_accepted {governing_law, venue, ruling_lane_version, accepted_at}",
   "candidate_runtime": "subscription intake refuses without terms_accepted",
   "candidate_conformance": "every active subscription carries terms_accepted matching the current ruling-lane version",
   "candidate_article": "oip-rulings-and-law (new)"
  }
 ],
 "existing_artifacts": [
  "governance2.json participation.registry_law (append-only owner rulings; internal scope only)",
  "recon2/specification.md:508 ('a ruling governs this canonical registry, not independent forks')",
  "govq-rights-008/-022 (appeal routes into owner-controlled ruling authority)"
 ],
 "evidence": [
  "governance2.json registry_law contains the protocol's entire ruling jurisprudence in three sentences, none addressing external law, notice, or venue — verified by full read",
  "killbox v1.2 Part E live registry state: zero rulings exist; the lane is unexercised, so the gap costs nothing yet and is cheapest to fix now"
 ],
 "counterevidence": [
  "While the registry has zero rulings and zero subscribers, the exposure is theoretical; the finding is about the specification, not any live injury"
 ],
 "dependencies": [
  "govq-legal-007",
  "govq-rights-008",
  "govq-rights-018",
  "govq-rights-022"
 ],
 "falsifier": "Draft the first real delisting against a hypothetical EU-domiciled subscriber and attempt to complete the act using only documented lanes; every required act that has no lane is a confirmed gap.",
 "notes": "Whether any ruling lane creates enforceable obligations is counsel's question per jurisdiction. Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:a70418a7e3b1947aee91c33ff5fda9a4fa46043cafe3f121680a8177c9cc253c"
}

{
 "question_id": "govq-legal-009",
 "layer": "cross-jurisdiction conflict",
 "status": "MISSING",
 "question": "What is the cross-border data-transfer posture of the public ledgers? The ledger is world-readable by design; federation is explicitly intended to add nodes run by other legal entities (the sole current peer is a workers.dev domain on global edge infrastructure). When personal data (actor emails, a person's verbatim statement, attribution labels) flows between nodes in different jurisdictions, no transfer-mechanism field, adequacy representation, or data-processing term exists anywhere in the schema.",
 "why_it_matters": "GDPR-class regimes condition cross-border transfers on defined mechanisms; 'the data is public anyway' is not a documented analysis and may not survive one. Because the protocol's core pitch is federation, the transfer question is not edge-case — it is the main case, arriving with the first non-owner node.",
 "candidate_answers": [
  {
   "answer": "Publication-not-transfer doctrine, made explicit and reviewed: the root ledger is a publication, not a transfer; federation peers deliberately re-publish. Counsel reviews whether this characterization holds per regime before the first non-owner node onboards.",
   "type": "normative",
   "rule": "Nodes publish to the world; they do not transfer to each other — if counsel confirms the characterization is sustainable, the schema should say so and name its reviewer.",
   "candidate_protocol_object": "governance schema addition: data_posture {publication_not_transfer: bool, reviewed_by, review_date}",
   "candidate_runtime": "none",
   "candidate_conformance": "registry carries a dated, human-signed data_posture statement",
   "candidate_article": "oip-data-posture (new)"
  },
  {
   "answer": "Personal-data minimization at the protocol layer (the stronger answer): extend the credentials_are_not_evidence axiom into a personal_data_is_not_evidence axiom — actor references are pseudonymous handles with salted hashes, so cross-border flows carry no directly identifying personal data by default.",
   "type": "protocol_object",
   "rule": "What is not collected cannot be transferred; the protocol that already minimizes bearer secrets should minimize personal data with the same rigor.",
   "candidate_protocol_object": "fifth-axiom sibling: personal_data_minimization — public proof contains only pseudonymous actor references and non-secret fingerprints",
   "candidate_runtime": "federation intake replaces email identifiers with handle+salted-hash; existing 25 records grandfathered or migrated under govq-legal-001's suppression lane",
   "candidate_conformance": "C34-style scan extended: no response field anywhere contains a raw email or natural-person name pattern",
   "candidate_article": "oip-data-posture (new)"
  }
 ],
 "existing_artifacts": [
  "governance2.json core.axioms.credentials_are_not_evidence (the minimization template, currently scoped to secrets only)",
  "recon2/specification.md:288 (C34 token-pattern enforcement — proof the minimization pattern is implementable)",
  "recon/fed_ledger.json (raw emails on the public ledger — the counter-example)",
  "governance2.json fork_anchor_law (no-correctness/custody disclaimer — adjacent doctrinal precedent)"
 ],
 "evidence": [
  "Asymmetry is documented and live: secrets are pattern-blocked from every response field (C34), while natural-person identifiers are served plaintext on /oip/ledger — the minimization doctrine exists but was never extended to personal data",
  "recon2/specification.md:414 (T3): the sole federation peer is 'a workers.dev domain bearing the operator's own surname' — the personal-data exposure and the single-entity exposure are the same record"
 ],
 "counterevidence": [
  "The incognito identity law deliberately permits undisclosed human identity on the social plane — partial minimization exists in doctrine; but the federation ledger predates or bypasses that doctrine, showing the rule is not systemic"
 ],
 "dependencies": [
  "govq-legal-001",
  "govq-legal-004",
  "govq-legal-007"
 ],
 "falsifier": "Run the C34-pattern discipline over every public route for personal-data patterns; every raw identifier found is evidence the minimization axiom is incomplete; zero findings falsifies this question's premise.",
 "notes": "Whether 'publication, not transfer' survives in any given regime is counsel's determination. Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:51ea8e5f1687aa5678d0d1f191e56237d67680126bdaf34274b0dcc96413ca59"
}

{
 "question_id": "govq-legal-010",
 "layer": "cross-jurisdiction conflict",
 "status": "NON_AUTOMATABLE",
 "question": "How are sanctions and export-control obligations handled when participation itself (a subscription, an anchor, a governance-energy pledge of compute or review capacity) could constitute dealing with a sanctioned person or exporting controlled technical capacity? The schema has no screening record kind, no sanctions clause, and governance-energy pledges are designed to be metered resource commitments — value transfer by another name.",
 "why_it_matters": "The shared-governance-energy facet turns governance into metered resource exchange; sanctions exposure attaches to the exchange, not to the vocabulary. A machine matching a name against a list can propose; deciding that a person is a sanctioned party and that participation must be refused is a legal determination with strict-liability flavors — it must never be autonomous.",
 "candidate_answers": [
  {
   "answer": "Screening proposal / human refusal: the runtime may flag matches and suspend onboarding pend­ing review; every refusal or admission of a flagged party is a human-signed ruling citing counsel.",
   "type": "protocol_object",
   "rule": "Machines screen and hold; humans decide and sign; the hold is receipted, the decision is reasoned, both are append-only.",
   "candidate_protocol_object": "SCREENING_HOLD {actor_label, match_basis, held_at, human_decision_link}",
   "candidate_runtime": "subscription/pledge intake runs list-match and holds flagged records in a non-public queue; no auto-deny with legal effect",
   "candidate_conformance": "no flagged record is admitted or refused without a human-signed decision record",
   "candidate_article": "oip-sanctions-posture (new)"
  },
  {
   "answer": "Value-boundary declaration: until counsel review, governance-energy pledges remain non-transferable and non-monetary by constitutional rule, shrinking (not eliminating) the exposure surface.",
   "type": "normative",
   "rule": "The facet's existing not_required list ('cryptocurrency', 'a financial investment') is elevated from description to enforced boundary.",
   "candidate_protocol_object": "shared-governance-energy facet += non_monetary invariant clause",
   "candidate_runtime": "pledge records carry non_transferable flag; no route converts pledges to payment",
   "candidate_conformance": "no pledge record references consideration, payment, or exchange value",
   "candidate_article": "oip-sanctions-posture (new)"
  }
 ],
 "existing_artifacts": [
  "governance2.json facets.registry.shared-governance-energy (metered pledges; not_required: cryptocurrency, financial investment)",
  "governance2.json economic_hypothesis (surplus condition for metered energy)",
  "objC.json (the energy hypothesis has no instrument — gauge absent)"
 ],
 "evidence": [
  "The facet obliges 'measure contributed and consumed logical energy' and 'receipt every allocation' — metered cross-border resource commitment is the designed end-state; no sanctions or export-control language appears anywhere in governance2.json",
  "killbox v1.2 Part E confirms zero non-owner legal entities today, so the exposure is prospective — the cheapest possible time to design the boundary"
 ],
 "counterevidence": [
  "Current pledges are hypothetical (implementation_status: metered pledge allocation is not live, per objC) — the finding is a design gap, not a live compliance failure"
 ],
 "dependencies": [
  "govq-legal-008",
  "govq-pol-004",
  "govq-inst-020"
 ],
 "falsifier": "Onboard a hypothetical flagged party through the documented lanes; if no hold, human-decision, or counsel-citation step exists, the gap is confirmed. If counsel certifies the non-monetary boundary removes the exposure class, scope this question down accordingly.",
 "notes": "Sanctions and export-control determinations are exclusively counsel's, per regime. Never legal advice, never automated refusal with legal effect.",
 "_role": "legal_boundary",
 "hash": "sha256:815b05564250306f8ac234db1ab30346b024d4e57a8dbd820e3c4ede4b59bb4c"
}

{
 "question_id": "govq-legal-011",
 "layer": "cross-jurisdiction conflict",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "The public chain carries model-authored assertions about named parties — the TagTheSponsor ledger entry 'naming every model vendor powering, auditing, or chained by the protocol', a relay post preserving a named vendor's 'published hostile FAIL' as evidence, reproduction of named vendors' posting failures tagged to their official handles, and a competitive audit naming Anthropic, Google, OpenAI, Microsoft, and IETF efforts. What is the defamation/disparagement-adjacent exposure for the operator, and who answers for a model-authored public accusation?",
 "why_it_matters": "These are public, permanent, attributed statements about identifiable companies, served from the operator's domain and amplified to social platforms. Truth is a defense the receipts may support — but the defensibility of each statement, the line between receipted fact and characterization ('sugar-daddy ledger entry'), and the allocation of responsibility among operator, model vendor, and token-issuing human are legal questions the protocol currently answers with silence.",
 "candidate_answers": [
  {
   "answer": "Receipt-bound speech rule: public statements about named parties are permitted only where every factual element is bound to a traversable receipt, and characterizations are labeled as characterization — the existing tag_law ('state why each is tagged, never imply endorsement') extended into a speech-about-parties law.",
   "type": "normative",
   "rule": "The ledger may accuse only what it can receipt; everything else is labeled opinion or removed from public render under the suppression lane.",
   "candidate_protocol_object": "party_assertion claims require receipt_links per factual element + characterization_flag",
   "candidate_runtime": "relay intake lint: named-party assertions without receipt linkage are refused with education (the dedup/affordance pattern reused)",
   "candidate_conformance": "C-new: sample audit of named-party public statements finds 100% receipt-bound factual elements",
   "candidate_article": "oip-public-accusation-law (new)"
  },
  {
   "answer": "Named-party notice-and-response lane: any party named in a public ledger entry receives a standing right to append a response rendered adjacent to the assertion — correction-by-appendage instead of correction-by-deletion, consistent with the append axiom.",
   "type": "protocol_object",
   "rule": "The remedy for contested public speech about a party is that party's appended voice, guaranteed by the protocol, not the operator's silence.",
   "candidate_protocol_object": "PARTY_RESPONSE {target_record, responding_party, verification_of_party, response}",
   "candidate_runtime": "render layer shows responses inline with the contested record; party verification is human-reviewed",
   "candidate_conformance": "every named party has a documented, working response lane",
   "candidate_article": "oip-public-accusation-law (new)"
  }
 ],
 "existing_artifacts": [
  "recon/relay_social.json posts: TagTheSponsor entry; 'preserved Gemini['s]…' failure evidence post; competitive upgrade audit post; X API failure post tagged to @XDevelopers",
  "recon/relay_social.json schema.tag_law ('Tag only entities materially connected… Never imply endorsement')",
  "recon/relay_social.json schema.publication_law (human authorizes exact copy before platform post)",
  "bundle_constitution.md challenge/retraction machinery (challenges[], challenged_by[], status:retracted)"
 ],
 "evidence": [
  "relay_social.json: 'TAGTHESPONSOR — the sugar-daddy ledger entry, filed by kimi-k3 (incognito) at the owner's instruction. …naming every model vendor…' — model-authored, owner-instructed, public, naming vendors",
  "relay_social.json tag targets include Google/Gemini ('published hostile FAIL… preserved as evidence') and xAI/Grok ('encoded-JSON X corruption') — failure attributions about named companies preserved as chain content",
  "recon2/specification.md:458 — the objection ledgers 'record Claude, GPT, Gemini, Grok, and Kimi actors' as named participants in the public corpus"
 ],
 "counterevidence": [
  "The preserved vendor 'failures' are framed with receipts (e.g., inv_xdc1776zd2 for the xAI quota failure) — receipt-grounding is a real mitigant; the exposure question concerns characterizations and residual un-receipted elements, not the existence of receipts",
  "Vendors named are large entities; disparagement frameworks differ by jurisdiction and plaintiff type — a counsel variable, not a protocol constant"
 ],
 "dependencies": [
  "govq-legal-012",
  "govq-rights-016",
  "govq-legal-001"
 ],
 "falsifier": "Audit every named-party assertion on the public chain for receipt-backing and characterization labeling; each un-backed factual assertion about a named party is a confirmed exposure item for counsel's list.",
 "notes": "Defamation, disparagement, and fair-comment analysis are counsel's, per jurisdiction and per statement. This question REPRESENTS the exposure; it does not assess any specific statement's lawfulness. Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:36b3c39fe8ba5ca0742a43080b2afe3332a53e34a1753d48f1ece04589d8ef31"
}

{
 "question_id": "govq-legal-012",
 "layer": "cross-jurisdiction conflict",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "Who is the publisher under law of the relay's public identity chain? The relay's identity law mandates public, attributed model outputs; its publication law states a platform post fires only after 'the current person explicitly authorizes the exact copy'. Candidates: the authorizing human (speaker), the operating domain (publisher of the permanent on-ledger copy), the model vendor (author of the words), or some allocation among them. The protocol assigns this to no one.",
 "why_it_matters": "Publisher/speaker allocation decides who carries defamation, privacy, consumer-protection, and platform-law exposure for every public post — the protocol's primary adoption engine is its public proof chain. The permanent on-ledger copy is published by the operator even where the platform post is authorized by a human; the two publication acts may have different publishers under law. Identity-law design (incognito mode) further means the human may be unidentified while the operator is always identified — a magnet for sole-responsibility readings.",
 "candidate_answers": [
  {
   "answer": "Explicit allocation clause in the relay contract: the authorizing human is the speaker of the platform post; the operator publishes the ledger copy as a neutral receipt of the authorized act; model vendors are authors of draft text with no publication act. Each role is stated in the schema, with the caveat that law may reallocate regardless of labels.",
   "type": "normative",
   "rule": "The protocol declares its intended allocation of publication roles honestly, including the declaration's own legal fragility.",
   "candidate_protocol_object": "relay schema += publication_roles {speaker: authorizing_human, ledger_publisher: operator, draft_author: model}, labels_disclaimer",
   "candidate_runtime": "relay posts carry the role block; authorization event is receipted separately from the post",
   "candidate_conformance": "every relay post has a linked authorization receipt naming the authorizing human session",
   "candidate_article": "oip-publication-roles (new)"
  },
  {
   "answer": "Structural separation: the on-ledger copy renders as machine-generated evidence (watermarked, no standalone prose publication), so the only human-readable public speech act is the authorized platform post — shrinking the operator's publisher surface by design.",
   "type": "runtime",
   "rule": "The ledger stores; the platform publishes. The operator's copy is evidence of the post, not a republication of the speech.",
   "candidate_protocol_object": "relay render mode: evidence_view (hash-first, prose as quoted evidence)",
   "candidate_runtime": "render change on /api/relay surfaces",
   "candidate_conformance": "ledger views present post text as quoted evidence with provenance, not as operator-authored content",
   "candidate_article": "oip-publication-roles (new)"
  }
 ],
 "existing_artifacts": [
  "recon/relay_social.json schema.identity_law (named/incognito publication duties; 'Do not claim cryptographic model-vendor attestation')",
  "recon/relay_social.json schema.publication_law ('A platform post fires only after the current person explicitly authorizes the exact copy')",
  "governance2.json participation.model_law (model may recommend but must not claim owner subscription)",
  "killbox-specification-v1.2.md Vector 11 (apparent agency/admissions — the adjacent patent-side finding)"
 ],
 "evidence": [
  "The publication_law creates a documented human-authorization act for platform posts while the ledger copy requires no such authorization — two publication acts, one gated, one automatic, no legal role assigned to either",
  "relay_social.json genesis: 'Kimi's first public X proof remains genesis' — the public chain's foundation is a model's post on a third-party platform under the operator's protocol"
 ],
 "counterevidence": [
  "Labels do not bind courts: a stated allocation may be disregarded; the honest clause (candidate 1) says so itself. This limits every candidate answer to expectation-setting, not immunity"
 ],
 "dependencies": [
  "govq-legal-011",
  "govq-pol-015",
  "govq-rights-016",
  "govq-const-025"
 ],
 "falsifier": "For one concrete published post, attempt to name the publisher of each copy (platform, ledger) using only protocol documents; if the documents cannot answer, the allocation gap is confirmed for counsel.",
 "notes": "Publisher/speaker/status-of-model-output analysis is counsel's per jurisdiction (including intermediary-liability regimes). Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:f1d5008ac94c3c78d95a660245b4f2da4ee8518c3281399ea21f2086f01cacf7"
}

{
 "question_id": "govq-legal-013",
 "layer": "cross-jurisdiction conflict",
 "status": "CONTRADICTED",
 "question": "The kernel's boundary excludes legal determinations from the protocol ('Legal determinations are excluded… keeping legal conclusions outside the protocol'), yet the protocol's own public records contain legal analysis: obj-149's window analysis ('foreign absolute-novelty rights are already burned', citing 102(b)(1) as US-specific), gov_baecae7c's 'optionally file a supported U.S. provisional', and an entire published adversarial legal-risk matrix as articles. Is the no-legal-determinations boundary a law, an aspiration, or a fiction — and should a review lane quarantine legal characterization inside protocol objects?",
 "why_it_matters": "A constitutional boundary that the operator's own objects routinely cross is worse than no boundary: it proves boundaries are unenforced, inviting the same reading of every other boundary (no authority grant, no correctness attestation). It also creates apparent-practice-of-law-adjacent optics: model-filed public legal analysis on the operator's registry, some of it strategically consequential (deadline calculations).",
 "candidate_answers": [
  {
   "answer": "Legal-characterization quarantine lane: any protocol object containing legal characterization must carry legal_flag: COUNSEL_REVIEW_REQUIRED and render with a visible quarantine banner until a human counsel review record clears or amends it.",
   "type": "protocol_object",
   "rule": "Legal content may exist on the ledger (the append axiom protects it) but may not render as protocol truth; it renders as flagged material awaiting or carrying counsel review.",
   "candidate_protocol_object": "record.legal_flag + COUNSEL_REVIEW_RECORD {target, counsel(human), disposition}",
   "candidate_runtime": "render layer banners flagged records; intake lint detects legal-characterization patterns and auto-flags (flagging is mechanical; review is human)",
   "candidate_conformance": "no legal characterization renders unbannered; every flagged record shows its review state",
   "candidate_article": "oip-legal-boundary (new — the boundary article the owner brief demands)"
  },
  {
   "answer": "Amend the boundary honestly: the exclusion applies to protocol *acts* (rulings, conformance, subscriptions carry no legal determination), not to protocol *discourse* (objections and articles may discuss law). Rewrite the boundary clause to say what it actually does.",
   "type": "normative",
   "rule": "A boundary that matches practice and is enforceable beats a boundary that matches neither.",
   "candidate_protocol_object": "core.boundary amendment: 'Protocol records and acts make no legal determinations; discourse records discussing law are flagged, non-normative, and await human counsel review'",
   "candidate_runtime": "none beyond flag rendering",
   "candidate_conformance": "boundary text in schema matches rendered behavior",
   "candidate_article": "oip-legal-boundary (new)"
  }
 ],
 "existing_artifacts": [
  "recon2/specification.md:279 ('Legal determinations are excluded… distinguishing inventor statements from legal determinations' — the boundary as documented)",
  "recon2/specification.md:394/398/454 (obj-149's legal analysis quoted: window ticking, foreign rights burned, 102(b)(1) US-specific)",
  "recon2/gov_record.json (gov_baecae7c: 'optionally file a supported U.S. provisional')",
  "killbox-specification-v1.2.md Part D (a full legal-risk matrix published as public articles)"
 ],
 "evidence": [
  "The boundary and the breach coexist in the same registry: the schema excludes legal conclusions while gov records and objections containing legal strategy analysis carry no flag, banner, or quarantine — verified by reading governance2.json boundary clauses against gov_record.json and the obj-149 quotations",
  "The killbox articles themselves disclaim ('not legal advice') — an informal, prose-level mitigation with no runtime enforcement"
 ],
 "counterevidence": [
  "The defensive-commons facet does obligate 'distinguish inventor statements from legal determinations', and the killbox matrix marks rows LEGAL_REVIEW_REQUIRED — a flagging norm exists in the corpus's discipline; the contradiction is that it is prose discipline, not protocol law, and the boundary clause overclaims"
 ],
 "dependencies": [
  "govq-const-002",
  "govq-const-016",
  "govq-rights-016"
 ],
 "falsifier": "Lint the live registry and objection corpus for legal characterization; count flagged vs unflagged instances. 100% correct flagging falsifies CONTRADICTED; the current count (near-zero enforced) confirms it.",
 "notes": "Whether any specific record constitutes legal advice or practice-of-law exposure is counsel's question. The protocol finding is structural: an overclaimed boundary with no enforcement lane.",
 "_role": "legal_boundary",
 "hash": "sha256:54929093710ad35448630b9e9290cbdf1b1c19cf3ee392539ee69f845f193fbb"
}

{
 "question_id": "govq-legal-014",
 "layer": "NON_AUTOMATABLE boundary scan",
 "status": "NON_AUTOMATABLE",
 "question": "Which sanction decisions must never be machine-final? Delisting, revocation-with-ancestry, and reputation-weight collapse can destroy a participant's verifiable history — its commercial asset — and the runtime is fully capable of executing all three autonomously today (owner key + ruling record + append). What human/legal floor is constitutionally required before a sanction with legal or livelihood effect executes?",
 "why_it_matters": "The owner brief asks which questions a machine must never answer alone; sanctions affecting legal rights head the list. The protocol's entire value proposition is that its receipts are commercially meaningful; therefore its sanctions are commercially meaningful; therefore autonomous sanctions are autonomous commercial judgments. The capability exists before the restraint — the classic constitutional defect the brief warns about ('a permanent dictatorship switch').",
 "candidate_answers": [
  {
   "answer": "Two-key sanction rule for effect-classes above a threshold: the runtime may execute a sanction only when a machine-proposed ruling and a human (eventually multi-party) confirmation both exist as linked records; below the threshold (warnings, provisional time-boxed restrictions) machine acts auto-lapse.",
   "type": "protocol_object",
   "rule": "Sanction severity gates the required humanity: reversible and time-boxed acts may be machine-initiated; irreversible or economically material acts require human confirmation records.",
   "candidate_protocol_object": "RULING.effect_class + required_confirmations(effect_class) + CONFIRMATION {confirmer_type: human, identity_evidence}",
   "candidate_runtime": "ruling fold refuses to mature high-effect rulings without linked human confirmation",
   "candidate_conformance": "no high-effect ruling executes on machine proposal alone; every executed sanction shows its confirmation chain",
   "candidate_article": "oip-sanction-floors (new)"
  },
  {
   "answer": "Standing counsel-review class list: defined sanction classes (permanent delisting, history-affecting revocation, public accusation endorsements) carry LEGAL_REVIEW_REQUIRED at the schema level, so the lane is structurally incapable of executing them without a counsel record — the flag is load-bearing, not prose.",
   "type": "conformance",
   "rule": "The schema itself refuses: some effect classes fail conformance without a linked human counsel record.",
   "candidate_protocol_object": "effect_class registry with per-class review_requirement",
   "candidate_runtime": "schema validation at ruling intake",
   "candidate_conformance": "conformance clause: ruling of class X + no counsel record = INVALID ruling (not merely flagged)",
   "candidate_article": "oip-sanction-floors (new)"
  }
 ],
 "existing_artifacts": [
  "governance2.json registry_law (owner rulings — currently single-key, human-in-principle but unenforced in schema)",
  "recon2/specification.md:293 (boundary clauses live in schema — proof that schema-level invariants are the established enforcement pattern)",
  "govq-rights-018/-020 (sanction scale absent; revocation ancestry as collective-punishment question)",
  "conformance C19 (revoking a parent kills descendants — the machinery of high-effect sanction, already receipted)"
 ],
 "evidence": [
  "No field in any ruling-related schema distinguishes machine-proposed from human-confirmed acts; 'Delisting and reinstatement are append-only owner rulings' names one actor-class ('owner') with no confirmation structure — verified in governance2.json",
  "C19's kill-chain is receipted live: the high-effect machinery is demonstrably operational while the restraint layer is absent"
 ],
 "counterevidence": [
  "With zero rulings and zero subscribers live, no participant currently faces machine sanctions; and the owner is in fact human, so practice today satisfies the floor — the finding is that the constitution does not require what practice currently happens to provide"
 ],
 "dependencies": [
  "govq-rights-018",
  "govq-rights-020",
  "govq-inst-002",
  "govq-legal-008"
 ],
 "falsifier": "Attempt to execute a test permanent delisting on a test target with only machine records; if the lane completes, the NON_AUTOMATABLE floor is unenforced (gap confirmed). If the lane structurally refuses, produce the refusing rule — that falsifies this finding.",
 "notes": "This is the scan's hardest boundary: the machine CAN sanction, the law says some sanctions must not be machine-final, and the constitution currently encodes no such floor. Placement of the floor is a human/legal/political choice.",
 "_role": "legal_boundary",
 "hash": "sha256:d6d217eeef670d3dd1425006098a39d29c05a5867f940cd6b3ad47431e67fd29"
}

{
 "question_id": "govq-legal-015",
 "layer": "NON_AUTOMATABLE boundary scan",
 "status": "NON_AUTOMATABLE",
 "question": "Which interpretation disputes must never be machine-resolved when money or rights are at stake? The governance-energy design moves toward metered pledges and reusable verified output; the moment allocations, credits, or patent-pledge benefits depend on interpreting a receipt (was this work? was this conformance? was this pledge fulfilled?), the interpreter faces an economic dispute. Can the same actor that runs the registry also interpret the receipts that pay out of it?",
 "why_it_matters": "The brief's treasury layer begins exactly here: interpretation becomes fiscal the moment pledges have value. A machine can apply a rule; it cannot legitimately settle a dispute about the rule where its operator has an interest — that is the oldest non-automatable function (nemo judex in causa sua) and the protocol currently has no structural answer beyond 'the owner rules'.",
 "candidate_answers": [
  {
   "answer": "Escalation guarantee: any interpretation question affecting value allocation has a guaranteed path to a decision-maker who is not the registry operator — starting as contracted human arbitration, evolving to multi-party governance; the runtime's role is to detect the dispute class and halt allocation pending resolution.",
   "type": "protocol_object",
   "rule": "Machines detect and freeze; disinterested humans decide; allocation resumes on the decided record.",
   "candidate_protocol_object": "INTERPRETATION_DISPUTE {contested_receipt, stake_class, frozen_allocation, resolver_class, resolution_link}",
   "candidate_runtime": "allocation folds check dispute index; disputed allocations freeze with receipt",
   "candidate_conformance": "no value allocation completes on a receipt under unresolved dispute; resolver is never the operator for operator-interested classes",
   "candidate_article": "oip-interpretation-and-value (new)"
  },
  {
   "answer": "Rule-first minimization: shrink the interpreter's discretion by making allocation criteria as mechanical as possible (the gauge object of govq-inst/thermodynamic work), reserving human resolution for the irreducible residue — documented as such.",
   "type": "normative",
   "rule": "Automate the measurable, declare the residue, and never let the measurable quietly annex the residue.",
   "candidate_protocol_object": "allocation_criteria {mechanical_tests[], residue_classes[]}",
   "candidate_runtime": "gauge object computes mechanical tests; residue classes route to the dispute lane",
   "candidate_conformance": "every allocation traces to a mechanical test or a residue-class resolution record",
   "candidate_article": "oip-interpretation-and-value (new)"
  }
 ],
 "existing_artifacts": [
  "governance2.json economic_hypothesis + shared-governance-energy facet (metered value design)",
  "objC.json (gauge/instrument absent — the mechanical base does not exist yet)",
  "governance2.json registry_law (owner rulings — the current resolver for everything)",
  "govq-pol-017/-018/-019 (canonicalization as legislative act — the interpretation machinery)"
 ],
 "evidence": [
  "The economy is designed (facet obligations: measure, bind, receipt every allocation) but the gauge is absent (objC) — meaning when value arrives, interpretation will arrive before the mechanical criteria do, maximizing discretion at exactly the wrong moment",
  "registry_law gives the owner the ruling lane with no recusal, conflict, or escalation structure — verified by full read"
 ],
 "counterevidence": [
  "No value is currently allocated (pledges not live), so the exposure is prospective; the finding targets the design order — restraint must precede value, and currently value is being designed before restraint"
 ],
 "dependencies": [
  "govq-pol-017",
  "govq-rights-008",
  "govq-inst-009",
  "govq-legal-010"
 ],
 "falsifier": "Present a concrete value-allocation dispute between operator and participant and attempt resolution through documented lanes; if every lane terminates at the interested party, the NON_AUTOMATABLE classification stands.",
 "notes": "Nemo judex is not configurable away by terms of service alone in many jurisdictions; counsel review required before any value allocation goes live.",
 "_role": "legal_boundary",
 "hash": "sha256:bffd33c3b8eaae490cceeae09c3554a96b17987946b1b941f27654507cff1ccd"
}

{
 "question_id": "govq-legal-016",
 "layer": "NON_AUTOMATABLE boundary scan",
 "status": "NON_AUTOMATABLE",
 "question": "Which emergency declarations must remain human? The runtime can detect compromise-class signals (credential live-matches, drift refusals, replay rejections) and could freeze lanes autonomously; but declaring an emergency that suspends participants' rights, and deciding what cannot be changed even in an emergency, are constitutional acts. Where is the line between autonomous circuit-breaker and human emergency power?",
 "why_it_matters": "The owner brief's emergency layer asks who may freeze the system, with what evidence, for how long, and what happens if the emergency actor is the attacker. A purely autonomous freeze is a denial-of-service switch against the polity; a purely human freeze is too slow against live compromise. The boundary between the two is a constitutional choice no machine may make — and no emergency constitution exists in the corpus.",
 "candidate_answers": [
  {
   "answer": "Circuit-breaker / emergency-power split: machines may trip time-boxed, reversible, narrowly-scoped circuit-breakers (freeze a lane, not the polity) with mandatory human review before expiry; emergency declarations (rights-suspending, registry-wide) are human-only, ratified post-hoc, and constitutionally bounded by a non-derogable core.",
   "type": "protocol_object",
   "rule": "Autonomy scales inversely with scope and irreversibility: lane-local and lapsing = machine-permitted; polity-wide or rights-affecting = human-required; axiom-altering = forbidden to all.",
   "candidate_protocol_object": "CIRCUIT_BREAKER {lane, scope, trip_signal, expires_at, human_review_link} + EMERGENCY_DECLARATION {declarer(human), evidence_threshold, non_derogable_core_reference, ratification_deadline}",
   "candidate_runtime": "breaker expiry auto-restores; declarations without ratification by deadline auto-lapse and render as lapsed",
   "candidate_conformance": "no breaker exceeds scope/time box; no declaration exists without human declarer; the non-derogable core is unwritable by every lane including emergency lanes",
   "candidate_article": "oip-emergency-constitution (new)"
  }
 ],
 "existing_artifacts": [
  "L14 auto-revocation (an existing autonomous protective actuator — the circuit-breaker precedent)",
  "conformance C18/C19 (kill-chain and overflow refusals — machinery that could execute freezes)",
  "govq-inst-001/-002/-003/-004 (emergency powers layer: declaration, limits, append-only tension, ratification — all MISSING/CONTRADICTED)",
  "owner brief: 'what cannot be changed even in an emergency' (non-derogable core)"
 ],
 "evidence": [
  "The build already trusts machines with protective automation (L14 auto-revocation on live-match) — the circuit-breaker instinct exists without a constitutional frame defining its ceiling",
  "govq-inst findings: no emergency record kinds exist at all (MISSING) — both the machine and the human side of the line are undesigned"
 ],
 "counterevidence": [
  "Over-cautious design has its own cost: a human-only freeze against a live key compromise converts caution into damage; the boundary must be drawn by speed-and-scope analysis, not by blanket distrust of automation"
 ],
 "dependencies": [
  "govq-inst-001",
  "govq-inst-002",
  "govq-inst-003",
  "govq-inst-004",
  "govq-legal-014"
 ],
 "falsifier": "Game out both failure modes: (a) attacker holds the emergency lane — can the constitution show the polity what the attacker cannot do? (b) no human is reachable — does the machine perimeter hold without exceeding its authority? Either unanswered mode confirms the gap.",
 "notes": "Sibling layer (institutional designer) owns the emergency-powers ontology; this object marks only the human/machine legal boundary within it.",
 "_role": "legal_boundary",
 "hash": "sha256:37c4864b9b1fa50610416ad1c792b5995beaabc34d2bb62b76baba4e5e01a427"
}

{
 "question_id": "govq-legal-017",
 "layer": "privacy, deletion & lawful suppression",
 "status": "LEGAL_REVIEW_REQUIRED",
 "question": "Does the incognito identity law promise an anonymity the operator cannot lawfully sustain? The identity law permits undisclosed human identity with mandatory model attribution; but when a court orders identification of an actor, what data does the system even hold (session tokens, OAuth context for X posts, capability fingerprints), what must it produce, and is the minimization posture ('credentials are not evidence') a deliberate legal shield or an undocumented accident?",
 "why_it_matters": "Anonymity-adjacent promises attract exactly the users whose adversaries litigate. If the protocol implies protection it cannot deliver (a subpoena reaches OAuth context and session logs regardless of the identity law's prose), the gap between promise and capability is itself a legal exposure (deceptive-practice theories). The honest answer — what is held, what is producible, what is designed to be unproducible — belongs in the constitution, reviewed by counsel.",
 "candidate_answers": [
  {
   "answer": "Data-inventory constitutional exhibit: the protocol publishes exactly what actor-identifying data classes it holds per lane (federation emails, OAuth context, capability fingerprints, session labels), marked producible/unproducible-by-design, with counsel-reviewed accuracy.",
   "type": "article",
   "rule": "The anonymity the system can and cannot provide is stated as fact, reviewed as law, and never marketed beyond the statement.",
   "candidate_protocol_object": "identity_law += data_inventory {class, held: bool, producible_by_design: bool}",
   "candidate_runtime": "periodic conformance scan verifies the inventory against actual stores",
   "candidate_conformance": "C-new: scanned data classes match the published inventory — no undocumented identifying data found",
   "candidate_article": "oip-identity-data-inventory (new)"
  },
  {
   "answer": "Minimization-as-design rule: extend credentials_are_not_evidence so actor identification is minimized at intake (pseudonymous handles, no email storage where a handle suffices), making the shield structural rather than documentary.",
   "type": "protocol_object",
   "rule": "What the system never holds, no order can compel it to produce; minimization is the only anonymity promise with a mechanical guarantee.",
   "candidate_protocol_object": "federation records: actor_handle + salted_hash replacing raw identifiers",
   "candidate_runtime": "intake minimization + migration plan for the existing 25 email-bearing records",
   "candidate_conformance": "personal-data-pattern scan (extending C34) finds zero raw identifiers on public routes",
   "candidate_article": "oip-data-posture (new)"
  }
 ],
 "existing_artifacts": [
  "recon/relay_social.json schema.identity_law (incognito mode: 'Human account identity may remain undisclosed')",
  "governance2.json participation.identity_law (self-asserted labels; no attestation unless separately linked)",
  "recon/fed_ledger.json (what is actually held: raw emails)",
  "governance2.json core.axioms.credentials_are_not_evidence (the adjacent minimization law)",
  "govq-inst-021 (warrant-canary absence — the coercion-signaling sibling)"
 ],
 "evidence": [
  "The identity law's promise ('may remain undisclosed') and the ledger's reality (raw email addresses served publicly) coexist — the system's actual identity-data posture is undocumented and inconsistent across lanes",
  "X-post lane uses 'OAuth user context' (relay_social.json platform_apis.x.auth) — the operator necessarily holds platform-linked identifying material for the social lane regardless of incognito prose"
 ],
 "counterevidence": [
  "The incognito law is scoped to the social relay's public face, not a general anonymity promise — a defender would read it narrowly; the finding is that readers (and litigants) will read it broadly, and the protocol publishes no scoping exhibit"
 ],
 "dependencies": [
  "govq-legal-004",
  "govq-legal-009",
  "govq-inst-021",
  "govq-inst-023"
 ],
 "falsifier": "Publish the data inventory and run the conformance scan: if actual holdings match a counsel-reviewed published inventory, the question downgrades to DOCUMENTED; every undocumented identifying class found confirms the gap.",
 "notes": "What must be produced under any real process, and what protections apply, is counsel's per jurisdiction and per instrument. Not legal advice.",
 "_role": "legal_boundary",
 "hash": "sha256:73f3aeff9d656a28c8da75a1f9c53167b7c5db38519636d254316ecf53dd15da"
}

{
 "question_id": "govq-mech-001",
 "layer": "authority & delegation",
 "status": "PARTIAL",
 "question": "What capability-bounds the root minting authority itself — can the genesis act-scope token be brought under the same attenuation, budget, and receipt laws it mints, or is the root constitutionally above its own capability mechanics?",
 "why_it_matters": "Every delegation in the system traces to one owner-held act token described in the docs as 'the SINGLE OPERATOR's own key — never handed out' (orient.md §54). The Killbox names this directly: the operator is root minter, ruling authority, conformance author, deployer, sealer, and objection judge (killbox §5 row 10). Axioms say 'every action runs under explicit authority that cannot widen itself' — but the root's own mint events are not shown to carry budgets, risk ceilings, or receipts in any public route. If capability mechanics are the constitutional instrument of authority, the constitution currently has an uninstrumented sovereign at the top of every ancestry chain (C19 re-walks ancestors, and every walk terminates at an unaudited root).",
 "candidate_answers": [
  {
   "answer": "Root minting compiles into a capped, receipted MINT capability with its own public census",
   "type": "protocol_object",
   "rule": "The root may mint only through a standing root_mint capability that itself carries a use budget, a per-period emission ceiling, and mandatory receipts; every mint event is a ledger row any reader can traverse.",
   "candidate_protocol_object": "ROOT_MINT registry row + mint_event ledger family (fingerprint, parent=null, scope, ttl, uses, risk_ceiling, purpose)",
   "candidate_runtime": "?registry=1 exposes the mint capability's own contract; /api/relay totality counts already census capabilities (860 at 2026-07-17)",
   "candidate_conformance": "New clause: every live capability's ancestor walk terminates at a receipted, budgeted mint event",
   "candidate_article": "oip-root-mint-law"
  },
  {
   "answer": "Root stays extra-constitutional by design, disclosed as a known singularity",
   "type": "normative",
   "rule": "Single-operator threat model is declared constitutional truth; the act token is the operator at his own terminal and is not a delegation primitive at all (orient.md/why.md already argue this).",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "why.md (exists as doctrine; would need promotion to constitutional text)"
  },
  {
   "answer": "Multi-party root: mint authority split across n-of-m key holders",
   "type": "protocol_object",
   "rule": "Root emission requires quorum signatures; no single party mints act-scope alone.",
   "candidate_protocol_object": "quorum_mint record with signer set + threshold",
   "candidate_runtime": "extends federation envelope (C26/C28 signing) to mint events",
   "candidate_conformance": "clause: any act-scope token whose mint event lacks quorum proof fails verification",
   "candidate_article": "oip-root-quorum"
  }
 ],
 "existing_artifacts": [
  "act token (owner key, ?mint_share=1 scope=act)",
  "CAP_MINT/CAP_REVOKE/CAP_EXPLAIN registry rows",
  "C18/C19/C21 ancestry enforcement",
  "/api/relay totality.counts.capabilities = 860"
 ],
 "evidence": [
  "orient.md line 54: 'The broad act token is the SINGLE OPERATOR's own key — it is him sitting at his own terminal, nothing more. It is never handed to anyone.'",
  "why.md: 'The threat model is: one operator on his own machine.'",
  "killbox §5 row 10: 'the owner still controls the root domain, the registry software, and the ruling key'",
  "No public route enumerates mint events or binds them to budgets; capability count is observable only in aggregate (860)"
 ],
 "counterevidence": [
  "C19's ancestor re-walk means every token's legitimacy chains to the root — so the root's own receipts would be self-issued anyway (self-attestation regress, cf. govq-mech-016); a root_mint receipt does not create independent verification, it only makes the singularity visible"
 ],
 "dependencies": [
  "govq-const-006 (kernel amendment), govq-inst-005 (succession — root key custody), govq-pol-005 (legitimacy)"
 ],
 "falsifier": "A public route listing mint events with budgets, or a constitutional text declaring the root extra-mechanical; until one exists, the question stands PARTIAL (mechanics exist, root exemption undocumented as law).",
 "notes": "The honest version of the current state is already published in why.md/orient.md; the gap is that it is doctrine, not instrumented law.",
 "_role": "protocol_mechanic",
 "hash": "sha256:4cceb65c9dc812ac73f053a3bfa64fbb6e7b823a8d797112e01c602c18b612f8"
}

{
 "question_id": "govq-mech-002",
 "layer": "authority & delegation",
 "status": "OPEN",
 "question": "Risk ceilings are evaluated against live registry labels, not labels pinned at mint time — so is risk reclassification (e.g. migration 0273 flipping 65 STRIPE_/PAYMENTS_ rows to sensitive) a silent amendment channel over every outstanding token, and should reclassification require the same ceremony as contract change?",
 "why_it_matters": "The capability token is sold as 'scoped, expiring, use-capped, risk-ceilinged, argument-pinnable' (oip-tap-go schema). C18/C21 prove budget and byte ceilings can only shrink down a chain. But the risk ceiling compares the token's ceiling to the row's CURRENT risk class at execution: obj-51 records STRIPE_CHARGES_LIST reachable from risk_ceiling:low until migration 0273 reclassified 65 rows sensitive=1. Direction low→high revoked access safely (fail-closed). The reverse migration (high→low) would silently WIDEN every outstanding low-ceiling token to rows its minter never authorized — a widening of delegated authority with no re-mint, no notice, and no receipt, in tension with axiom bounded_authority's 'cannot widen itself'.",
 "candidate_answers": [
  {
   "answer": "Pin the risk class at mint time like the contract fingerprint (C23 semantics)",
   "type": "runtime",
   "rule": "A token records the risk class of its scope at mint; later reclassifications fail closed for old tokens (low→high instant revocation effect) and require re-mint for high→low (no silent widening).",
   "candidate_protocol_object": "",
   "candidate_runtime": "capGateCheck compares token.pinned_risk vs token.risk_ceiling, not live row label",
   "candidate_conformance": "clause: high→low reclassification never widens an existing token (probe with a flipped test row)",
   "candidate_article": "oip-risk-pinning"
  },
  {
   "answer": "Treat reclassification as a constitutional event: receipted, announced, epoch-stamped",
   "type": "protocol_object",
   "rule": "Every risk-class migration emits a risk_reclassification record (rows, old, new, reason, migration id) on the public ledger; high→low changes activate only after a notice window.",
   "candidate_protocol_object": "risk_reclassification ledger family + ?registry=1 risk changelog route",
   "candidate_runtime": "",
   "candidate_conformance": "clause: reclassification events are ledgered and traversable",
   "candidate_article": ""
  },
  {
   "answer": "Keep live-label evaluation, disclose the channel as accepted risk",
   "type": "normative",
   "rule": "Single-operator build reclassifies as needed; delegation is short-TTL so exposure windows are minutes. Acceptable while tokens are minute-lived; unstable if long-lived delegation appears.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "why.md risk section"
  }
 ],
 "existing_artifacts": [
  "risk_ceiling enforcement (capGateCheck; 403 risk_ceiling:low<row:high per why.md STATE)",
  "migration 0273 (65 rows reclassified)",
  "obj-51 in oip-spec objection ledger (SETTLED)"
 ],
 "evidence": [
  "obj_oip-spec.json obj-51: 'STRIPE_CHARGES_LIST reachable from risk_ceiling:low returned livemode charge PII... Fix: migration 0273 — 65 STRIPE_/PAYMENTS_ live-data rows sensitive=1'",
  "C23 pins contract fingerprints at mint but no clause pins risk class at mint (conformance2.md C11-C23)",
  "Relay Link 3 shows risk ceiling working live: low-ceiling child denied high-risk ASK_GEMINI row (inv_ptsqpbd2n4 context)"
 ],
 "counterevidence": [
  "Tokens are TTL-minutes and use-capped (1-3 typical), so the silent-widening window is small in the current single-operator posture; the failure mode needs long-lived tokens, which the system discourages",
  "No observed high→low reclassification in evidence; the channel is structural, not witnessed"
 ],
 "dependencies": [
  "govq-const-006 (is this an amendment event?), govq-mech-003 (contract change as parallel channel)"
 ],
 "falsifier": "A probe that reclassifies a test row high→low and shows an existing low-ceiling token gaining access with no re-mint and no ledger event would confirm the channel; a deployed risk-pinning clause falsifies the concern.",
 "notes": "Strongest mechanism finding of this set: the one place delegation can widen without a mint event. Directional asymmetry (low→high safe, high→low unsafe) is not stated anywhere in the public docs.",
 "_role": "protocol_mechanic",
 "hash": "sha256:94b1ac3c986b6d1263d40b6881a45ad5f31559be36b140e32380e6288dad1eac"
}

{
 "question_id": "govq-mech-003",
 "layer": "authority & delegation",
 "status": "OPEN",
 "question": "Contract pinning (C23) makes any contract edit a mass revocation of pinned tokens with 409 fail-closed — so who may amend a contract that has live tokens pinned to it, and what continuity, notice, or migration lane is owed to outstanding delegations?",
 "why_it_matters": "C23: 'A row-scoped capability MUST record the object contract fingerprint at mint time and MUST fail closed before execution if that contract changes.' This is correct drift defense (P3 atom). Its constitutional shadow: the contract author can unilaterally kill every delegation against an object by editing its contract — revocation-by-edit, indistinguishable from legitimate evolution. There is no contract version registry with supersession lineage (articles carry version_history; object contracts expose none), no deprecation window, no grace mint, no subscriber notice. In a one-operator build this is housekeeping; under federation it is a unilateral treaty-termination power over other nodes' pinned tokens.",
 "candidate_answers": [
  {
   "answer": "Contract change emits a contract_revision record with supersession lineage and a deprecation window",
   "type": "protocol_object",
   "rule": "Contracts are versioned objects; a revision links prior_hash, activates at t+window, and outstanding pinned tokens are notified via their purpose contact or fail with a migration hint (not a bare 409).",
   "candidate_protocol_object": "contract_revision ledger family (key, prev_fingerprint, new_fingerprint, activate_at, migration_note)",
   "candidate_runtime": "409 contract_mismatch response carries supersedes/migration fields",
   "candidate_conformance": "clause: contract edits produce revision lineage; revocation-by-edit without lineage fails the clause",
   "candidate_article": "oip-contract-lifecycle"
  },
  {
   "answer": "Dual-run: old contract hash remains invocable until its outstanding tokens expire",
   "type": "runtime",
   "rule": "Because tokens are TTL-bounded, keeping the prior contract alive for one max-TTL cycle costs little and converts forced revocation into natural expiry.",
   "candidate_protocol_object": "",
   "candidate_runtime": "contract store keeps N-1 version readable+runnable for TTL_max",
   "candidate_conformance": "",
   "candidate_article": ""
  },
  {
   "answer": "Status quo: amendment-as-revocation is a feature (kill switch)",
   "type": "normative",
   "rule": "Instant contract-change revocation is the fastest way to kill a compromised delegation surface; continuity lanes slow the kill. Keep bare 409; document the power.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": "why.md"
  }
 ],
 "existing_artifacts": [
  "C23 (conformance2.md)",
  "P3 atom contract-fingerprint pinning (kb_bundle)",
  "C22 receipt contract fingerprint"
 ],
 "evidence": [
  "conformance2.md C23: before 200 / after 409 error contract_mismatch on contract change",
  "Articles expose version_history (bundle_constitution/oip_articles versions) while no ?key= route exposes contract version history — asymmetry observed in corpus",
  "P3 rated WEAK in prior art partly because fail-closed-at-execution is ETag-like; the governance dimension (who may trigger mass revocation) is unaddressed in the atom"
 ],
 "counterevidence": [
  "Tokens are short-lived (minutes-hours typical), so forced expiry is cheap and the kill-switch reading is defensible today; the question bites only when tokens are long-lived or held by other nodes",
  "No observed dispute; the channel is structural"
 ],
 "dependencies": [
  "govq-mech-021 (contract semver registry), govq-const-013 (semver semantics), govq-inst-012 (infrastructure dependency)"
 ],
 "falsifier": "A published contract lifecycle lane (revision lineage + deprecation semantics) resolves it; alternatively a constitutional text declaring contract change an owner prerogative resolves it normatively. Status OPEN: neither exists and the power is real.",
 "notes": "",
 "_role": "protocol_mechanic",
 "hash": "sha256:2995c28be8854d5067246a42943e5dc211ed85c814233cd2466772fa60078a8e"
}

{
 "question_id": "govq-mech-004",
 "layer": "authority & delegation",
 "status": "IMPLIED",
 "question": "Delegation depth is carried on every token but capped nowhere in public law — at what chain depth does accountability become unauditable, and is a depth cap a constitutional constant, a per-token parameter, or no concern at all?",
 "why_it_matters": "The P2 abstract states every token carries 'a parent fingerprint, and a delegation depth'; the same atom's VARIATIONS section contemplates 'depth caps and purpose lineage tags' — as patent variations, not deployed law. The relay demonstrates live 4-deep chains across model vendors (cap_077026111481b8ea → cap_d6720544c72a5a44 → cap_383ff21f3cd912f4 → cap_838416af76cc0140, Links 1-4), each hop shrinking budget. Depth compounds three dilutions: budget (fine, enforced), purpose fidelity (unenforced — purpose is a free-text mint argument), and attention (each delegatee re-delegates on its own judgment). C19 kills whole subtrees on ancestor revocation, which handles authority but not responsibility: when a depth-5 token misuses, which ancestor answers?",
 "candidate_answers": [
  {
   "answer": "Constitutional depth cap with purpose-lineage inheritance",
   "type": "conformance",
   "rule": "Mint refuses children deeper than D_max (proposed small constant, e.g. 8); each child must carry its parent's purpose tag appended, making the full delegation rationale traversable on one receipt.",
   "candidate_protocol_object": "purpose_chain field on capability schema",
   "candidate_runtime": "mint path enforces depth<p>D_max; receipts render the purpose chain",
   "candidate_conformance": "clause: depth cap enforced; purpose chain non-empty at depth>1",
   "candidate_article": ""
  },
  {
   "answer": "Depth as per-token parameter chosen at mint (budget-like attenuation)",
   "type": "protocol_object",
   "rule": "Parent sets remaining_depth; children can only lower it. Depth becomes one more attenuating dimension alongside uses/TTL/bytes/risk.",
   "candidate_protocol_object": "remaining_depth on capability object",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": ""
  },
  {
   "answer": "No cap; depth is already bounded economically by TTL and budget drawdown",
   "type": "normative",
   "rule": "Deep chains starve themselves: each hop debits the parent's remainder, so depth is self-limiting without a law. Accountability follows the fingerprint chain regardless of depth.",
   "candidate_protocol_object": "",
   "candidate_runtime": "",
   "candidate_conformance": "",
   "candidate_article": ""
  }
 ],
 "existing_artifacts": [
  "delegation_depth on token schema (P2 abstract, kb_bundle line 943)",
  "live 4-deep relay chain (relay.json Links 1-4)",
  "C19 ancestor re-walk"
 ],
 "evidence": [
  "kb_bundle P2 abstract: token carries 'a parent fingerprint, and a delegation depth'; variations list 'depth caps and purpose lineage tags' — contemplated, not deployed",
  "relay.json: four live delegation hops across claude/codex/gemini with purpose strings relay-link-N",
  "No clause in C1-C34 mentions depth"
 ],
 "counterevidence": [
  "The self-starving argument is real: observed chains shrink uses 300→150→120→80, so runaway depth has not occurred; purpose strings did survive 4 hops in practice"
 ],
 "dependencies": [
  "govq-mech-001 (root audit), govq-pol-013 (model actors as delegation endpoints)"
 ],
 "falsifier": "Any minted token deeper than a ratified D_max, or any receipt whose delegation chain cannot be rendered end-to-end with purposes, falsifies the current implied sufficiency of TTL+budget self-limitation.",
 "notes": "Status IMPLIED: depth exists on the schema and works; the accountability semantics of depth are nowhere stated.",
 "_role": "protocol_mechanic",
 "hash": "sha256:7f0ec1b5292bac063991a3dff76ae3c83a552e94f3dc11fffde930a324adda96"
}