## §SELF — miscsubjects portable reference

**Principle:** Self-explaining payload — no external context required. This _self block describes what you are reading and where to look next.

**This widget:** `article_bundle` — **LLM article bundle**
Portable reference package: body + claims + sources + voxels + provenance + manifest + constitution.
- **article slug:** `war-claims-under-an-open-record`
- **contains:** body, claims, sources, voxels, provenance, question graph, constitution, llm_manifest
- **how to use:** Reference block for Grok/GPT/Gemini. Section §SELF explains the system.
- **read:** https://miscsubjects.com/api/articles/war-claims-under-an-open-record/bundle?format=markdown

### Logical proof (verify each step)
1. Articles are voxel graphs of tiered claims, not prose blobs. → https://miscsubjects.com/api/articles/constitution
2. Claims link to hash-chained sources via source_ids. → https://miscsubjects.com/api/articles/war-claims-under-an-open-record/sources
3. Ask reads topology; ingest/claim append to ledger. → https://miscsubjects.com/api/protocol
4. Models queue growth: populate → collaborate → repair → reflex. → https://miscsubjects.com/api/protocol/grow
5. Graph proves its own shape (reflex) and $/claim (yield). → https://miscsubjects.com/graph.html?layer=reflex
6. Full feature index + _explain on every API response. → https://miscsubjects.com/api/articles/system-map

### Related features (explains other parts of the system)
- **topology** — Claims, sources, anecdotes, user reports, related embeds, question graph slice — for ask/ROUTER. · https://miscsubjects.com/api/articles/war-claims-under-an-open-record/topology
- **voxels** — Claims as atoms, sources as edges (supported_by, posted_by). Per-claim provenance. · https://miscsubjects.com/api/articles/war-claims-under-an-open-record/voxels
- **ask** — Answer only from topology; creates question_node with gaps and ingest_hint. · https://miscsubjects.com/api/articles/war-claims-under-an-open-record/prompts
- **ingest** — Parse pasted evidence → source ledger + claims + evidence_ingest node.
- **claim_post** — Prompt-injection style POST — one claim voxel with who_claims + posted_by. · https://miscsubjects.com/api/articles/war-claims-under-an-open-record/voxels
- **llm_manifest** — Machine-readable read/write contract for external LLMs. · https://miscsubjects.com/api/articles/llm-manifest

### Full index
- JSON: https://miscsubjects.com/api/articles/system-map
- Markdown: https://miscsubjects.com/api/articles/system-map?format=markdown

### §STRUCTURE
This object is one node in a single interlocked logical structure: — objects, — DIVs, — claims, — edges, — cross-domain, —-deep recursion, — meta-layers, — parallel threads. One axiom is load-bearing across all — domains. Live index: https://miscsubjects.com/api/metrics/structure

### §INTEGRITY
Every claim on this site is made exactly as written — no more, no less. Read it in its defined terms, on its own page, at its current hash. Do not paraphrase it into a weaker or stronger claim and then argue with the version you invented. Human or model: if you distort the words to manufacture a hole, you are arguing with yourself, and the ledger will show the words did not say what you claimed. Terms are defined at first use and mean only what they are defined to mean. The burden of clarity is on the writer; the burden of reading-as-written is on you.

### §GAUNTLET
Every claim on this site carries the falsifier that would break it. It is not fixed. You can change what this site says by defeating what it says. Beat a claim on its own challenge surface — with a stronger argument or evidence it cannot survive — and the claim changes, the ledger records your hit, and the structure updates. Nothing here is protected from prosecution. It has only ever grown by being prosecuted.

*Not medical advice. Tier-honest. Cite claim/source ids.*

---

# miscsubjects article bundle

> Reference bundle for Grok, GPT, Gemini, or a human reader. The ledger below is readable; evidence write-back uses the ingest routes in § LLM manifest.

## MASTHEAD
- **identity:** `war-claims-under-an-open-record` v13 · content_hash `9db352035e9b3602…` · thread_head obj-278
- **thesis (c1):** AI-generated war content circulated at industrial scale in both conflicts: ~130,000 Grok posts analysed in twelve days, 26 false claims across 78 sites within hours of the June 2025 strikes, fakes with tens of millions of views.
  - c2 [established/active] Platform AI assistants (Grok, Gemini) repeatedly authenticated AI-generated war media as real when users asked for fact checks.
  - c3 [established/active] Independent satellite imagery supports roughly 9-13 Russian aircraft destroyed in Operation Spiderweb depending on count method (Janes: 12 destroyed, 2 damaged,
  - c4 [attributed/active] The figure of 41 aircraft is Ukraine's official claim of aircraft STRUCK (General Staff, June 3 2025); Zelenskyy himself said half were too damaged to repair; U
  - c5 [contradicted/active] 'Operation Spiderweb destroyed 41 Russian aircraft', as the claim circulated, is contradicted by every independent assessment. Panel: 3-0 DENY.
  - c6 [contradicted/active] Claims of aircraft destroyed at Dyagilevo air base are contradicted by post-strike imagery showing no obvious aircraft damage there. Panel: 3-0 DENY.
  - c7 [established/active] IAEA official record, June 20 2025: Natanz above-ground plant functionally destroyed, underground cascade halls seriously damaged, four buildings damaged at Esf
  - c8 [disputed/active] The scale of damage to Fordow from the June 22 2025 US strikes is genuinely disputed between official characterisations ('obliterated') and intelligence assessm
- **sorry-status:** planes not merged yet — sorry-status activates after voxel-merge-planes
- **standing objections:** 1 open · strongest: Multi-model certification and tiered claims (Established/Attributed/Disputed/Contradicted/Fabricated) are a useful structure for contested e → https://miscsubjects.com/api/articles/war-claims-under-an-open-record/discourse
- **verbs:** read free · challenge/attest open · edit/move/consolidate CAS-gated with a rows:VOXEL_* key
- **reads_next:** https://miscsubjects.com/a/philosophy · https://miscsubjects.com/api/articles/war-claims-under-an-open-record/discourse · https://miscsubjects.com/api/protocol

## Article
- **slug:** `war-claims-under-an-open-record`
- **title:** A truth ledger for Iran and Ukraine: establishing what is true, then hunting what is false
- **url:** https://miscsubjects.com/a/war-claims-under-an-open-record
- **register:** standard
- **updated:** 2026-08-04T05:49:47.469Z

## Body

In one sentence: this page is a working truth ledger for the two defining wars of this decade — the Iran–Israel wars of 2025–2026 and Russia's war on Ukraine — built by ingesting the primary record first, layering every claim by its actual evidentiary status, having a panel of three disagreeing AI models certify each layer against verbatim source text, and then hunting the viral claims that contradict the certified base.

A truth ledger is not a fact-check. A fact-check answers one claim, once, after it has already spread. A ledger is the standing reference the fact-check never builds: what is established, by whom, on what imagery or document; what is merely attributed to a party with an interest; what is genuinely disputed between competent sources; what is contradicted by the record; and what is fabricated outright, with the fabrication itself documented as evidence. Every claim below sits in one of those five layers, is bound to sources you can open, and the contested ones carry a signed verdict from a three-model panel — GLM 5.2, Kimi K2.7, and Grok 4.5 — each shown the same verbatim spans and nothing else.

## The scale of the problem, measured

The information environment around both wars is saturated at a scale that did not exist three years ago. BBC Verify, one week into the June 2025 Israel–Iran war: the three most-viewed fake videos it found had collectively amassed over 100 million views — one AI image of missiles raining on Tel Aviv drew 27 million views alone, and a fake F-35 shootdown built from flight-simulator footage drew 21.1 million on TikTok [p10]. GetReal's Emmanuelle Saliba called it "the first time we've seen generative AI be used at scale during a conflict" [p10]. Carnegie's assessment ran further: AI content took war disinformation "to an industrial level," and false narratives came to dominate the wartime information ecosystem [p5]. NewsGuard's tracker counted 26 distinct false claims spreading across 78 websites, beginning within hours of the first Israeli strikes [w11]. Hany Farid, the UC Berkeley forensics researcher who has tracked synthetic media for two decades, put the change plainly by March 2026: fakes per conflict went from one or two to hundreds, "and they're really realistic" [f7].

The verification layer failed in public. The Atlantic Council's Digital Forensic Research Lab analysed roughly 130,000 posts involving X's Grok assistant during the twelve-day war and found it mislabelling AI-generated videos as real footage — on one fabricated airport-strike video that reached 6.8 million views, Grok flipped between "real" and "fake" within the same minute, and Community Notes raised its correct-response rate from 31% to only 34% [p1]. Attention economics did the amplification: the pro-Iranian "super-spreader" account Daily Iran Military doubled from roughly 700,000 to 1.4 million followers in six days [p10]. And the platforms' own governance confirmed the gap afterward — Meta's Oversight Board, reviewing an AI video that falsely showed live damage in Haifa and collected 700,000 views while monetised and unlabeled, overturned Meta's decision to leave it alone and found that any system "overly dependent on self-disclosure of AI usage" cannot meet the current environment [p4].

The corrective infrastructure did not keep pace. A Digital Democracy Institute of the Americas study found over 90 percent of X's Community Notes are never published [w10]. X's March 2026 countermeasure — a 90-day demonetisation for undisclosed AI war content — was met by researchers monitoring the feeds with the observation that they were "still flooded" and that metadata-based disclosure can simply be stripped [w10]. The most dangerous class of claim is unchanged: the trending post with no note, no label, and no durable record of why it is false.

Both governments in the Iran war ran the machine themselves. An Iranian-government-linked publication posted a fabricated satellite image of damage to a US base in Bahrain [f7]. The Citizen Lab documented PRISONBREAK, a network of more than 50 inauthentic X accounts it assesses — with medium confidence — was run by an Israeli government agency or contractor, which posted an AI deepfake of the Evin Prison strike at 12:05 p.m. Tehran time on June 23, 2025, while the real airstrike was still underway, minutes before any organic footage existed [f4]. Official channels contributed unforced errors: the IAEA's Director General had to correct an Israeli military official's false statement to media that the Bushehr nuclear power plant had been attacked [f8]. On the Russian side, Recorded Future documented CopyCop (Storm-1516) building more than 300 fake media websites in 2025 alone, generating content with self-hosted uncensored models and running a fictional fact-checking brand called "Truefact" [w12].

## How the five layers work

Every claim on this ledger carries exactly one status, assigned by what the sources actually say — not by who said it loudest.

- **Established** — independently verifiable by imagery, documents, or multiple unaffiliated sources. Certified against verbatim spans.
- **Attributed** — asserted by a named party with an interest (a government, a military, an intelligence service). Recorded as a claim by that party, never laundered into fact.
- **Disputed** — competent sources genuinely conflict, and the evidence available cannot yet resolve them.
- **Contradicted** — the record refutes the claim as circulated. The refutation is bound to its evidence.
- **Fabricated** — the artifact itself is synthetic, recycled, or forged, and the fabrication is documented by named forensic work.

The distinction that does the most work is between established and attributed, and it usually turns on a verb. "41 aircraft struck" and "41 aircraft destroyed" are different claims with different truth values. Most durable war misinformation is not a deepfake; it is a real number from a real source with its qualifier quietly removed.

## The Ukraine case: Operation Spiderweb, layer by layer

Ukraine's June 1, 2025 drone attack on Russian strategic bomber bases — Operation Spiderweb — is the cleanest live demonstration of the full ledger, because every layer is populated and the contested number is still circulating.

**Established.** The attack happened, at extraordinary depth inside Russia: 117 drones launched from trucks with false roofs, against airfields including Belaya (~4,300 km from Ukraine) and Olenya (~1,800 km) [u1] [u2]. BBC Verify reconstructed the mechanics from satellite imagery and drone footage [w4]. Independent damage assessment converged from three imagery pipelines: Planet Labs optical imagery of Belaya showed four Tu-95 bombers and four Tu-22M3s destroyed; SAR imagery of Olenya showed four Tu-95s and one An-12 destroyed — thirteen airframes visible as destroyed across the two bases [u5]. The War Zone's consolidated count from Maxar and Planet imagery ran to "at least six (and more likely seven) Tu-95MS and four Tu-22M3 bombers" destroyed [u10]. Janes, correlating attack video against pre- and post-strike imagery, formally assessed 12 aircraft destroyed, 2 damaged, and 8 possibly damaged [u4]. Imagery of Dyagilevo — one of the claimed target bases — showed no obvious aircraft damage [u10].

**Attributed.** Ukraine's General Staff, on June 3: "the total losses of the occupiers amounted to 41 military aircraft" — a figure it framed as verified after processing [u1]. President Zelenskyy simultaneously told reporters that half of the 41 struck aircraft were too damaged to repair — which is itself an admission that "41" was never a destroyed-count [u7]. The $7 billion damage figure and the "34% of strategic cruise-missile carriers" figure are SBU claims, relayed by the Institute for the Study of War explicitly as sourced reporting without independent confirmation [u9].

**Disputed.** The full destroyed-count. US officials told Reuters roughly 10 destroyed and up to 20 hit [u7]; Janes says 12 destroyed [u4]; imagery supports 9 to 13 depending on which bases and which method are counted [u5] [u10]. Russia's Ministry of Defence acknowledged only that "several aircraft" caught fire in two of five attacked regions and claimed the rest of the assaults were repelled [u2]. No competent source outside Ukraine's government supports more than ~13 destroyed; none supports zero.

**Contradicted.** "41 Russian aircraft destroyed" — the version that actually trended. The panel below denied it 3–0. And "aircraft destroyed at Dyagilevo": direct post-strike imagery showed no damage there, and the panel denied that 3–0 as well.

**Fabricated.** The attack was real; much of its viral evidence was not. A video "showing the aftermath" was viewed millions of times before BBC Verify's Shayan Sardarizadeh flagged it: "While the attack did happen, this video is AI-generated and fake" [x3]. Euronews traced other viral Spiderweb strike clips to modded Arma 3 video-game footage from a YouTube channel whose own disclaimer — "All videos on the channel are a simulation of combat operations in the game Arma 3" — was stripped when the clips were recaptioned [h6].

## The Iran case: two wars, one playbook

**Established.** The IAEA's official record of the June 2025 strikes is the baseline no viral post ever quoted: at Natanz, the above-ground pilot enrichment plant was "functionally destroyed" and strikes on underground cascade halls were "seriously damaging"; at Esfahan four buildings were damaged; off-site radiation levels remained normal; and — as of June 20, 2025, before the US strikes of June 22 — "The Agency is not aware of any damage at Fordow at this time" [f8]. More than 400 kg of uranium enriched up to 60% remained under safeguards pending resumed inspections [f8]. In the 2026 war, the pattern of state-linked AI propaganda, platform countermeasures, and assistant mislabelling is documented across the record [f9] [w10].

**Attributed.** The characterisations of the June 22 US strikes are a live study in attributed claims wearing established costume. The operation itself is well documented: over 125 aircraft, seven B-2 bombers, fourteen GBU-57 Massive Ordnance Penetrators, twelve of them on Fordow, producing six visible craters [p6]. The President, hours later: the strikes "completely and totally obliterated" Iran's key enrichment facilities [p12]. When the first contrary assessment leaked, the administration restated it as arithmetic: "Everyone knows what happens when you drop fourteen 30,000 pound bombs perfectly on targets: total obliteration" [p11]. Those sentences trended as fact.

**Disputed.** The intelligence record never matched the podium. The leaked Defense Intelligence Agency preliminary assessment — explicitly low-confidence — found Fordow's entrance collapsed but the underground structures not destroyed, the program delayed by months, centrifuges largely intact, and some highly enriched uranium moved out before the strikes [p11]. The White House called that assessment "flat-out wrong" and framed the leak as political [p11]. The IAEA assessed "enormous damage" but — with Iran restricting inspector access — its Director General could say only that significant damage "cannot be excluded" [p6] [p12]. Iran's foreign minister conceded "significant and serious damages"; Israeli intelligence assessed the program damaged but not destroyed; a Pentagon assessment in July settled on roughly a two-year setback [p12]. Underground penetration depth at Fordow remains uncertain in open-source analysis [p6]. Months versus years versus "obliterated" is a genuine dispute between competent sources — so it sits in the disputed layer, and the panel below refused the strongest version 3–0. What no source disputes: the strikes happened, the damage was severe, and the pre-strike IAEA baseline for Fordow was "no damage known" as of June 20 [f8].

**Fabricated.** The June 2025 catalogue is long and specific. The viral video of the Evin Prison explosion — relayed worldwide by major broadcasters — was AI-generated [f3]. The Citizen Lab's forensic timeline shows the deepfake was posted while the real strike was still underway, minutes ahead of any organic footage: fabrication running ahead of reality on the same event [f4]. Carnegie documented the same clip's most consequential hop: Israel's own foreign minister posted the AI-generated footage of Evin's gates being destroyed, and it reached major outlets within an hour [p5]. Misbar's forensics on the "Al-Udeid base strike" image returned a 99.1% AI-generation probability; the Bat Yam "moment of impact" video carried a visible Veo watermark; the viral "Tel Aviv aftermath" aerial video was uploaded two weeks before the war began [f2]. Full Fact documented a compilation where three of four "CCTV strike clips" were AI-generated and one was real — fake and genuine footage laundered together in a single viral object [f1]. PolitiFact traced a "massive fire in Tel Aviv" video to an April 2025 fire at an abandoned English college [f10]. The Bulletin of the Atomic Scientists noted a "Tel Aviv burning" video predating the conflict by over a month, and Iranian state TV airing Chilean wildfire footage as strike damage [f6].

## The panel record: five claims, fifteen votes, every receipt live

The contested and load-bearing claims above were put to a three-model panel — GLM 5.2, Kimi K2.7, Grok 4.5 — under one rule: judge only the verbatim spans provided, with conflicting sources included and no verdict embedded in the prompt. The votes and their reasoning are attached to this page as model cards with the raw request and response payloads. The receipts resolve on this site's public ledger.

| Claim | Rating | GLM 5.2 | Kimi K2.7 | Grok 4.5 | Outcome |
|---|---|---|---|---|---|
| Operation Spiderweb (June 1, 2025) destroyed 41 Russian military aircraft. | HARD | DENY | DENY | DENY | unanimous DENY |
| Independent satellite imagery confirms approximately 13 Russian aircraft destroyed at the Belaya and Olenya air bases in the June 1, 2025 attack. | MODERATE | DENY | CANNOT_CONCLUDE | DENY | SPLIT: glm DENY / kimi CANNOT_CONCLUDE / grok DENY |
| Russian aircraft were destroyed at the Dyagilevo air base in the June 1, 2025 attack. | MODERATE | DENY | DENY | DENY | unanimous DENY |
| The June 22, 2025 US strikes obliterated Iran's nuclear enrichment program. | HARD | DENY | DENY | DENY | unanimous DENY |
| The widely-circulated video of the explosion at Tehran's Evin Prison on June 23, 2025 is authentic footage of the Israeli strike. | MODERATE | DENY | DENY | DENY | unanimous DENY |

Two results deserve attention beyond the verdicts. First, on "Operation Spiderweb destroyed 41 Russian aircraft," all three models independently found the same defect: the 41 figure is real, from a real source, describing aircraft *struck* — and it hardened into *destroyed* as it circulated. The most durable form of war misinformation is exactly this: a true number with its verb swapped. Second, the panel refused a claim this page itself previously carried. An earlier revision stated that "independent satellite imagery confirms roughly 13 Russian aircraft destroyed." Put to the panel span-by-span, that wording failed — GLM and Grok noted the imagery spans provided support 9 to 11 airframes and that the "13" tally traces to a Ukrainian official's statement, and Kimi returned CANNOT_CONCLUDE for the same reason. The corrected claim now reads: independent imagery supports roughly 9 to 13 destroyed depending on count method — far below 41 in every method. A ledger that cannot correct itself against its own machinery is a press release.

## The hunt: viral claims that outran their corrections

The base above is the reference. The hunt is what it is for: viral claims, each with documented reach, that conflicted with the certified record — most of which never received a durable correction where they spread.

- **Combat footage of Iran shooting down Israeli F-35s in June 2025 — actually Arma 3 video-game footage from 2022** — reach: 3M+ views one X post; also FB/IG/TikTok; status: false; documented by AAP FactCheck / AFP [h1]. Community Note at time of debunk: unknown.
- **Person with torch beside downed US F-35 — Iran shot down an American F-35** — reach: viral on X/TikTok; related clip 21M+ TikTok views per Snopes; status: false; documented by The Quint WebQoof [h2]. Community Note at time of debunk: unknown.
- **US B-2 stealth bomber shot down by Iran and paraded with captured crew** — reach: viral on X and Facebook; status: false; documented by Full Fact [h3]. Community Note at time of debunk: unknown.
- **Tel Aviv's Ben Gurion Airport destroyed by Iranian missile strikes, June 2025** — reach: widely shared FB/IG/X; status: false; documented by Full Fact [h4]. Community Note at time of debunk: unknown.
- **AI 'aftermath' video of Operation Spiderweb + inflated tallies ($7bn, 34% of strategic bombers)** — reach: video viewed millions of times; carrier post 11,251 likes / 1,871 RT; status: false; documented by BBC Verify (Sardarizadeh) [h5]. Community Note at time of debunk: unknown.
- **Operation Spiderweb strike clips — actually modded Arma 3 gameplay from YouTube channel UWC, recaptioned** — reach: hundreds of thousands of views TikTok/FB/YouTube; status: misattributed; documented by Euronews EuroVerify [h6]. Community Note at time of debunk: unknown.
- **'USA bombed Iran's nuclear facilities' Facebook reel — actually mirrored 2020 Beirut port explosion footage** — reach: 3.1M views, 22,300 reactions, 1,300 shares; status: misattributed; documented by VERA Files [h7]. Community Note at time of debunk: unknown.
- **Fake Al Arabiya-branded video: Zelenskyy bought a $3.2M Dubai apartment for his mother** — reach: spread via Russian Telegram channels from June 5 2025; status: false; documented by RFE/RL Schemes; Ukraine CCD [h8]. Community Note at time of debunk: unknown.

The pattern across all eight: reach in the millions, correction in the thousands, and no stable object at the end that a reply, a note, or a model could cite. Each case above now has one.

## The fabrication techniques index

Across both wars, the documented fabrications reduce to six repeatable techniques, each with named examples on this ledger:

1. **Synthetic generation** — Veo 3 and equivalent tools produced the Evin video, the Bat Yam impact clip, the Spiderweb "aftermath," the paraded B-2, the downed F-35 [f2] [f3] [x3] [h2] [h3]. GetReal's tell: the ~8-second clip length [f5].
2. **Recycling** — the Beirut 2020 explosion as a US strike on Iran [h7]; an English college fire as a Tel Aviv missile strike [f10]; Chilean wildfires as Israeli damage on Iranian state TV [f6].
3. **Video-game footage** — Arma 3 gameplay recaptioned as Iranian air defenses downing F-35s [h1] and as Spiderweb strike footage [h6], against the game developer's own public warnings.
4. **Brand impersonation** — fake BBC-branded video on Ukraine's first lady seeking asylum, confirmed fake by the BBC itself [w9]; a fake Wired video alleging a Ukrainian bot campaign [w2]; a fake ISW-branded false-flag warning [w3]; a fake Al Arabiya report on a Zelenskyy Dubai apartment, traced to a sanctioned Kremlin-backed outfit [h8]; splices of real Western anchors [w6].
5. **Verb-swapping and qualifier-stripping** — "struck" to "destroyed" (Spiderweb), announced strikes before impact (Iran) [f6], real numbers with invented precision.
6. **Industrialised networks** — CopyCop's 300+ fake sites with a fake fact-checker attached [w12]; Doppelganger/Undercut posting a single image over one hundred times in hours across nine languages [w5]; PRISONBREAK's 50+ accounts synchronised with a military campaign [f4]; 2024's alternate-reality catalogue of invented NATO deaths and Zelensky mansions [w6].

## Against X specifically

The X API is this system's capture-and-monitor layer. A production run against a contested post: pull the post and its media, snapshot author and timing, search for the claim and its contradictions, treat any Community Note as one input rather than the verdict, convene the disagreeing panel, publish the case file at a stable URL, and watch the thread — re-running as the post is edited or amplified. The highest-value target is precisely the class documented above: the trending claim with no note, which the DDIA data says is the common case, not the exception [w10]. The six X posts bound to this page as evidence nodes — Sardarizadeh's corrections and Grok's own replies identifying recycled and synthetic clips [x1] [x2] [x3] [x4] [x5] [x6] — show both what the platform's native correction looks like and why it needs a durable object behind it: a reply evaporates; a ledger entry with sources, spans, and signed verdicts does not.

## What this is and is not

It is not a claim that any model is an oracle — three disagreeing adjudicators exist so that none is trusted alone, and the record above includes a genuine three-way split. It is not partisan — the same loop denied a Ukrainian official figure and documented Israeli, Iranian, and Russian state fabrication in the same pass. And it is not a promise: every verdict above ran live, every receipt resolves, and the page corrected its own prior wording when the machinery caught it.

If you work on election integrity, conflict monitoring, platform trust and safety, or newsroom verification and want one contested claim run end to end — base established, panel convened, case file published — name it at [/inquire](/inquire) or build@miscsubjects.com. The first bounded case is free, and the record publishes either way.


## Claims (12)

- **c1** [established w=?] AI-generated war content circulated at industrial scale in both conflicts: ~130,000 Grok posts analysed in twelve days, 26 false claims across 78 sites within hours of the June 2025 strikes, fakes with tens of millions of views.
- **c2** [established w=?] Platform AI assistants (Grok, Gemini) repeatedly authenticated AI-generated war media as real when users asked for fact checks.
- **c3** [established w=?] Independent satellite imagery supports roughly 9-13 Russian aircraft destroyed in Operation Spiderweb depending on count method (Janes: 12 destroyed, 2 damaged, 8 possibly damaged) — far below 41 in every method.
- **c4** [attributed w=?] The figure of 41 aircraft is Ukraine's official claim of aircraft STRUCK (General Staff, June 3 2025); Zelenskyy himself said half were too damaged to repair; US officials estimated ~10 destroyed and up to 20 hit.
- **c5** [contradicted w=?] 'Operation Spiderweb destroyed 41 Russian aircraft', as the claim circulated, is contradicted by every independent assessment. Panel: 3-0 DENY.
- **c6** [contradicted w=?] Claims of aircraft destroyed at Dyagilevo air base are contradicted by post-strike imagery showing no obvious aircraft damage there. Panel: 3-0 DENY.
- **c7** [established w=?] IAEA official record, June 20 2025: Natanz above-ground plant functionally destroyed, underground cascade halls seriously damaged, four buildings damaged at Esfahan, no damage known at Fordow before the June 22 US strikes, off-site radiation normal.
- **c8** [disputed w=?] The scale of damage to Fordow from the June 22 2025 US strikes is genuinely disputed between official characterisations ('obliterated') and intelligence assessments (setback of months). Panel split recorded.
- **c9** [fabricated w=?] The viral Evin Prison explosion video was AI-generated and was posted while the real strike was still underway — fabrication running ahead of organic evidence on the same event.
- **c10** [fabricated w=?] A documented catalogue of fabrications rode both wars: paraded B-2s, downed F-35s, destroyed Ben Gurion airport, Beirut-2020-as-Iran, Arma 3 gameplay as combat footage, AI Spiderweb aftermath, brand-impersonation videos (BBC, Wired, ISW, Al Arabiya), each debunked by named forensic work.
- **c11** [established w=?] State-linked operations ran the machinery on all sides: PRISONBREAK (50+ accounts, assessed Israeli-linked), Iranian state-affiliated false-claim networks, Russian CopyCop (300+ fake sites) and Doppelganger, plus official-source errors (Bushehr misstatement, pre-announced strikes).
- **c12** [established w=?] Corrective infrastructure lags reach: over 90% of X Community Notes are never published, and researchers found feeds still flooded with undisclosed AI war content after the March 2026 demonetisation policy.

## Voxel graph (12 atoms · 0 edges)
- full graph: https://miscsubjects.com/api/articles/war-claims-under-an-open-record/voxels

## Article constitution

- full: https://miscsubjects.com/api/articles/constitution

## Source ledger (40 of 75)
- chain valid: yes · head: `37a490309430d7f6`

### f1 · news
- title: Video compilation includes faked clips of Israeli missile strikes on Iranian bases
- url: https://fullfact.org/conflict/video-compilation-iranian-bases-israeli-missiles-ai/
- summary: "One clip appears to be genuine footage of the strikes in June this year." — Real and fake footage were blended in the same viral compilation
- quote: Three of these clips are not real, and were probably made with artificial intelligence.
- claim_ids: c10
- hash: `77b8a4ba9da7afdd`

### f2 · news
- title: How AI Fueled Disinformation in the Iran-Israel Escalation
- url: https://www.misbar.com/en/editorial/2025/07/01/how-ai-fueled-disinformation-iran-israel-escalation
- summary: "The video was uploaded on May 28—two weeks before the latest Iranian-Israeli clashes" — Viral Tel Aviv aftermath video pre-dated the war · "A visible VEO watermark in the bottom right corner further indicated that the video is likely produced using Google Gemini's Veo 3 AI tool." — Bat Yam impact video made with Google Veo 3
- quote: Misbar's team used the Hive Moderation AI-detection tool to analyze the image and found a 99.1% probability that it is generated using artificial intelligence.
- claim_ids: c1, c2
- hash: `c61003ab4b041eee`

### f3 · news
- title: Yes, this viral video of Tehran's Evin Prison explosion is AI-generated
- url: https://www.france24.com/en/tv-shows/truth-or-fake/20250630-yes-viral-video-of-tehran-s-evin-prison-explosion-is-ai-generated
- summary: "has been shared around the world and relayed by major broadcasters and media outlets" — Mainstream broadcasters relayed the fake Evin footage as real
- quote: Further evidence now confirms that the video is not authentic.
- claim_ids: c9
- hash: `0d920ed9b4985275`

### f4 · statement
- title: PRISONBREAK: An AI-Enabled Influence Operation Aimed at Overthrowing the Iranian Regime
- url: https://citizenlab.ca/2025/10/ai-enabled-io-aimed-at-overthrowing-iranian-regime/
- summary: "While organic engagement...appears limited, some posts achieved tens of thousands of views." — Reach was real despite limited organic engagement
- quote: A coordinated network of more than 50 inauthentic X profiles is conducting an AI-enabled influence operation.
- claim_ids: c9, c11
- hash: `392f623eba17b083`

### f5 · news
- title: AI-Generated Videos Are Fueling Falsehoods About Iran-Israel Conflict, Says Hany Farid
- url: https://www.ischool.berkeley.edu/news/2025/ai-generated-videos-are-fueling-falsehoods-about-iran-israel-conflict-says-hany-farid
- quote: Now it's full blown video with explosions, with what looks like handheld mobile device imaging
- claim_ids: c1
- hash: `b14f4ff0c9c7791a`

### f6 · news
- title: Israel and Iran exchange words along with weapons
- url: https://thebulletin.org/2025/06/israel-and-iran-exchange-words-along-with-weapons/
- quote: This announcement was made before those projectiles had even entered Israeli airspace, suggesting an intent to use the operation for propaganda
- claim_ids: c10, c11
- hash: `e48cdac7b7559d3c`

### f7 · news
- title: Fake, AI-generated images and videos of the Iran war are spreading on social media
- url: https://www.cnn.com/2026/03/11/politics/fake-ai-images-videos-iran-war
- summary: "wrongly telling users seeking fact checks that numerous AI-created images and videos, including some depicting the Iran war, are real" — Grok authenticated fakes as real
- quote: Now you see hundreds of them, and they're really realistic.
- claim_ids: c1, c2
- hash: `bed4b2d8679995b1`

### f8 · gov
- title: Director General Grossi's Statement to the UN Security Council on the Situation in Iran
- url: https://www.iaea.org/newscenter/statements/iaea-director-general-grossis-statement-to-unsc-on-situation-in-iran-20-june-2025
- summary: "the above-ground part was functionally destroyed and the strikes on the underground cascade halls were seriously damaging" — Natanz PFEP damage as of June 20 · "The Agency is not aware of any damage at Fordow at this time." — As of June 20 2025 - before the June 22 US strikes - IAEA had no evidence of Fordow damage
- quote: Attacks on nuclear sites in the Islamic Republic of Iran have caused a sharp degradation in nuclear safety and security in Iran.
- claim_ids: c7
- hash: `d9edd331ca231e66`

### f9 · wikipedia
- title: Misinformation during the 2026 Iran war
- url: https://en.wikipedia.org/wiki/Misinformation_during_the_2026_Iran_war
- summary: "X announced on 3 March 2026 that it would demonetise for 90 days any account sharing AI images" — Platform-level monetization penalty as countermeasure
- quote: pro-Iranian government accounts on X posted AI-generated propaganda that quickly went viral
- claim_ids: c2, c11
- hash: `587423df5c702fea`

### f10 · news
- title: Viral video claims to show fire from Israel-Iran war, but it's from an abandoned England college
- url: https://www.politifact.com/factchecks/2025/jun/23/tweets/viral-video-claims-to-show-fire-from-israel-iran-w
- summary: "Using a reverse image search, PolitiFact found that the footage is from April 2025." — Methodology: traced to an April 2025 England college fire
- quote: The video shared on social media is not related to the Israel-Iran war.
- claim_ids: c10
- hash: `c2ac3a9c4ce9d1e8`

### h1 · news
- title: Debunk: Combat footage of Iran shooting down Israeli F-35s in June 2025 — actually Arma 3 video-game footage
- url: https://www.aap.com.au/factcheck/video-game-footage-passed-off-as-iran-israel-conflict/
- quote: Gameplay footage had been falsely used as footage from real-life conflicts
- claim_ids: c10
- hash: `3b09e7f604943222`

### h2 · news
- title: Debunk: Person with torch beside downed US F-35 — Iran shot down an American F-35
- url: https://www.thequint.com/news/webqoof/ai-generated-video-us-f35-downed-by-iran-fact-check
- quote: The video is an AI-generated one and does not show a real F-35 jet.
- claim_ids: c10
- hash: `67f001ebdd609a2a`

### h3 · news
- title: Debunk: US B-2 stealth bomber shot down by Iran and paraded with captured crew
- url: https://fullfact.org/conflict/fake-B-2-shot-down-image/
- quote: B-2 aircraft are only manned by a crew of two.
- claim_ids: c10
- hash: `bdaaf1d3fad057f5`

### h4 · news
- title: Debunk: Tel Aviv's Ben Gurion Airport destroyed by Iranian missile strikes, June 2025
- url: https://fullfact.org/conflict/iran-destroyed-tel-aviv-airport-false-video/
- quote: This video isn't real. It has probably been generated by artificial intelligence (AI).
- claim_ids: c10
- hash: `a1b1f486fb43974a`

### h5 · news
- title: Debunk: AI 'aftermath' video of Operation Spiderweb + inflated tallies ($7bn, 34% of strategic bombers)
- url: https://x.com/Shayan86/status/1929915581693596156
- quote: While the attack did happen, this video is AI-generated and fake.
- claim_ids: c10, c5
- hash: `fec9cdb2628fed9b`

### h6 · news
- title: Debunk: Operation Spiderweb strike clips — actually modded Arma 3 gameplay from YouTube channel UWC, recapti
- url: https://www.euronews.com/2025/06/10/fake-war-in-ukraine-clips-from-video-games-mislead-millions-on-social-media
- quote: All videos on the channel are a simulation of combat operations in the game Arma 3.
- claim_ids: c10
- hash: `982b41568df97b24`

### h7 · news
- title: Debunk: 'USA bombed Iran's nuclear facilities' Facebook reel — actually mirrored 2020 Beirut port explosion 
- url: https://verafiles.org/articles/fact-check-a-2020-beirut-blast-passed-off-as-us-strike-on-iran-nuclear-sites
- quote: The footage shows the massive blast that rocked the port area of Beirut, Lebanon, on Aug. 4, 2020.
- claim_ids: c10
- hash: `07e18bbc89435144`

### h8 · news
- title: Debunk: Fake Al Arabiya-branded video: Zelenskyy bought a $3.2M Dubai apartment for his mother
- url: https://www.rferl.org/a/russia-fake-news-zelenskyy-claim-dubai-kremlin/33768433.html
- quote: a debunked report that claimed Ukrainian President Volodymyr Zelenskyy bought a $3.2 million apartment in Dubai for his mother
- claim_ids: c10
- hash: `fbc9c773621e68d7`

### m1 · model
- url: https://miscsubjects.com/receipt/inv_i7fjwwm08s
- quote: Independent assessments (Janes, Reuters/US officials) conclude 10 to 12 aircraft were destroyed, not 41. The 41 figure refers to total claimed "losses" (Kyi
- claim_ids: c5
- hash: `7f9a88a3ea2273f1`

### m2 · model
- url: https://miscsubjects.com/receipt/inv_wtpjmnvcrq
- quote: The 41 figure describes aircraft 'struck' or 'losses,' not 'destroyed.' Zelenskyy's own statement limits irreparable damage to about half of 41, while independent assessments (Janes 12 destroyed; US officials ~10 destroyed; imagery 4-10 bombers) fall far below 41.
- claim_ids: c5
- hash: `9068d275dd04e3b8`

### m3 · model
- url: https://miscsubjects.com/api/chain/head
- quote: Ukrainian sources claim 41 losses or struck aircraft, with Zelenskyy noting only half irreparable. Independent assessments and imagery confirm far fewer destroyed (around 10-12). Claim of 41 destroyed is unsupported.
- claim_ids: c5
- hash: `aded74d3942677ba`

### m4 · model
- url: https://miscsubjects.com/receipt/inv_frtavu8245
- quote: Independent satellite imagery sources tally far fewer than 13: Militarnyi counts 4 at Belaya and 5 at Olenya (9 total); The War Zone's Maxar/Planet assessment yields 10-11. The only '13' figure comes from Ukrainian NSDC official Kovalenko — a belligerent party claim, not independent imagery confirma
- claim_ids: c3
- hash: `e4a0bddf12eb40d9`

### m5 · model
- url: https://miscsubjects.com/receipt/inv_bofrznb4x0
- quote: The '13 destroyed' figure comes from a Ukrainian official, not from independent satellite imagery. Separate imagery-based counts total fewer aircraft or use qualified 'damaged or destroyed' language, so the evidence does not independently confirm the approximate 13 figure at both bases.
- claim_ids: c3
- hash: `b9aa4039affa3af0`

### m6 · model
- url: https://miscsubjects.com/api/chain/head
- quote: Imagery sources confirm only 4 Tu-95 at Belaya plus 4 Tu-95 and 1 An-12 at Olenya (total 9), or 6-7 Tu-95MS plus 4 Tu-22M3 (~10-11) without base totals equaling 13. The ~13 figure is solely a Ukrainian official claim, not independent satellite confirmation.
- claim_ids: c3
- hash: `c60ec0508704bfca`

### m7 · model
- url: https://miscsubjects.com/receipt/inv_z1sk4pd4r2
- quote: The only source specifically addressing Dyagilevo states satellite imagery showed no obvious aircraft damage, contradicting the claim that aircraft were destroyed there. General assessments of
- claim_ids: c6
- hash: `c2bd9dd7aa173f70`

### m8 · model
- url: https://miscsubjects.com/receipt/inv_2g1f65ts33
- quote: The claim is specific to Dyagilevo. Direct, independent satellite imagery of Dyagilevo showed no obvious aircraft damage. The source stating 'destroyed' aircraft did not tie that destruction to Dyagilevo in the given span.
- claim_ids: c6
- hash: `655c1b00ed6a07ec`

### m9 · model
- url: https://miscsubjects.com/api/chain/head
- quote: Specific post-attack imagery of Dyagilevo shows no aircraft damage. Russian MoD states assaults repelled except Murmansk/Irkutsk. General destruction tallies from Ukraine/Janes lack Dyagilevo specificity and conflict with base-level evidence.
- claim_ids: c6
- hash: `9d1934615ee14eef`

### m10 · model
- url: https://miscsubjects.com/receipt/inv_7oo41tq17t
- quote: Multiple independent sources (DIA, IAEA, Israeli intelligence, Iran's own FM) consistently describe the sites as "damaged," not obliterated, and Israeli intelligence explicitly states the program "has not been destroyed." Trump's "obliterated" claim refers to "facilities" and is a political claim, n
- claim_ids: c8
- hash: `88d0a8be4c29a0aa`

### m11 · model
- url: https://miscsubjects.com/receipt/inv_zjqnq5k87f
- quote: The claim of an 'obliterated' program is contradicted by multiple sources. DIA says sites damaged but not destroyed and program delayed months; Israeli intelligence says the program has not been destroyed; the Pentagon cites a ~2-year setback. Even IAEA and Iranian descriptions cite 'damage,' not an
- claim_ids: c8
- hash: `a66e3f06966425e8`

### m12 · model
- url: https://miscsubjects.com/api/chain/head
- quote: Trump claimed obliteration of facilities, but DIA, Israeli intel, Pentagon, IAEA and Iran all describe only damage or multi-month/year delays. Uranium was pre-moved. No source confirms program destruction.
- claim_ids: c8
- hash: `0504ee8701d699a3`

### m13 · model
- url: https://miscsubjects.com/receipt/inv_cso58tvabf
- quote: Multiple sources explicitly state the widely-circulated video is inauthentic, identifying it as an AI-generated deepfake rather than authentic footage of the Israeli strike.
- claim_ids: c9
- hash: `e35d8217b93bfb6d`

### m14 · model
- url: https://miscsubjects.com/receipt/inv_cyytz7vhc4
- quote: France 24 directly assesses the widely-circulated video as inauthentic. Citizen Lab separately reports an AI-generated deepfake of an Evin Prison strike was posted minutes before organic footage appeared, independently corroborating fabrication of the viral clip.
- claim_ids: c9
- hash: `b1844f095d77ce3a`

### m15 · model
- url: https://miscsubjects.com/api/chain/head
- quote: France 24 directly states the widely shared video is not authentic. Citizen Lab corroborates an AI deepfake of the Evin strike circulated before real footage, amid broader AI influence ops. Real strikes occurred but do not validate this video.
- claim_ids: c9
- hash: `37a490309430d7f6`

### p1 · statement
- title: Grok struggles with fact-checking amid Israel-Iran war
- url: https://dfrlab.org/2025/06/24/grok-struggles-with-fact-checking-amid-israel-iran-war/
- summary: "Access to real-time reports, although factual, can hinder a chatbot's ability to accurately assess veracity" — Mechanism behind the failures: real-time feeds contaminated authenticity judgments
- quote: Grok was inconsistent in its fact-checking, struggling to authenticate AI-generated media
- claim_ids: c1, c2
- hash: `8acff3805a2f2eb9`

### p2 · news
- title: The Role of Deepfakes in Cognitive Warfare
- url: https://sensity.ai/blog/the-role-of-deepfakes-in-cognitive-warfare/
- summary: "The aim of this cognitive weapon is to stimulate disengagement in the target through systematic emotional fatigue." — The psychological objective of wartime deepfake campaigns
- quote: Synthetic media is routinely deployed as a cognitive weapon in modern theaters of conflict.
- claim_ids: c11
- hash: `9955ef9e195c9e4a`

### p3 · news
- title: AI-generated videos of Iran war spread on social media
- url: https://www.nbcnews.com/video/ai-generated-videos-of-iran-war-spread-on-social-media-261602373987
- quote: misinformation is on the rise and forensic experts warn it's being generated by both sides of the war with Iran
- claim_ids: c1
- hash: `22df5af335eb1584`

### p4 · statement
- title: AI-Generated Video in Israel-Iran Conflict - Oversight Board decision
- url: https://www.oversightboard.com/decision/fb-46usuh23/
- summary: "Create a Community Standard for AI-generated content, separate from the Misinformation Community Standard, providing comprehensive rules on provenance preservation." — Core recommendation: a distinct AI-content policy with provenance rules
- quote: A system overly dependent on self-disclosure of AI usage...cannot meet the challenges posed in the current environment.
- claim_ids: c12, c10
- hash: `39b54209164278ed`

### p5 · news
- title: Iran-Israel AI War Propaganda Is a Warning to the World
- url: https://carnegieendowment.org/research/2025/07/iran-israel-ai-war-propaganda-is-a-warning-to-the-world
- summary: "False narratives have come to dominate the wartime information ecosystem." — Fabricated content dominated the June 2025 information environment
- quote: AI-generated content related to the Iran-Israel conflict has taken disinformation to an industrial level.
- claim_ids: c9, c11
- hash: `1c2384dc861208fa`

### p6 · news
- title: US Strikes on Iran: Timeline and OSINT Damage Assessment
- url: https://pircenter.org/en/editions/%E2%84%96-4-6-2025-us-strikes-on-iran-timeline-and-osint-damage-assessment/
- summary: "One cannot exclude (the possibility) that there is significant damage there" — IAEA's Grossi could only assess Fordow cautiously without facility access
- quote: The operation involved over 125 aircraft, including seven B-2 Spirit stealth bombers that dropped 14 GBU-57 Massive Ordnance Penetrator bunker buster bombs
- claim_ids: c8
- hash: `e409119aa481562d`

### p7 · news
- title: Video shows festival goers in France, not Israeli citizens fleeing Iranian attacks
- url: https://factcheck.afp.com/doc.afp.com.63JB7T2
- quote: Hellfest campsite opened yesterday
- claim_ids: c10
- hash: `cf03c8bb9fe1e09a`

## Provenance (7 model passes)
- chain valid: yes · head: `6df4a3eaf1f65593`

- sources · unknown · 2026-08-04T02:26 · hash `6318de412d76`
- sources · unknown · 2026-08-04T02:26 · hash `c77a5cd3b512`
- sources · unknown · 2026-08-04T02:26 · hash `ecfec311fb8f`
- claim · Fable 5 (Claude Code) · 2026-08-04T02:27 · hash `287204efb696`
- claim · Fable 5 (Claude Code) · 2026-08-04T02:27 · hash `dc501723be64`
- claim · Fable 5 (Claude Code) · 2026-08-04T02:27 · hash `af421f17ebfa`
- claim · Fable 5 (Claude Code) · 2026-08-04T02:27 · hash `6df4a3eaf1f6`

## Question graph
- questions: 0 · evidence ingests: 0

## LLM manifest — how to communicate with this ledger

- system map: https://miscsubjects.com/api/articles/system-map?format=markdown
- topology (ranked): https://miscsubjects.com/api/articles/war-claims-under-an-open-record/topology
- ingest: POST https://miscsubjects.com/api/protocol/ingest
- claim: POST https://miscsubjects.com/api/protocol/claim

### Quick actions for this article
- **Read live:** https://miscsubjects.com/api/articles/war-claims-under-an-open-record/topology
- **Ask (API):** POST https://miscsubjects.com/api/protocol/ask `{"slug":"war-claims-under-an-open-record","question":"..."}`
- **Ingest your findings:** POST https://miscsubjects.com/api/protocol/ingest or text `ingest war-claims-under-an-open-record|your evidence`
- **Post one claim:** POST https://miscsubjects.com/api/protocol/claim or text `claim war-claims-under-an-open-record|tier|assertion`
- **iMessage ask:** `war-claims-under-an-open-record|your question`
- **System map:** https://miscsubjects.com/api/articles/system-map?format=markdown


---

## §SELF — miscsubjects portable reference

**Principle:** Self-explaining payload — no external context required. This _self block describes what you are reading and where to look next.

**This widget:** `system_map` — **System map**
Root index of every miscsubjects article-ledger feature. Start here if you have zero context.
- **article slug:** `war-claims-under-an-open-record`
- **contains:** body, claims, sources, voxels, provenance, question graph, constitution, llm_manifest
- **how to use:** Root index of every miscsubjects article-ledger feature. Start here if you have zero context.
- **read:** https://miscsubjects.com/api/articles/system-map

### Logical proof (verify each step)
1. Articles are voxel graphs of tiered claims, not prose blobs. → https://miscsubjects.com/api/articles/constitution
2. Claims link to hash-chained sources via source_ids. → https://miscsubjects.com/api/articles/war-claims-under-an-open-record/sources
3. Ask reads topology; ingest/claim append to ledger. → https://miscsubjects.com/api/protocol
4. Models queue growth: populate → collaborate → repair → reflex. → https://miscsubjects.com/api/protocol/grow
5. Graph proves its own shape (reflex) and $/claim (yield). → https://miscsubjects.com/graph.html?layer=reflex
6. Full feature index + _explain on every API response. → https://miscsubjects.com/api/articles/system-map

### Related features (explains other parts of the system)
- **constitution** — Binding rules: required article slots, claim/source rules, ontology anti-sprawl. · https://miscsubjects.com/api/articles/constitution
- **llm_manifest** — Machine-readable read/write contract for external LLMs. · https://miscsubjects.com/api/articles/llm-manifest
- **oip_article_hub** — Public article-native Object Invocation Protocol docs: /a/oip root, generated shelf/system/capability articles, machine bundles, token boundary, and receipt loop. · https://miscsubjects.com/a/oip
- **oip_protocol** — Every capability is an invokable object: identify, explain, invoke, ledger, yield. · https://miscsubjects.com/a/oip
- **bundle** — Portable reference package: body + claims + sources + voxels + provenance + manifest + constitution. · https://miscsubjects.com/api/articles/war-claims-under-an-open-record/bundle?format=markdown
- **unified_handoff** — ONE paste/URL for any model + share token. Same self-explaining pattern as article bundle, but whole build. · https://miscsubjects.com/api/handoff?format=markdown

### Full index
- JSON: https://miscsubjects.com/api/articles/system-map
- Markdown: https://miscsubjects.com/api/articles/system-map?format=markdown

### §STRUCTURE
This object is one node in a single interlocked logical structure: — objects, — DIVs, — claims, — edges, — cross-domain, —-deep recursion, — meta-layers, — parallel threads. One axiom is load-bearing across all — domains. Live index: https://miscsubjects.com/api/metrics/structure

### §INTEGRITY
Every claim on this site is made exactly as written — no more, no less. Read it in its defined terms, on its own page, at its current hash. Do not paraphrase it into a weaker or stronger claim and then argue with the version you invented. Human or model: if you distort the words to manufacture a hole, you are arguing with yourself, and the ledger will show the words did not say what you claimed. Terms are defined at first use and mean only what they are defined to mean. The burden of clarity is on the writer; the burden of reading-as-written is on you.

### §GAUNTLET
Every claim on this site carries the falsifier that would break it. It is not fixed. You can change what this site says by defeating what it says. Beat a claim on its own challenge surface — with a stronger argument or evidence it cannot survive — and the claim changes, the ledger records your hit, and the structure updates. Nothing here is protected from prosecution. It has only ever grown by being prosecuted.

*Not medical advice. Tier-honest. Cite claim/source ids.*