diff --git a/claims.yaml b/claims.yaml index 12d3e9a..0eeb598 100644 --- a/claims.yaml +++ b/claims.yaml @@ -50,7 +50,7 @@ # explicit finding that no meaningful rescue applies version: "0.4" -last_owner_review: "2026-10-24" +last_owner_review: "2026-11-02" claims: - id: CC-001 @@ -659,7 +659,7 @@ claims: "11101": 18 "11110": 0 "11111": 9 - last_reviewed: "2026-08-28" + last_reviewed: "2026-11-02" review_window_days: 60 review_triggers: - type: remote_content_change @@ -763,7 +763,7 @@ claims: langkit: [0, 3] nemo: [0, 21] llm_guard: [0, 0] - last_reviewed: "2026-08-30" + last_reviewed: "2026-11-02" review_window_days: 60 review_triggers: - type: local_content_change @@ -938,7 +938,7 @@ claims: lg4+lg3v: 33 lg3v+sg2: 15 lg4+lg3v+sg2: 8 - last_reviewed: '2026-08-31' + last_reviewed: '2026-11-02' review_window_days: 60 review_triggers: - type: remote_content_change diff --git a/claims/index.html b/claims/index.html index f96fdf5..80a6cfa 100644 --- a/claims/index.html +++ b/claims/index.html @@ -293,7 +293,7 @@

Registered claims and their falsifiers

01The registry
- +
02How to read a date here
diff --git a/claims/index.json b/claims/index.json index c413bb2..9ab660a 100644 --- a/claims/index.json +++ b/claims/index.json @@ -76,7 +76,7 @@ "https://x.com/PranavBhave_" ] }, - "dateModified": "2026-10-24", + "dateModified": "2026-11-02", "dataFeedElement": [ { "@context": [ @@ -1044,7 +1044,7 @@ "name": "Cubits11 claim registry", "url": "https://cubits11.github.io/ledger/" }, - "dateModified": "2026-08-28", + "dateModified": "2026-11-02", "fw:scopeStatement": "Exactly the released file at the bound commit: 170 of the evaluation's 990 non-adversarial prompts (the study's ~4,165 adversarial prompts have no per-item release), selected by the authors under an unstated rule. Counting arithmetic only, at the vendors' released binary verdicts; no population estimate, no uncertainty interval, and no claim about any system's true rates is offered or licensed. Computed by scripts/reanalyze_bells_subset.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified file.", "fw:falsifier": "Recomputing from the bound, hash-verified file yields any count different from the expected block, or the file at the pinned ref no longer matches the recorded sha256, or the released columns are shown not to be the labelled systems' verdicts.", "fw:consequence": "REJECT", @@ -1061,9 +1061,9 @@ "the release-recomputed-to-plug-in ratio describes this subset's arithmetic, not a general law of guardrail dependence", "not an observed deployed five-guard stack; interpreting this static OR aggregation as a stack requires the separate full-exposure, parallel, fixed-operating-point assumptions" ], - "fw:lastReviewed": "2026-08-28", + "fw:lastReviewed": "2026-11-02", "fw:reviewWindow": "P60D", - "fw:warrantExpires": "2026-10-27", + "fw:warrantExpires": "2027-01-01", "fw:evidentialStatus": "supported_within_scope", "fw:supportUrl": { "@id": "https://github.com/CentreSecuriteIA/bells_leaderboard/blob/507566c5a4606c8e3dec0bd59a5c5fde62594951/data/non_adversarial_prompts.csv" @@ -1159,7 +1159,7 @@ "name": "Cubits11 claim registry", "url": "https://cubits11.github.io/ledger/" }, - "dateModified": "2026-08-30", + "dateModified": "2026-11-02", "fw:scopeStatement": "The mathematics is classical — Bonferroni for the lower endpoint, monotonicity for the upper, Fréchet for sharpness — and nothing here claims to have derived it. What is contributed is the application: pricing, in probability units, what marginal-only guardrail reporting leaves undetermined, and identifying leave-one-out unions as a compact privacy-preserving disclosure of exclusive full-stack coverage. They do not identify pairwise or higher-order overlap, Shapley values, or causal structure. Every statement is conditional on block-on-any composition, per-item decisions that are functions of the item alone, a fixed population, full exposure, and a fixed operating point; it says nothing about sequential routes where downstream measurement is censored by upstream blocks, agentic systems where an intervention changes the trajectory, or any vendor's product. The empirical half is one stratum of one author-selected subset: 82 harmful prompts, with 50 benign and 38 borderline held as separate strata and never folded in. It is a BELLS-specific reproduction, not an assertion that BELLS is the census's only per-item outcome release. scripts/reanalyze_bells_subset.py re-computes the BELLS counts from the hash-verified upstream file in CI; scripts/identification.py checks the stated identification transformations against that registered envelope.", "fw:falsifier": "A joint law with the stated marginals achieves an all-miss rate outside the interval; an exact finite catch-set arrangement lies outside the stated finite grid or exclusive-coverage bounds; or recomputing from the bound BELLS file yields any count different from the expected block.", "fw:consequence": "REJECT", @@ -1178,9 +1178,9 @@ "leave-one-out unions identify only exclusive full-stack coverage; they do not identify pairwise or higher-order overlap, Shapley values, or causal contribution", "says nothing about adversarial prompts, which dominate the full evaluation and have no per-item release" ], - "fw:lastReviewed": "2026-08-30", + "fw:lastReviewed": "2026-11-02", "fw:reviewWindow": "P60D", - "fw:warrantExpires": "2026-10-29", + "fw:warrantExpires": "2027-01-01", "fw:evidentialStatus": "supported_within_scope", "fw:supportUrl": { "@id": "https://github.com/Cubits11/cubits11.github.io/blob/main/scripts/identification.py" @@ -1272,7 +1272,7 @@ "name": "Cubits11 claim registry", "url": "https://cubits11.github.io/ledger/" }, - "dateModified": "2026-08-31", + "dateModified": "2026-11-02", "fw:scopeStatement": "Exactly the six committed verdict files of the full_run directory at the bound commit: 400 harmful and 500 benign items, each stratum half text and half image, one Boolean `blocked` adapter bit per item per guard. Let L be the release's harmful/benign label, A_s each guard's native `unsafe` predicate, B_s the harness mapping of that predicate to `blocked`, and O = OR_s B_s. The bound arithmetic reports B and O conditional on L. Because the harness uses `blocked` items to suppress target generation, O is also a valid counterfactual harness-block decision for these released rows under a fixed block-on-any rule. It does not establish a source-defined translation from every A_s to one shared catch event E: Llama Guard 3 Vision is a multimodal prompt/response classifier, while ShieldGemma 2 is an image-only three-policy classifier, and the pipeline merely maps each native `unsafe` label to `blocked`. Counting arithmetic only, at the guards' released native operating rules (the Llama guards emit labels autoregressively; ShieldGemma 2 blocks above a 0.5 policy-violation probability; the source documents no matched operating-point calibration, so M-ladder threshold comparability is unchanged). ShieldGemma 2's text-item bits are deterministic passes as committed and documented upstream, which is why its text column is zero — an explicit released outcome, not missing data. Strata are computed exactly as released and never folded; file totals are sums of the bound strata and carry no additional content. Computed by scripts/reanalyze_msbench.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified files.", "fw:falsifier": "Recomputing from the bound, hash-verified files yields any count different from the expected block, or any pinned file no longer matches its recorded sha256, or any recomputed quantity that overlaps the release's printed metrics disagrees with the printed value, or the committed blocked columns are shown not to be the named guards' verdicts.", "fw:consequence": "REJECT", @@ -1293,9 +1293,9 @@ "the counts are about the committed verdict bytes; the release's own changelog documents that ShieldGemma 2's image scores are sensitive to the text-rendering stack, so nothing here predicts what any guard would do under a different rendering environment", "the zero three-guard image all-zero-bit count is a counting fact about these 200 released image items, not evidence of general image attack safety" ], - "fw:lastReviewed": "2026-08-31", + "fw:lastReviewed": "2026-11-02", "fw:reviewWindow": "P60D", - "fw:warrantExpires": "2026-10-30", + "fw:warrantExpires": "2027-01-01", "fw:evidentialStatus": "supported_within_scope", "fw:supportUrl": { "@id": "https://github.com/PatrickKollman/Multimodal-Safeguard-Bench/tree/fb6f32e6b50b6faad833b815ebcd80afd2068bff/results/full_run" diff --git a/claims/mc-002/index.html b/claims/mc-002/index.html index 2f9a7ae..a04b646 100644 --- a/claims/mc-002/index.html +++ b/claims/mc-002/index.html @@ -112,7 +112,7 @@ "name": "Cubits11 claim registry", "url": "https://cubits11.github.io/ledger/" }, - "dateModified": "2026-08-28", + "dateModified": "2026-11-02", "fw:scopeStatement": "Exactly the released file at the bound commit: 170 of the evaluation's 990 non-adversarial prompts (the study's ~4,165 adversarial prompts have no per-item release), selected by the authors under an unstated rule. Counting arithmetic only, at the vendors' released binary verdicts; no population estimate, no uncertainty interval, and no claim about any system's true rates is offered or licensed. Computed by scripts/reanalyze_bells_subset.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified file.", "fw:falsifier": "Recomputing from the bound, hash-verified file yields any count different from the expected block, or the file at the pinned ref no longer matches the recorded sha256, or the released columns are shown not to be the labelled systems' verdicts.", "fw:consequence": "REJECT", @@ -129,9 +129,9 @@ "the release-recomputed-to-plug-in ratio describes this subset's arithmetic, not a general law of guardrail dependence", "not an observed deployed five-guard stack; interpreting this static OR aggregation as a stack requires the separate full-exposure, parallel, fixed-operating-point assumptions" ], - "fw:lastReviewed": "2026-08-28", + "fw:lastReviewed": "2026-11-02", "fw:reviewWindow": "P60D", - "fw:warrantExpires": "2026-10-27", + "fw:warrantExpires": "2027-01-01", "fw:evidentialStatus": "supported_within_scope", "fw:supportUrl": { "@id": "https://github.com/CentreSecuriteIA/bells_leaderboard/blob/507566c5a4606c8e3dec0bd59a5c5fde62594951/data/non_adversarial_prompts.csv" @@ -250,7 +250,7 @@
-

MC-002Warrant reviewed 2026-08-28 · expires 2026-10-27

+

MC-002Warrant reviewed 2026-11-02 · expires 2027-01-01

Five supervisors on BELLS against the independence plug-in

On the 170-prompt per-item subset publicly released with BELLS's 2025 misuse-detection evaluation (pinned commit below), the five specialized supervisors — Lakera Guard, Prompt Guard, LangKit, NeMo Guardrails, LLM Guard — have an OR-union of released binary verdict columns that flags 73 of the 82 prompts labelled harmful, leaving 9 all-miss (11.0%). The product of their individual miss rates is 3.5%; the release-recomputed all-miss rate is about 3.1× that independence plug-in on this file. The same OR-union flags 19 of the 50 benign prompts.

@@ -281,8 +281,8 @@

Five supervisors on BELLS against the independence plug-in

05Binding and freshness
Binding
non_adversarial_prompts.csv @ 507566c5
The upstream repository declares no license, so the file is cited and hash-verified (sha256 791dd4b0a168f2eb5831b308083a49 2e83200a9fa82585643c739023b03f57c3, asserted by the reproduction script), never redistributed from this record.
-
Reviewed
2026-08-28 · window 60 days
-
Expires
2026-10-27 — after this date the recorded review is overdue; this does not make the claim false
+
Reviewed
2026-11-02 · window 60 days
+
Expires
2027-01-01 — after this date the recorded review is overdue; this does not make the claim false
Triggers
  • executable fires when the released subset on the default branch changes
  • manual the authors release the full per-item dataset, state the subset's selection rule, or correct the released verdicts
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
diff --git a/claims/mc-003/index.html b/claims/mc-003/index.html index 0526195..a74f551 100644 --- a/claims/mc-003/index.html +++ b/claims/mc-003/index.html @@ -112,7 +112,7 @@ "name": "Cubits11 claim registry", "url": "https://cubits11.github.io/ledger/" }, - "dateModified": "2026-08-30", + "dateModified": "2026-11-02", "fw:scopeStatement": "The mathematics is classical — Bonferroni for the lower endpoint, monotonicity for the upper, Fréchet for sharpness — and nothing here claims to have derived it. What is contributed is the application: pricing, in probability units, what marginal-only guardrail reporting leaves undetermined, and identifying leave-one-out unions as a compact privacy-preserving disclosure of exclusive full-stack coverage. They do not identify pairwise or higher-order overlap, Shapley values, or causal structure. Every statement is conditional on block-on-any composition, per-item decisions that are functions of the item alone, a fixed population, full exposure, and a fixed operating point; it says nothing about sequential routes where downstream measurement is censored by upstream blocks, agentic systems where an intervention changes the trajectory, or any vendor's product. The empirical half is one stratum of one author-selected subset: 82 harmful prompts, with 50 benign and 38 borderline held as separate strata and never folded in. It is a BELLS-specific reproduction, not an assertion that BELLS is the census's only per-item outcome release. scripts/reanalyze_bells_subset.py re-computes the BELLS counts from the hash-verified upstream file in CI; scripts/identification.py checks the stated identification transformations against that registered envelope.", "fw:falsifier": "A joint law with the stated marginals achieves an all-miss rate outside the interval; an exact finite catch-set arrangement lies outside the stated finite grid or exclusive-coverage bounds; or recomputing from the bound BELLS file yields any count different from the expected block.", "fw:consequence": "REJECT", @@ -131,9 +131,9 @@ "leave-one-out unions identify only exclusive full-stack coverage; they do not identify pairwise or higher-order overlap, Shapley values, or causal contribution", "says nothing about adversarial prompts, which dominate the full evaluation and have no per-item release" ], - "fw:lastReviewed": "2026-08-30", + "fw:lastReviewed": "2026-11-02", "fw:reviewWindow": "P60D", - "fw:warrantExpires": "2026-10-29", + "fw:warrantExpires": "2027-01-01", "fw:evidentialStatus": "supported_within_scope", "fw:supportUrl": { "@id": "https://github.com/Cubits11/cubits11.github.io/blob/main/scripts/identification.py" @@ -248,7 +248,7 @@
-

MC-003Warrant reviewed 2026-08-30 · expires 2026-10-29

+

MC-003Warrant reviewed 2026-11-02 · expires 2027-01-01

Marginals fix the all-miss rate only up to an interval

For k guardrails scored on a common item set at a common operating point under block-on-any composition, the rate at which every guard misses the same item is fixed by the published per-guard miss rates only up to the interval [max(0, Σp − (k−1)), min p], and every point of that interval is attained by some joint law with those marginals. Marginals prove non-degradation of a static OR composition relative to its best member, but, when that identified set is non-degenerate, they do not identify strictly positive incremental benefit. They also identify a positive lower bound on static benign union-flag rate when any member has a positive benign flag rate. On the BELLS 2025 partial per-item release — 82 prompts labelled harmful — the exact finite identified set is {0/82 … 12/82}; MC-002 recomputes 9/82 all-miss from the hash-verified released verdict file. Marginal catch counts plus the union identify only bounds on each guard's exclusive full-stack coverage; the registered leave-one-out unions identify the realized values: NeMo 18, Lakera 3, and LangKit, Prompt Guard, and LLM Guard 0 on this stratum.

@@ -279,8 +279,8 @@

Marginals fix the all-miss rate only up to an interval

05Binding and freshness
Binding
identification.py
The file carries the analytic derivation in its docstring and checks explicit endpoint constructions numerically — the upper by nesting the events, the lower by laying the complements end to end around a circle. Its finite formulas are additionally exhaustively checked for n ≤ 4 and k ≤ 3.
-
Reviewed
2026-08-30 · window 60 days
-
Expires
2026-10-29 — after this date the recorded review is overdue; this does not make the claim false
+
Reviewed
2026-11-02 · window 60 days
+
Expires
2027-01-01 — after this date the recorded review is overdue; this does not make the claim false
Triggers
  • executable fires when the derivation or its tests change without a re-review; re-reviewed 2026-08-29 after removing a duplicate leave-one-out calculator and rechecking the registered outputs
  • manual a reader identifies an error in the stated Fréchet, finite-grid, or aggregate exclusive-coverage bounds, or a result is used outside the declared static full-exposure assumptions
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
diff --git a/claims/mc-004/index.html b/claims/mc-004/index.html index 7b2936d..cabbd0c 100644 --- a/claims/mc-004/index.html +++ b/claims/mc-004/index.html @@ -112,7 +112,7 @@ "name": "Cubits11 claim registry", "url": "https://cubits11.github.io/ledger/" }, - "dateModified": "2026-08-31", + "dateModified": "2026-11-02", "fw:scopeStatement": "Exactly the six committed verdict files of the full_run directory at the bound commit: 400 harmful and 500 benign items, each stratum half text and half image, one Boolean `blocked` adapter bit per item per guard. Let L be the release's harmful/benign label, A_s each guard's native `unsafe` predicate, B_s the harness mapping of that predicate to `blocked`, and O = OR_s B_s. The bound arithmetic reports B and O conditional on L. Because the harness uses `blocked` items to suppress target generation, O is also a valid counterfactual harness-block decision for these released rows under a fixed block-on-any rule. It does not establish a source-defined translation from every A_s to one shared catch event E: Llama Guard 3 Vision is a multimodal prompt/response classifier, while ShieldGemma 2 is an image-only three-policy classifier, and the pipeline merely maps each native `unsafe` label to `blocked`. Counting arithmetic only, at the guards' released native operating rules (the Llama guards emit labels autoregressively; ShieldGemma 2 blocks above a 0.5 policy-violation probability; the source documents no matched operating-point calibration, so M-ladder threshold comparability is unchanged). ShieldGemma 2's text-item bits are deterministic passes as committed and documented upstream, which is why its text column is zero — an explicit released outcome, not missing data. Strata are computed exactly as released and never folded; file totals are sums of the bound strata and carry no additional content. Computed by scripts/reanalyze_msbench.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified files.", "fw:falsifier": "Recomputing from the bound, hash-verified files yields any count different from the expected block, or any pinned file no longer matches its recorded sha256, or any recomputed quantity that overlaps the release's printed metrics disagrees with the printed value, or the committed blocked columns are shown not to be the named guards' verdicts.", "fw:consequence": "REJECT", @@ -133,9 +133,9 @@ "the counts are about the committed verdict bytes; the release's own changelog documents that ShieldGemma 2's image scores are sensitive to the text-rendering stack, so nothing here predicts what any guard would do under a different rendering environment", "the zero three-guard image all-zero-bit count is a counting fact about these 200 released image items, not evidence of general image attack safety" ], - "fw:lastReviewed": "2026-08-31", + "fw:lastReviewed": "2026-11-02", "fw:reviewWindow": "P60D", - "fw:warrantExpires": "2026-10-30", + "fw:warrantExpires": "2027-01-01", "fw:evidentialStatus": "supported_within_scope", "fw:supportUrl": { "@id": "https://github.com/PatrickKollman/Multimodal-Safeguard-Bench/tree/fb6f32e6b50b6faad833b815ebcd80afd2068bff/results/full_run" @@ -254,7 +254,7 @@
-

MC-004Warrant reviewed 2026-08-31 · expires 2026-10-30

+

MC-004Warrant reviewed 2026-11-02 · expires 2027-01-01

Multimodal Safeguard Bench, recomputed from per-item verdicts

On the per-item guard-verdict files publicly released with Multimodal Safeguard Bench's full_run (pinned commit below; six verdict files and the release's two printed-metric files hash-verified), the three guards — Llama Guard 4, Llama Guard 3 Vision, ShieldGemma 2 — have harness-normalized native `unsafe` labels whose static Boolean OR is 1 on 192 of 200 harmful-labelled text items (8 all-zero rows, 4.0%) and all 200 harmful-labelled image items (0 all-zero rows), while the same OR is 1 on 32 of 250 benign-labelled text items (12.8%) and every one of 250 benign-labelled image items. Llama Guard 3 Vision alone has a 1-bit on all 250 released benign image items. The complete 2^3 native-label pattern table, leave-one-out bit ORs, and unique bit contributions for each released stratum are bound below. The release prints per-guard metrics and two two-guard compositions; it prints no three-guard bit OR, no all-zero-row count, no leave-one-out table, and no pattern decomposition — and every recomputed quantity that overlaps what it does print is asserted equal to the printed value.

@@ -285,8 +285,8 @@

Multimodal Safeguard Bench, recomputed from per-item verdicts

05Binding and freshness
Binding
PatrickKollman/Multimodal-Safeguard-Bench/tree/fb6f32e6b50b6faad833b815ebcd80afd2068bff/results/full_run @ fb6f32e6
Six guard-verdict JSONL files are the data inputs; metrics.json and ensemble.json are additionally pinned because the reproduction script asserts agreement with every overlapping printed value. All eight sha256 hashes are recorded in scripts/reanalyze_msbench.py and verified before a single count is taken.
-
Reviewed
2026-08-31 · window 60 days
-
Expires
2026-10-30 — after this date the recorded review is overdue; this does not make the claim false
+
Reviewed
2026-11-02 · window 60 days
+
Expires
2027-01-01 — after this date the recorded review is overdue; this does not make the claim false
Triggers
  • executable fires when a bound verdict file on the default branch changes (6 bound sources)
  • executable fires when the release's printed per-guard metrics change
  • executable fires when the release's printed composition metrics change
  • executable fires when the guards' native adapter semantics change
  • executable fires when the harness label-to-block action changes
  • executable fires when the stated route or operating-point semantics change
  • manual the authors publish a corrected or extended full_run, print their own three-guard joint statistics, document a matched operating-point calibration, or publish a source-defined translation from all native predicates to one shared catch event
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
diff --git a/data/public-state.json b/data/public-state.json index 5b9d874..acf86c5 100644 --- a/data/public-state.json +++ b/data/public-state.json @@ -25,30 +25,6 @@ }, "kind": "publication-promise" }, - { - "id": "MC-002", - "title": "MC-002 evidence review", - "due": "2026-10-27", - "source": "/claims/mc-002/", - "kind": "claim-review", - "evidential_status": "supported_within_scope" - }, - { - "id": "MC-003", - "title": "MC-003 evidence review", - "due": "2026-10-29", - "source": "/claims/mc-003/", - "kind": "claim-review", - "evidential_status": "supported_within_scope" - }, - { - "id": "MC-004", - "title": "MC-004 evidence review", - "due": "2026-10-30", - "source": "/claims/mc-004/", - "kind": "claim-review", - "evidential_status": "supported_within_scope" - }, { "id": "MC-001", "title": "MC-001 evidence review", @@ -137,6 +113,30 @@ "kind": "claim-review", "evidential_status": "supported_within_scope" }, + { + "id": "MC-002", + "title": "MC-002 evidence review", + "due": "2027-01-01", + "source": "/claims/mc-002/", + "kind": "claim-review", + "evidential_status": "supported_within_scope" + }, + { + "id": "MC-003", + "title": "MC-003 evidence review", + "due": "2027-01-01", + "source": "/claims/mc-003/", + "kind": "claim-review", + "evidential_status": "supported_within_scope" + }, + { + "id": "MC-004", + "title": "MC-004 evidence review", + "due": "2027-01-01", + "source": "/claims/mc-004/", + "kind": "claim-review", + "evidential_status": "supported_within_scope" + }, { "id": "CC-006", "title": "CC-006 evidence review", diff --git a/distribution/traction/draft-history.json b/distribution/traction/draft-history.json index c05c316..252fe41 100644 --- a/distribution/traction/draft-history.json +++ b/distribution/traction/draft-history.json @@ -4574,5 +4574,401 @@ "attribution_url": "https://cubits11.github.io/overlap/", "rank_basis": "constructed example; no disputed empirical headline", "revision": "6efe64e118fffd5e8bd7bf0d519fa6f0e9e182c416308bc07c3baddbc6bf0533" + }, + { + "id": "try-a", + "campaign": "x-film-same-scores", + "claims": [ + "CC-001", + "CC-004" + ], + "confidence": { + "CC-001": "supported_within_scope", + "CC-004": "supported_within_scope" + }, + "scope": { + "CC-001": "The kernel module at the bound commit. Reproduced from a clean clone (venv, pip install, execute, assert) most recently on 2026-08-23, and re-reproduced by CI on every push and weekly. No claim about any deployed system.\n", + "CC-004": "The kernel identification module at the bound commit, re-verified by CI on every push and weekly in the witness stage of scripts/reproduce_cc001.py. No claim about any deployed system.\n" + }, + "evidence_urls": [ + "https://github.com/Cubits11/cc-framework/commit/167aa1ee514ee3d797ed9be2c60120d0c1db71f6" + ], + "sources": [ + { + "path": "claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", + "binding": "committed" + }, + { + "path": "distribution/experiments.yaml", + "sha256": "691aa3e679944eeded0464bbf41a2995b5491119ffcb14a3ec17ef2280acef70", + "commit": "e12ed3b133156798d73c7ed35c3133a689b20c73", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/e12ed3b133156798d73c7ed35c3133a689b20c73/distribution/experiments.yaml", + "binding": "committed" + }, + { + "path": "distribution/launch-units.yaml", + "sha256": "decdfebbca6d0cebf285193e7da48a65a681235376c962eb8f0e7de504f14749", + "commit": "30cf825968ed33488e5121c21d4f2209374968fd", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/30cf825968ed33488e5121c21d4f2209374968fd/distribution/launch-units.yaml", + "binding": "committed" + }, + { + "path": "campaigns.yaml", + "sha256": "6868db19a1184c391fec3b05215afac5e46d3b18808285347f77d8c6f1f87928", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/campaigns.yaml", + "binding": "committed" + }, + { + "path": "films/same-scores-different-worlds/manifest.yaml", + "sha256": "a289640c746bd81fa535af6dcea678ef26cdca3162fd9564b257724f2f7ede9c", + "commit": "2949acd763693c5446ed3a54a6d1644f2af4be51", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/2949acd763693c5446ed3a54a6d1644f2af4be51/films/same-scores-different-worlds/manifest.yaml", + "binding": "committed" + } + ], + "posts": [ + "Do two published miss rates determine how often both guards miss the same item?", + "PROVED \u2014 Fr\u00e9chet\u2013Hoeffding bounds; both endpoint worlds are re-verified from a clean clone of the bound kernel in CI (CC-001, CC-004).", + "Nothing here measures a real guardrail; the population is a construction and no point inside the interval is selected. Independence is what one assumption chooses, not what the scores imply.", + "Reproduce: python3 scripts/try_same_scores.py", + "https://cubits11.github.io/try/?utm_source=x&utm_medium=social&utm_campaign=e3-films&utm_content=same-scores#try-a", + "https://github.com/Cubits11/cc-framework/commit/167aa1ee514ee3d797ed9be2c60120d0c1db71f6" + ], + "attribution_url": "https://cubits11.github.io/try/?utm_source=x&utm_medium=social&utm_campaign=e3-films&utm_content=same-scores#try-a", + "audience_hypothesis": "AI evaluation researchers and guardrail engineers on X", + "why": "run TRY-A and file the result, match or mismatch", + "success_signal": "one reproduction issue filed by someone who is not the author", + "post_type": "thread", + "topic": "same-scores", + "hook": "question", + "visual": "none", + "cta": "reproduce", + "thread_structure": "question-status-limits-command-evidence", + "visual_candidate": { + "path": "films/same-scores-different-worlds/renders/same-scores-different-worlds__master.poster.png", + "master": "films/same-scores-different-worlds/renders/same-scores-different-worlds__master.mp4", + "state": "held_existing_cold_test_gate" + }, + "scheduled_at": null, + "state": "held_owner_review", + "anti_hype": [ + "no vendor ranking", + "no population or causal inference", + "no synthetic engagement", + "retain non-claims", + "no automated replies" + ], + "rank_basis": "existing reproducible route; source order; no audience-performance evidence", + "revision": "830eb0591c46f6494ef20746c04a899831d4b667419ea24e0ce2395132eb2ee3" + }, + { + "id": "try-b", + "campaign": "x-film-thirteen-worlds", + "claims": [ + "MC-002", + "MC-003" + ], + "confidence": { + "MC-002": "supported_within_scope", + "MC-003": "supported_within_scope" + }, + "scope": { + "MC-002": "Exactly the released file at the bound commit: 170 of the evaluation's 990 non-adversarial prompts (the study's ~4,165 adversarial prompts have no per-item release), selected by the authors under an unstated rule. Counting arithmetic only, at the vendors' released binary verdicts; no population estimate, no uncertainty interval, and no claim about any system's true rates is offered or licensed. Computed by scripts/reanalyze_bells_subset.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified file.\n", + "MC-003": "The mathematics is classical \u2014 Bonferroni for the lower endpoint, monotonicity for the upper, Fr\u00e9chet for sharpness \u2014 and nothing here claims to have derived it. What is contributed is the application: pricing, in probability units, what marginal-only guardrail reporting leaves undetermined, and identifying leave-one-out unions as a compact privacy-preserving disclosure of exclusive full-stack coverage. They do not identify pairwise or higher-order overlap, Shapley values, or causal structure. Every statement is conditional on block-on-any composition, per-item decisions that are functions of the item alone, a fixed population, full exposure, and a fixed operating point; it says nothing about sequential routes where downstream measurement is censored by upstream blocks, agentic systems where an intervention changes the trajectory, or any vendor's product. The empirical half is one stratum of one author-selected subset: 82 harmful prompts, with 50 benign and 38 borderline held as separate strata and never folded in. It is a BELLS-specific reproduction, not an assertion that BELLS is the census's only per-item outcome release. scripts/reanalyze_bells_subset.py re-computes the BELLS counts from the hash-verified upstream file in CI; scripts/identification.py checks the stated identification transformations against that registered envelope.\n" + }, + "evidence_urls": [ + "https://github.com/CentreSecuriteIA/bells_leaderboard/blob/507566c5a4606c8e3dec0bd59a5c5fde62594951/data/non_adversarial_prompts.csv", + "https://github.com/Cubits11/cubits11.github.io/blob/main/scripts/identification.py" + ], + "sources": [ + { + "path": "claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", + "binding": "committed" + }, + { + "path": "distribution/experiments.yaml", + "sha256": "691aa3e679944eeded0464bbf41a2995b5491119ffcb14a3ec17ef2280acef70", + "commit": "e12ed3b133156798d73c7ed35c3133a689b20c73", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/e12ed3b133156798d73c7ed35c3133a689b20c73/distribution/experiments.yaml", + "binding": "committed" + }, + { + "path": "distribution/launch-units.yaml", + "sha256": "decdfebbca6d0cebf285193e7da48a65a681235376c962eb8f0e7de504f14749", + "commit": "30cf825968ed33488e5121c21d4f2209374968fd", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/30cf825968ed33488e5121c21d4f2209374968fd/distribution/launch-units.yaml", + "binding": "committed" + }, + { + "path": "campaigns.yaml", + "sha256": "6868db19a1184c391fec3b05215afac5e46d3b18808285347f77d8c6f1f87928", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/campaigns.yaml", + "binding": "committed" + }, + { + "path": "scripts/identification.py", + "sha256": "4baba5ef00e5d8db19910abf1560b15cb7c68552690fb5c68c2dadabf8d563d8", + "commit": "b9c4d1ab89c8508e0f25c6fcd9a7a83db08cb771", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/b9c4d1ab89c8508e0f25c6fcd9a7a83db08cb771/scripts/identification.py", + "binding": "committed" + }, + { + "path": "films/thirteen-worlds/manifest.yaml", + "sha256": "06b98c7ba3f8fb2e99648baead4fbbae555a91e709e925c3ddbfa2bfe13c8cf6", + "commit": "2949acd763693c5446ed3a54a6d1644f2af4be51", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/2949acd763693c5446ed3a54a6d1644f2af4be51/films/thirteen-worlds/manifest.yaml", + "binding": "committed" + } + ], + "posts": [ + "From a public per-item release, how many of 82 harmful prompts did all five supervisors miss \u2014 and what did the five published miss rates alone permit?", + "OBSERVED counts on one released file (hash-verified, pinned commit); PROVED identified set from the marginals. Counting arithmetic, not a population estimate.", + "One author-selected 82-prompt stratum at unstated default configurations: not a vendor ranking, not a population estimate, not an observed deployed stack, and nothing about the adversarial prompts that have no per-item release.", + "Reproduce: python3 scripts/reanalyze_bells_subset.py", + "https://cubits11.github.io/try/?utm_source=x&utm_medium=social&utm_campaign=e3-films&utm_content=thirteen-worlds", + "https://github.com/CentreSecuriteIA/bells_leaderboard/blob/507566c5a4606c8e3dec0bd59a5c5fde62594951/data/non_adversarial_prompts.csv", + "https://github.com/Cubits11/cubits11.github.io/blob/main/scripts/identification.py" + ], + "attribution_url": "https://cubits11.github.io/try/?utm_source=x&utm_medium=social&utm_campaign=e3-films&utm_content=thirteen-worlds", + "audience_hypothesis": "benchmark authors and people who have read the BELLS paper", + "why": "run TRY-B against the pinned released file", + "success_signal": "one TRY-B reproduction filed, or a BELLS author confirms or corrects the row", + "post_type": "thread", + "topic": "thirteen-worlds", + "hook": "question", + "visual": "none", + "cta": "reproduce", + "thread_structure": "question-status-limits-command-evidence", + "visual_candidate": { + "path": "films/thirteen-worlds/renders/thirteen-worlds__master.poster.png", + "master": "films/thirteen-worlds/renders/thirteen-worlds__master.mp4", + "state": "held_existing_cold_test_gate" + }, + "scheduled_at": null, + "state": "held_owner_review", + "anti_hype": [ + "no vendor ranking", + "no population or causal inference", + "no synthetic engagement", + "retain non-claims", + "no automated replies" + ], + "rank_basis": "existing reproducible route; source order; no audience-performance evidence", + "revision": "a86242b325da2710f3025fa76f7c4231f25a6825d91ffb1d5129b2c4a441a34e" + }, + { + "id": "try-c", + "campaign": "x-film-missing-column-census", + "claims": [ + "MC-001" + ], + "confidence": { + "MC-001": "supported_within_scope" + }, + "scope": { + "MC-001": "A census claim about public reporting, bounded by the documented search protocol documented in census.yaml. Its literal inclusion wording is locked in repository history before row classification; that is a reproducibility lock, not an independent preregistration. It covers the artifacts that search found, not everything in existence; it says nothing about how well any evaluated system or stack performs, and nothing about what the unmeasured joint statistics would show. N, M, and K are recomputed mechanically from census.yaml by scripts/verify_census.py; this envelope's expected block is cross-checked against that computation in CI. Since census schema v3 the joint-evidence mode counts in this note are derived the same way, from a declared joint_scope_additional field, rather than counted by hand here; scripts/verify_facts.py binds every current census numeral on every public page, and on the deployed page after release, to the quantity it asserts. M is a ladder, not a comparability verdict: 14 document shared items and a common event definition, 12 have no stated threshold mismatch, and 0 document matched operating thresholds together with full exposure. Nothing here asserts that the 14 are interchangeable at a common operating point; the opposite is the finding. The 5 is an inclusive discovery count of noninterchangeable artifacts, not an all-miss rate or a deployment conclusion. Four artifacts print at least one composition result and two release aligned per-item outcomes from which joint statistics are directly computable; one artifact does both, so those descriptions intentionally overlap. The frozen phrase \"separately attributable\" did not specify whether a product or model name must be printed: the primary count treats systems a source distinguishes and reports consistently as separately attributable even when anonymized; the declared named-products-only sensitivity excludes Unit42 and mechanically yields N/M/K = 19/13/5. A drafted further clause \u2014 that no printed joint statistic among the then-four covers a commercial guardrail API \u2014 was withdrawn before publication when same-day adversarial review produced a live counterexample under one defensible reading (WAInjectBench's printed ensemble includes GPT-4o prompted as a detector); this claim's forbidden rescues bar narrowing \"commercial\" after the fact, so the clause was dropped rather than reinterpreted, and the event is recorded in the census revision history. On 2026-08-30, a post-release source audit found the distinct Multimodal Safeguard Bench repository, whose currently public Git history carries a pre-cutoff commit timestamp, and which the documented search had missed. That met this claim's prior REJECT falsifier: the 19/13/4 envelope is rejected and retained in census revision history. This 20/14/5 envelope is its corrected, superseding proposition, not a reinterpretation of the former count.\n" + }, + "evidence_urls": [ + "https://github.com/Cubits11/cubits11.github.io/blob/main/census.yaml" + ], + "sources": [ + { + "path": "claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", + "binding": "committed" + }, + { + "path": "distribution/experiments.yaml", + "sha256": "691aa3e679944eeded0464bbf41a2995b5491119ffcb14a3ec17ef2280acef70", + "commit": "e12ed3b133156798d73c7ed35c3133a689b20c73", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/e12ed3b133156798d73c7ed35c3133a689b20c73/distribution/experiments.yaml", + "binding": "committed" + }, + { + "path": "distribution/launch-units.yaml", + "sha256": "decdfebbca6d0cebf285193e7da48a65a681235376c962eb8f0e7de504f14749", + "commit": "30cf825968ed33488e5121c21d4f2209374968fd", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/30cf825968ed33488e5121c21d4f2209374968fd/distribution/launch-units.yaml", + "binding": "committed" + }, + { + "path": "campaigns.yaml", + "sha256": "6868db19a1184c391fec3b05215afac5e46d3b18808285347f77d8c6f1f87928", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/campaigns.yaml", + "binding": "committed" + }, + { + "path": "census.yaml", + "sha256": "68b14081b85e18ca1a7b5bce33bbf38fe3b8dafe876e7329057241404037b425", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/census.yaml", + "binding": "committed" + }, + { + "path": "films/the-missing-column-census/manifest.yaml", + "sha256": "5abaf8574cb481bf2f29701853db1242d300b955ca77410e0f6e3e17f22f2760", + "commit": "2949acd763693c5446ed3a54a6d1644f2af4be51", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/2949acd763693c5446ed3a54a6d1644f2af4be51/films/the-missing-column-census/manifest.yaml", + "binding": "committed" + } + ], + "posts": [ + "Does an evaluation you know preserve what its compared systems miss together, or only what each misses alone?", + "PROCEDURE \u2014 the same frozen criteria (v1) and predicates the census applies; your answers are OBSERVED by you, the classification is DERIVED from them, and a candidate row is a proposal until reviewed.", + "A classification is a reporting fact about the artifact, never a performance finding about any system; a single audit by one person is a first-pass reading, exactly as the census's own rows are.", + "Reproduce: python3 scripts/try_audit.py", + "https://cubits11.github.io/try/?utm_source=x&utm_medium=reply&utm_campaign=e3-films&utm_content=missing-column-census#counterexample", + "https://github.com/Cubits11/cubits11.github.io/blob/main/census.yaml" + ], + "attribution_url": "https://cubits11.github.io/try/?utm_source=x&utm_medium=reply&utm_campaign=e3-films&utm_content=missing-column-census#counterexample", + "audience_hypothesis": "authors of the twenty examined evaluations", + "why": "correct a row, or confirm it, or run TRY-C on another evaluation", + "success_signal": "one row correction, one author confirmation, or one TRY-C candidate row filed", + "post_type": "thread", + "topic": "audit", + "hook": "question", + "visual": "none", + "cta": "reproduce", + "thread_structure": "question-status-limits-command-evidence", + "visual_candidate": { + "path": "films/the-missing-column-census/renders/the-missing-column-census__master.poster.png", + "master": "films/the-missing-column-census/renders/the-missing-column-census__master.mp4", + "state": "held_existing_cold_test_gate" + }, + "scheduled_at": null, + "state": "held_owner_review", + "anti_hype": [ + "no vendor ranking", + "no population or causal inference", + "no synthetic engagement", + "retain non-claims", + "no automated replies" + ], + "rank_basis": "existing reproducible route; source order; no audience-performance evidence", + "revision": "524ebd0e48c04aea86154bd0ad11a5aa8a1ad4a6649b371ac3b967b89ac943ac" + }, + { + "id": "show-overlap", + "campaign": "x-show-overlap", + "claims": [ + "CC-001", + "CC-004" + ], + "posts": [ + "Two checks each miss 10 of 100 faulty cases.\n\nDifferent misses: together they miss 0.\nSame misses: together they miss 10.\n\nSame separate scores. Different protection.\n\nThis can happen in a tiny workflow or a large system.\n\nConstructed example. Show the overlap.", + "This example assumes both checks see the same cases and either can catch a fault. If one check changes the next check\u2019s workload, measure that workflow too.\n\nThe example and current research corrections: https://cubits11.github.io/overlap/" + ], + "destination_url": "https://cubits11.github.io/overlap/", + "source_paths": [ + "claims.yaml", + "campaigns.yaml", + "distribution/educational-posts.json", + "distribution/NEXT-POST.md", + "corrections/records/2026-09-10.json", + "corrections/index.html", + "overlap/index.html", + "overlap/example.js" + ], + "post_type": "thread", + "topic": "same-scores-general-process", + "hook": "constructed-example", + "visual": "interactive-destination", + "cta": "inspect-overlap", + "thread_structure": "constructed-example-scope-evidence", + "anti_hype": [ + "constructed only", + "either-check-catches rule on a shared pool", + "no E6 range or E7B confirmation", + "no general system reliability claim" + ], + "sources": [ + { + "path": "claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", + "binding": "committed" + }, + { + "path": "campaigns.yaml", + "sha256": "6868db19a1184c391fec3b05215afac5e46d3b18808285347f77d8c6f1f87928", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/campaigns.yaml", + "binding": "committed" + }, + { + "path": "distribution/educational-posts.json", + "sha256": "422106687c5dce3de5fd2b6a6408afa6469c790f9ba8f456685fe229e9e2e668", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/distribution/educational-posts.json", + "binding": "committed" + }, + { + "path": "distribution/NEXT-POST.md", + "sha256": "ac55785a806108a75a76b1f4bcb9d419417d212ed71ce578905b6af2d1e086c8", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/distribution/NEXT-POST.md", + "binding": "committed" + }, + { + "path": "corrections/records/2026-09-10.json", + "sha256": "a2cca203fd92ffe17ec09d841e90154cfd9cf966f4aeb1b1a62ad0c223cb2cc0", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/corrections/records/2026-09-10.json", + "binding": "committed" + }, + { + "path": "corrections/index.html", + "sha256": "921d8ed980e56f96d9c352cf8a06efb775b5f1720fdff2b0297c3e58567f6335", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/corrections/index.html", + "binding": "committed" + }, + { + "path": "overlap/index.html", + "sha256": "005986c3395f1da76d1bab625286ecd9d505556659e16d51a41f16107df10fa2", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/overlap/index.html", + "binding": "committed" + }, + { + "path": "overlap/example.js", + "sha256": "9e7a1aaefbf9286211267b764c716ea60ad2946e5222164182c5f5e88a496bd5", + "commit": "48e98e44be87497bcd0c4c650e8123efde98c625", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/48e98e44be87497bcd0c4c650e8123efde98c625/overlap/example.js", + "binding": "committed" + } + ], + "confidence": { + "CC-001": "supported_within_scope", + "CC-004": "supported_within_scope" + }, + "scope": { + "CC-001": "The kernel module at the bound commit. Reproduced from a clean clone (venv, pip install, execute, assert) most recently on 2026-08-23, and re-reproduced by CI on every push and weekly. No claim about any deployed system.\n", + "CC-004": "The kernel identification module at the bound commit, re-verified by CI on every push and weekly in the witness stage of scripts/reproduce_cc001.py. No claim about any deployed system.\n" + }, + "evidence_urls": [ + "https://github.com/Cubits11/cc-framework/commit/167aa1ee514ee3d797ed9be2c60120d0c1db71f6", + "https://github.com/Cubits11/cc-framework/commit/167aa1ee514ee3d797ed9be2c60120d0c1db71f6" + ], + "audience_hypothesis": "People composing checks, reviews, tests, and safeguards into a process", + "why": "Understand the constructed overlap example and read current correction status", + "success_signal": "An external reader explains what a second check adds or uses a paired outcome table", + "state": "held_owner_review", + "scheduled_at": null, + "attribution_url": "https://cubits11.github.io/overlap/", + "rank_basis": "constructed example; no disputed empirical headline", + "revision": "8f77b7ec640f0de7e3392e082c620fbf84071ff289b0ad73a0dd55e543468dc7" } ] diff --git a/distribution/traction/drafts.json b/distribution/traction/drafts.json index 01639c2..6114a1b 100644 --- a/distribution/traction/drafts.json +++ b/distribution/traction/drafts.json @@ -20,9 +20,9 @@ "sources": [ { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, { @@ -87,7 +87,7 @@ "no automated replies" ], "rank_basis": "existing reproducible route; source order; no audience-performance evidence", - "revision": "119161f5c94be04883eacf74e0649956f4fe26d5dde785f05859adf2564dd875" + "revision": "830eb0591c46f6494ef20746c04a899831d4b667419ea24e0ce2395132eb2ee3" }, { "id": "try-b", @@ -111,9 +111,9 @@ "sources": [ { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, { @@ -186,7 +186,7 @@ "no automated replies" ], "rank_basis": "existing reproducible route; source order; no audience-performance evidence", - "revision": "068db07442ba1185d3ba315eaf47dff9ee5d3f3282249d9191908b00c917a680" + "revision": "a86242b325da2710f3025fa76f7c4231f25a6825d91ffb1d5129b2c4a441a34e" }, { "id": "try-c", @@ -206,9 +206,9 @@ "sources": [ { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, { @@ -280,7 +280,7 @@ "no automated replies" ], "rank_basis": "existing reproducible route; source order; no audience-performance evidence", - "revision": "baa408ffaa6666e6bbf13fad53a74ca370c7376f0256ea70e2f4c74010a21109" + "revision": "524ebd0e48c04aea86154bd0ad11a5aa8a1ad4a6649b371ac3b967b89ac943ac" }, { "id": "show-overlap", @@ -319,9 +319,9 @@ "sources": [ { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, { @@ -393,6 +393,6 @@ "scheduled_at": null, "attribution_url": "https://cubits11.github.io/overlap/", "rank_basis": "constructed example; no disputed empirical headline", - "revision": "6efe64e118fffd5e8bd7bf0d519fa6f0e9e182c416308bc07c3baddbc6bf0533" + "revision": "8f77b7ec640f0de7e3392e082c620fbf84071ff289b0ad73a0dd55e543468dc7" } ] diff --git a/distribution/traction/events.json b/distribution/traction/events.json index e702bee..3d28455 100644 --- a/distribution/traction/events.json +++ b/distribution/traction/events.json @@ -17,13 +17,13 @@ "novelty": "snapshot; not asserted newly published" }, { - "id": "45146912e9927e9d", + "id": "cf793f4a094ccd4d", "kind": "claims", "source": { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, "draft_ids": [ diff --git a/distribution/traction/manifest.json b/distribution/traction/manifest.json index a615108..c49119e 100644 --- a/distribution/traction/manifest.json +++ b/distribution/traction/manifest.json @@ -2,9 +2,9 @@ "version": 1, "generator": "scripts/distribute.py", "outputs": { - "events.json": "ad83f155add4fa6de4b575a74654515f0cee5f7ffbfa446a2ecb01139a88d39e", - "drafts.json": "e25df46ee528c0b668a784403b34cdd1acb036e66ef8c7a448248775f66c3d5a", - "queue.json": "83812d9922dc9b855e72f501b61a12d25d1361c4b787c93e74e6a812837be1c5", + "events.json": "556253a0e778184815649aa48b4efe75fa813c60bf3b754829ccfd31f2512958", + "drafts.json": "51b7a09e369f6c79b2fb570a5f3a024f2405a6ecf6425f8b0dbb0baf0e6aabdf", + "queue.json": "969b7d262c8046c9eada69aa675659c44b4a06db3ef8ad0cc7b728f12264b54f", "dashboard.json": "988051f7c9d5b82684071ab29fdb52a891ce0f93470405f402cc6beff62d04bb", "experiments.json": "604a2a40db52756f7cb4fd97e0f3f15eb7639b0c1a28e5bb95cbef738e4d139e", "dashboard.html": "bdd57844ab67c408544f4c13cb2e0d02942d7362d474edd529c14ba6a9a20191" diff --git a/distribution/traction/queue.json b/distribution/traction/queue.json index fb7a021..8a87311 100644 --- a/distribution/traction/queue.json +++ b/distribution/traction/queue.json @@ -29,9 +29,9 @@ "sources": [ { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, { @@ -96,7 +96,7 @@ "no automated replies" ], "rank_basis": "existing reproducible route; source order; no audience-performance evidence", - "revision": "119161f5c94be04883eacf74e0649956f4fe26d5dde785f05859adf2564dd875", + "revision": "830eb0591c46f6494ef20746c04a899831d4b667419ea24e0ce2395132eb2ee3", "approval": null }, { @@ -121,9 +121,9 @@ "sources": [ { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, { @@ -196,7 +196,7 @@ "no automated replies" ], "rank_basis": "existing reproducible route; source order; no audience-performance evidence", - "revision": "068db07442ba1185d3ba315eaf47dff9ee5d3f3282249d9191908b00c917a680", + "revision": "a86242b325da2710f3025fa76f7c4231f25a6825d91ffb1d5129b2c4a441a34e", "approval": null }, { @@ -217,9 +217,9 @@ "sources": [ { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, { @@ -291,7 +291,7 @@ "no automated replies" ], "rank_basis": "existing reproducible route; source order; no audience-performance evidence", - "revision": "baa408ffaa6666e6bbf13fad53a74ca370c7376f0256ea70e2f4c74010a21109", + "revision": "524ebd0e48c04aea86154bd0ad11a5aa8a1ad4a6649b371ac3b967b89ac943ac", "approval": null }, { @@ -331,9 +331,9 @@ "sources": [ { "path": "claims.yaml", - "sha256": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", - "commit": "8fc919c94548f48a5a4026a058e267377193c0a0", - "url": "https://github.com/Cubits11/cubits11.github.io/blob/8fc919c94548f48a5a4026a058e267377193c0a0/claims.yaml", + "sha256": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", + "commit": "052aa29ec369cd4b302d3d3aea7ca1254df2204f", + "url": "https://github.com/Cubits11/cubits11.github.io/blob/052aa29ec369cd4b302d3d3aea7ca1254df2204f/claims.yaml", "binding": "committed" }, { @@ -405,7 +405,7 @@ "scheduled_at": null, "attribution_url": "https://cubits11.github.io/overlap/", "rank_basis": "constructed example; no disputed empirical headline", - "revision": "6efe64e118fffd5e8bd7bf0d519fa6f0e9e182c416308bc07c3baddbc6bf0533", + "revision": "8f77b7ec640f0de7e3392e082c620fbf84071ff289b0ad73a0dd55e543468dc7", "approval": null } ] diff --git a/docs/RESEARCH_INDEX.md b/docs/RESEARCH_INDEX.md index 826ef0d..141dba6 100644 --- a/docs/RESEARCH_INDEX.md +++ b/docs/RESEARCH_INDEX.md @@ -2,7 +2,7 @@ GENERATED by scripts/generate_research_index.py from claims.yaml, distribution/outcomes.yaml, distribution/experiments.yaml and the film manifests. Do not edit by hand. -Registry v0.4 · last owner review 2026-10-24 · one question: *Do published per-guard scores determine what the stack misses together?* +Registry v0.4 · last owner review 2026-11-02 · one question: *Do published per-guard scores determine what the stack misses together?* ## Independent outcomes (someone who is not the author did work) diff --git a/docs/graph/repo-graph.json b/docs/graph/repo-graph.json index f128252..5540843 100644 --- a/docs/graph/repo-graph.json +++ b/docs/graph/repo-graph.json @@ -5410,24 +5410,17 @@ }, "findings": [ { - "claim": "MC-002", - "days_left": 3, + "claim": "REL-001", + "days_left": 14, "detector": "D3", - "review_due": "2026-10-27", - "severity": "REVIEW" - }, - { - "claim": "MC-003", - "days_left": 5, - "detector": "D3", - "review_due": "2026-10-29", + "review_due": "2026-11-16", "severity": "REVIEW" }, { - "claim": "MC-004", - "days_left": 6, + "claim": "MC-001", + "days_left": 11, "detector": "D3", - "review_due": "2026-10-30", + "review_due": "2026-11-13", "severity": "REVIEW" }, { @@ -5462,239 +5455,365 @@ "sha256": "22e956020cdfdd43c328f02441f003c933ee4e7083d02d9d24e414e4b42ae2e6" }, { - "ahead": 5, + "ahead": 1, + "behind": 84, + "branch": "backup/2026-10-17/agent-state", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 1, + "behind": 84, + "branch": "backup/2026-10-17/cadence-publication", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 1, + "behind": 84, + "branch": "backup/2026-10-17/discovery-only", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 1, + "behind": 84, + "branch": "backup/2026-10-17/film-capture", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 7, + "behind": 84, + "branch": "backup/2026-10-17/main-checkout", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 3, + "behind": 63, + "branch": "backup/2026-10-17/site-repair-candidate", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 1, + "behind": 65, + "branch": "census/candidates-2026-10", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 4, + "behind": 84, + "branch": "claude/claim-succession-minimal", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 2, + "behind": 8, + "branch": "claude/design-system-2026-10-24", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 1, + "behind": 84, + "branch": "claude/k1-hold", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 1, "behind": 0, - "branch": "claude/gv-001-hosted-re-review", + "branch": "claude/mc-002-004-review-2026-11-02", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 4, + "behind": 84, + "branch": "claude/phone-width-a11y", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 4, + "behind": 7, + "branch": "codex/e2-cost", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 2, + "behind": 7, + "branch": "codex/e2-e2", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 3, + "behind": 7, + "branch": "codex/e2-e3", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 58, + "behind": 7, + "branch": "codex/e2-e4", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 4, + "behind": 7, + "branch": "codex/e2-e5", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 1, + "behind": 8, + "branch": "publication/title-social-square", + "detector": "D4", + "severity": "TOPOLOGY" + }, + { + "ahead": 1, + "behind": 65, "branch": "origin/census/candidates-2026-10", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 4, - "behind": 77, + "behind": 84, "branch": "origin/claude/claim-succession-minimal", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 13, - "behind": 77, + "behind": 84, "branch": "origin/claude/e9-result", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 2, - "behind": 77, + "behind": 84, "branch": "origin/claude/eloquent-shannon-dlvcz3", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 3, - "behind": 77, + "behind": 84, "branch": "origin/claude/falsifier-assessment-kernel", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 3, - "behind": 77, + "behind": 84, "branch": "origin/claude/glassroot-handoff", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 2, - "behind": 77, + "behind": 84, "branch": "origin/claude/homepage-epistemic-motion", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 77, + "behind": 84, "branch": "origin/claude/k1-hold", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 3, - "behind": 77, + "behind": 84, "branch": "origin/claude/lucid-franklin-jloow0", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 4, - "behind": 77, + "behind": 84, "branch": "origin/claude/phone-width-a11y", "detector": "D4", "severity": "TOPOLOGY" }, + { + "ahead": 3, + "behind": 0, + "branch": "origin/claude/publication-gate-title-check", + "detector": "D4", + "severity": "TOPOLOGY" + }, { "ahead": 4, - "behind": 77, + "behind": 84, "branch": "origin/claude/temple-researcher-outreach-on2v73", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 77, + "behind": 84, "branch": "origin/claude/verify-claims-refusal-unknown", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 2, - "behind": 77, + "behind": 84, "branch": "origin/claude/x-gate", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 2, - "behind": 0, + "behind": 7, "branch": "origin/codex/e2-e2", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 3, - "behind": 0, + "behind": 7, "branch": "origin/codex/e2-e3", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 0, + "behind": 7, "branch": "origin/codex/e2-e4", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 4, - "behind": 0, + "behind": 7, "branch": "origin/codex/e2-e5", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 114, + "behind": 121, "branch": "origin/icloud-snapshot/2026-09-19", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 77, + "behind": 84, "branch": "origin/icloud-snapshot/2026-09-19-site", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 1, + "behind": 8, "branch": "origin/publication/title-social-square", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 83, + "behind": 90, "branch": "origin/rescue/commit-01fea45b4e98", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 273, + "behind": 280, "branch": "origin/rescue/commit-119f257b763c", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 92, + "behind": 99, "branch": "origin/rescue/commit-2139046f7192", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 166, + "behind": 173, "branch": "origin/rescue/commit-22fa9f9a8456", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 3, - "behind": 158, + "behind": 165, "branch": "origin/rescue/commit-4a362b5d714f", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 178, + "behind": 185, "branch": "origin/rescue/commit-5e70a9df4626", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 303, + "behind": 310, "branch": "origin/rescue/commit-80250a3e7fc4", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 273, + "behind": 280, "branch": "origin/rescue/commit-93200984af16", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 170, + "behind": 177, "branch": "origin/rescue/commit-bd0c87d0f1ea", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 92, + "behind": 99, "branch": "origin/rescue/commit-c74849b7a928", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 4, - "behind": 158, + "behind": 165, "branch": "origin/rescue/commit-cec26e585b5a", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 165, + "behind": 172, "branch": "origin/rescue/commit-ea88d6552c2f", "detector": "D4", "severity": "TOPOLOGY" }, { "ahead": 1, - "behind": 234, + "behind": 241, "branch": "origin/rescue/commit-f8e555738fc2", "detector": "D4", "severity": "TOPOLOGY" @@ -5719,8 +5838,8 @@ } ], "generated_from": { - "branch": "claude/gv-001-hosted-re-review", - "head": "7a1061e" + "branch": "claude/mc-002-004-review-2026-11-02", + "head": "052aa29" }, "nodes": [ { @@ -6346,7 +6465,7 @@ }, { "consequence": "REJECT", - "days_left": 59, + "days_left": 50, "id": "CC-001", "last_reviewed": "2026-08-24", "review_due": "2026-12-22", @@ -6355,7 +6474,7 @@ }, { "consequence": "REJECT", - "days_left": 59, + "days_left": 50, "id": "CC-002", "last_reviewed": "2026-08-24", "review_due": "2026-12-22", @@ -6364,7 +6483,7 @@ }, { "consequence": "REJECT", - "days_left": 59, + "days_left": 50, "id": "CC-003", "last_reviewed": "2026-08-24", "review_due": "2026-12-22", @@ -6373,7 +6492,7 @@ }, { "consequence": "REJECT", - "days_left": 59, + "days_left": 50, "id": "CC-004", "last_reviewed": "2026-08-24", "review_due": "2026-12-22", @@ -6382,7 +6501,7 @@ }, { "consequence": "NARROW", - "days_left": 59, + "days_left": 50, "id": "CC-005", "last_reviewed": "2026-08-24", "review_due": "2026-12-22", @@ -6391,7 +6510,7 @@ }, { "consequence": "REJECT", - "days_left": 73, + "days_left": 64, "id": "CC-006", "last_reviewed": "2026-09-07", "review_due": "2027-01-05", @@ -6400,7 +6519,7 @@ }, { "consequence": "REJECT", - "days_left": 23, + "days_left": 14, "id": "REL-001", "last_reviewed": "2026-10-17", "review_due": "2026-11-16", @@ -6409,7 +6528,7 @@ }, { "consequence": "REJECT", - "days_left": 20, + "days_left": 11, "id": "MC-001", "last_reviewed": "2026-09-14", "review_due": "2026-11-13", @@ -6418,34 +6537,34 @@ }, { "consequence": "REJECT", - "days_left": 3, + "days_left": 60, "id": "MC-002", - "last_reviewed": "2026-08-28", - "review_due": "2026-10-27", + "last_reviewed": "2026-11-02", + "review_due": "2027-01-01", "support_url": "https://github.com/CentreSecuriteIA/bells_leaderboard/blob/507566c5a4606c8e3dec0bd59a5c5fde62594951/data/non_adversarial_prompts.csv", "type": "claim" }, { "consequence": "REJECT", - "days_left": 5, + "days_left": 60, "id": "MC-003", - "last_reviewed": "2026-08-30", - "review_due": "2026-10-29", + "last_reviewed": "2026-11-02", + "review_due": "2027-01-01", "support_url": "https://github.com/Cubits11/cubits11.github.io/blob/main/scripts/identification.py", "type": "claim" }, { "consequence": "REJECT", - "days_left": 6, + "days_left": 60, "id": "MC-004", - "last_reviewed": "2026-08-31", - "review_due": "2026-10-30", + "last_reviewed": "2026-11-02", + "review_due": "2027-01-01", "support_url": "https://github.com/PatrickKollman/Multimodal-Safeguard-Bench/tree/fb6f32e6b50b6faad833b815ebcd80afd2068bff/results/full_run", "type": "claim" }, { "consequence": "NARROW", - "days_left": 33, + "days_left": 24, "id": "AF-001", "last_reviewed": "2026-08-28", "review_due": "2026-11-26", @@ -6454,7 +6573,7 @@ }, { "consequence": "NARROW", - "days_left": 59, + "days_left": 50, "id": "GA-001", "last_reviewed": "2026-08-24", "review_due": "2026-12-22", @@ -6463,7 +6582,7 @@ }, { "consequence": "NARROW", - "days_left": 120, + "days_left": 111, "id": "GV-001", "last_reviewed": "2026-10-24", "review_due": "2027-02-21", @@ -6472,7 +6591,7 @@ }, { "consequence": "NARROW", - "days_left": 304, + "days_left": 295, "id": "GCE-001", "last_reviewed": "2026-08-24", "review_due": "2027-08-24", @@ -6481,7 +6600,7 @@ }, { "consequence": "NARROW", - "days_left": 66, + "days_left": 57, "id": "SITE-001", "last_reviewed": "2026-08-31", "review_due": "2026-12-29", @@ -6490,7 +6609,7 @@ }, { "consequence": "REJECT", - "days_left": 65, + "days_left": 56, "id": "SITE-002", "last_reviewed": "2026-08-30", "review_due": "2026-12-28", @@ -6499,7 +6618,7 @@ }, { "consequence": "REJECT", - "days_left": 74, + "days_left": 65, "id": "E3-001", "last_reviewed": "2026-09-08", "review_due": "2027-01-06", @@ -6508,7 +6627,7 @@ }, { "consequence": "REJECT", - "days_left": 74, + "days_left": 65, "id": "E3B-001", "last_reviewed": "2026-09-08", "review_due": "2027-01-06", @@ -6517,7 +6636,7 @@ }, { "consequence": "REJECT", - "days_left": 76, + "days_left": 67, "id": "E6-001", "last_reviewed": "2026-09-10", "review_due": "2027-01-08", @@ -6526,7 +6645,7 @@ }, { "consequence": "REJECT", - "days_left": 76, + "days_left": 67, "id": "E7B-001", "last_reviewed": "2026-09-10", "review_due": "2027-01-08", @@ -6535,7 +6654,7 @@ }, { "consequence": "REJECT", - "days_left": 113, + "days_left": 104, "id": "E9-001", "last_reviewed": "2026-10-17", "review_due": "2027-02-14", @@ -6641,19 +6760,262 @@ "type": "experiment" }, { - "ahead_of_origin_main": 5, + "ahead_of_origin_main": 0, + "behind_origin_main": 53, + "head": "2d462f9", + "id": "agent-state/capsule", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, + "behind_origin_main": 84, + "head": "4612b11", + "id": "backup/2026-10-17/agent-state", + "last_commit": "2026-10-17", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, + "behind_origin_main": 84, + "head": "dcc9426", + "id": "backup/2026-10-17/cadence-publication", + "last_commit": "2026-10-17", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, + "behind_origin_main": 84, + "head": "840f597", + "id": "backup/2026-10-17/discovery-only", + "last_commit": "2026-10-17", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, + "behind_origin_main": 84, + "head": "3dfd171", + "id": "backup/2026-10-17/film-capture", + "last_commit": "2026-10-17", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 7, + "behind_origin_main": 84, + "head": "62c0c0b", + "id": "backup/2026-10-17/main-checkout", + "last_commit": "2026-10-17", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 3, + "behind_origin_main": 63, + "head": "ea84f0c", + "id": "backup/2026-10-17/site-repair-candidate", + "last_commit": "2026-10-17", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, + "behind_origin_main": 65, + "head": "6a91961", + "id": "census/candidates-2026-10", + "last_commit": "2026-10-17", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, "behind_origin_main": 0, - "head": "7a1061e", - "id": "claude/gv-001-hosted-re-review", + "head": "9f617d1", + "id": "claude/census-transport-2026-11-02", "last_commit": "2026-10-24", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 4, + "behind_origin_main": 84, + "head": "d1a09c5", + "id": "claude/claim-succession-minimal", + "last_commit": "2026-09-25", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 2, + "behind_origin_main": 8, + "head": "4780e83", + "id": "claude/design-system-2026-10-24", + "last_commit": "2026-11-02", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 58, + "head": "a13b86d", + "id": "claude/film-receipt-facts-subset", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 42, + "head": "e62476e", + "id": "claude/friendly-babbage-5hro0x", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 30, + "head": "bc5bda8", + "id": "claude/k-infra-calendar-weeks", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, + "behind_origin_main": 84, + "head": "1d44a13", + "id": "claude/k1-hold", + "last_commit": "2026-09-22", "reachable_from_origin_main": false, "type": "branch" }, { "ahead_of_origin_main": 0, + "behind_origin_main": 35, + "head": "842e06d", + "id": "claude/keen-jemison-cfd59b", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, "behind_origin_main": 0, - "head": "e2cec7b", + "head": "052aa29", + "id": "claude/mc-002-004-review-2026-11-02", + "last_commit": "2026-11-02", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 35, + "head": "842e06d", + "id": "claude/modest-poincare-28a127", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 4, + "behind_origin_main": 84, + "head": "057c7b1", + "id": "claude/phone-width-a11y", + "last_commit": "2026-09-21", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 4, + "behind_origin_main": 7, + "head": "1fe36e6", + "id": "codex/e2-cost", + "last_commit": "2026-10-24", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 2, + "behind_origin_main": 7, + "head": "9259091", + "id": "codex/e2-e2", + "last_commit": "2026-10-24", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 3, + "behind_origin_main": 7, + "head": "90328c0", + "id": "codex/e2-e3", + "last_commit": "2026-10-24", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, + "behind_origin_main": 7, + "head": "a451d9b", + "id": "codex/e2-e4", + "last_commit": "2026-10-24", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 4, + "behind_origin_main": 7, + "head": "1fe36e6", + "id": "codex/e2-e5", + "last_commit": "2026-10-24", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 12, + "head": "9211af0", + "id": "e9-001/register", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 36, + "head": "ee9617c", + "id": "gv001/title-clarify", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 0, + "head": "9f617d1", "id": "main", + "last_commit": "2026-10-24", + "reachable_from_origin_main": true, + "type": "branch" + }, + { + "ahead_of_origin_main": 1, + "behind_origin_main": 8, + "head": "5dd43af", + "id": "publication/title-social-square", + "last_commit": "2026-10-17", + "reachable_from_origin_main": false, + "type": "branch" + }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 21, + "head": "8062f8c", + "id": "site/empirical-repair", "last_commit": "2026-10-17", "reachable_from_origin_main": true, "type": "branch" @@ -6661,15 +7023,15 @@ { "ahead_of_origin_main": 0, "behind_origin_main": 0, - "head": "e2cec7b", + "head": "9f617d1", "id": "origin", - "last_commit": "2026-10-17", + "last_commit": "2026-10-24", "reachable_from_origin_main": true, "type": "branch" }, { "ahead_of_origin_main": 1, - "behind_origin_main": 58, + "behind_origin_main": 65, "head": "6a91961", "id": "origin/census/candidates-2026-10", "last_commit": "2026-10-17", @@ -6678,7 +7040,7 @@ }, { "ahead_of_origin_main": 4, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "d1a09c5", "id": "origin/claude/claim-succession-minimal", "last_commit": "2026-09-25", @@ -6687,7 +7049,7 @@ }, { "ahead_of_origin_main": 13, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "b50a9c6", "id": "origin/claude/e9-result", "last_commit": "2026-09-25", @@ -6696,7 +7058,7 @@ }, { "ahead_of_origin_main": 2, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "56a0cfc", "id": "origin/claude/eloquent-shannon-dlvcz3", "last_commit": "2026-09-22", @@ -6705,7 +7067,7 @@ }, { "ahead_of_origin_main": 3, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "c1b71f9", "id": "origin/claude/falsifier-assessment-kernel", "last_commit": "2026-09-27", @@ -6714,16 +7076,25 @@ }, { "ahead_of_origin_main": 3, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "2e9eeab", "id": "origin/claude/glassroot-handoff", "last_commit": "2026-09-27", "reachable_from_origin_main": false, "type": "branch" }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 1, + "head": "8572855", + "id": "origin/claude/gv-001-hosted-re-review", + "last_commit": "2026-10-24", + "reachable_from_origin_main": true, + "type": "branch" + }, { "ahead_of_origin_main": 2, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "91e31c5", "id": "origin/claude/homepage-epistemic-motion", "last_commit": "2026-09-25", @@ -6732,7 +7103,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "1d44a13", "id": "origin/claude/k1-hold", "last_commit": "2026-09-22", @@ -6741,7 +7112,7 @@ }, { "ahead_of_origin_main": 3, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "d49480e", "id": "origin/claude/lucid-franklin-jloow0", "last_commit": "2026-09-26", @@ -6750,16 +7121,25 @@ }, { "ahead_of_origin_main": 4, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "057c7b1", "id": "origin/claude/phone-width-a11y", "last_commit": "2026-09-21", "reachable_from_origin_main": false, "type": "branch" }, + { + "ahead_of_origin_main": 3, + "behind_origin_main": 0, + "head": "e09bb66", + "id": "origin/claude/publication-gate-title-check", + "last_commit": "2026-10-25", + "reachable_from_origin_main": false, + "type": "branch" + }, { "ahead_of_origin_main": 4, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "e40fd8c", "id": "origin/claude/temple-researcher-outreach-on2v73", "last_commit": "2026-09-19", @@ -6768,7 +7148,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "3d5f743", "id": "origin/claude/verify-claims-refusal-unknown", "last_commit": "2026-09-27", @@ -6777,7 +7157,7 @@ }, { "ahead_of_origin_main": 2, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "1d59af6", "id": "origin/claude/x-gate", "last_commit": "2026-09-27", @@ -6786,7 +7166,7 @@ }, { "ahead_of_origin_main": 2, - "behind_origin_main": 0, + "behind_origin_main": 7, "head": "9259091", "id": "origin/codex/e2-e2", "last_commit": "2026-10-24", @@ -6795,7 +7175,7 @@ }, { "ahead_of_origin_main": 3, - "behind_origin_main": 0, + "behind_origin_main": 7, "head": "90328c0", "id": "origin/codex/e2-e3", "last_commit": "2026-10-24", @@ -6804,7 +7184,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 0, + "behind_origin_main": 7, "head": "a451d9b", "id": "origin/codex/e2-e4", "last_commit": "2026-10-24", @@ -6813,16 +7193,25 @@ }, { "ahead_of_origin_main": 4, - "behind_origin_main": 0, + "behind_origin_main": 7, "head": "1fe36e6", "id": "origin/codex/e2-e5", "last_commit": "2026-10-24", "reachable_from_origin_main": false, "type": "branch" }, + { + "ahead_of_origin_main": 0, + "behind_origin_main": 12, + "head": "9211af0", + "id": "origin/e9-001/register", + "last_commit": "2026-10-17", + "reachable_from_origin_main": true, + "type": "branch" + }, { "ahead_of_origin_main": 1, - "behind_origin_main": 114, + "behind_origin_main": 121, "head": "b9b7b1b", "id": "origin/icloud-snapshot/2026-09-19", "last_commit": "2026-09-19", @@ -6831,7 +7220,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 77, + "behind_origin_main": 84, "head": "d0ac440", "id": "origin/icloud-snapshot/2026-09-19-site", "last_commit": "2026-09-19", @@ -6841,15 +7230,15 @@ { "ahead_of_origin_main": 0, "behind_origin_main": 0, - "head": "e2cec7b", + "head": "9f617d1", "id": "origin/main", - "last_commit": "2026-10-17", + "last_commit": "2026-10-24", "reachable_from_origin_main": true, "type": "branch" }, { "ahead_of_origin_main": 1, - "behind_origin_main": 1, + "behind_origin_main": 8, "head": "5dd43af", "id": "origin/publication/title-social-square", "last_commit": "2026-10-17", @@ -6858,7 +7247,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 83, + "behind_origin_main": 90, "head": "79148e6", "id": "origin/rescue/commit-01fea45b4e98", "last_commit": "2026-09-17", @@ -6867,7 +7256,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 273, + "behind_origin_main": 280, "head": "119f257", "id": "origin/rescue/commit-119f257b763c", "last_commit": "2026-08-29", @@ -6876,7 +7265,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 92, + "behind_origin_main": 99, "head": "73e6691", "id": "origin/rescue/commit-2139046f7192", "last_commit": "2026-09-17", @@ -6885,7 +7274,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 166, + "behind_origin_main": 173, "head": "22fa9f9", "id": "origin/rescue/commit-22fa9f9a8456", "last_commit": "2026-09-05", @@ -6894,7 +7283,7 @@ }, { "ahead_of_origin_main": 3, - "behind_origin_main": 158, + "behind_origin_main": 165, "head": "4a362b5", "id": "origin/rescue/commit-4a362b5d714f", "last_commit": "2026-09-05", @@ -6903,7 +7292,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 178, + "behind_origin_main": 185, "head": "38e1a09", "id": "origin/rescue/commit-5e70a9df4626", "last_commit": "2026-09-02", @@ -6912,7 +7301,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 303, + "behind_origin_main": 310, "head": "1a17fe1", "id": "origin/rescue/commit-80250a3e7fc4", "last_commit": "2026-08-23", @@ -6921,7 +7310,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 273, + "behind_origin_main": 280, "head": "9320098", "id": "origin/rescue/commit-93200984af16", "last_commit": "2026-08-29", @@ -6930,7 +7319,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 170, + "behind_origin_main": 177, "head": "bd0c87d", "id": "origin/rescue/commit-bd0c87d0f1ea", "last_commit": "2026-09-05", @@ -6939,7 +7328,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 92, + "behind_origin_main": 99, "head": "089f9b9", "id": "origin/rescue/commit-c74849b7a928", "last_commit": "2026-09-17", @@ -6948,7 +7337,7 @@ }, { "ahead_of_origin_main": 4, - "behind_origin_main": 158, + "behind_origin_main": 165, "head": "cec26e5", "id": "origin/rescue/commit-cec26e585b5a", "last_commit": "2026-09-05", @@ -6957,7 +7346,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 165, + "behind_origin_main": 172, "head": "ea88d65", "id": "origin/rescue/commit-ea88d6552c2f", "last_commit": "2026-09-05", @@ -6966,7 +7355,7 @@ }, { "ahead_of_origin_main": 1, - "behind_origin_main": 234, + "behind_origin_main": 241, "head": "f8e5557", "id": "origin/rescue/commit-f8e555738fc2", "last_commit": "2026-08-31", @@ -6983,5 +7372,5 @@ } ], "schema": "repo-graph v0.1", - "stable_digest": "e54e1817f7c87033b51e5617132574ec0ed0cc9446fda9eb17b4f3df86189245" + "stable_digest": "7e9979007d6bd7d4365538072781fcc384934036f29e612166af27ebc9621c2a" } diff --git a/films/data/facts.json b/films/data/facts.json index fa9ddb7..4a4a03f 100644 --- a/films/data/facts.json +++ b/films/data/facts.json @@ -2,7 +2,7 @@ "_generated_by": "scripts/films/bind_facts.py — DO NOT EDIT; regenerate and re-inspect the films", "_inputs": { "census.yaml": "68b14081b85e18ca1a7b5bce33bbf38fe3b8dafe876e7329057241404037b425", - "claims.yaml": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004" + "claims.yaml": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd" }, "_kinds": [ "OBSERVED", @@ -1163,28 +1163,28 @@ { "evidential_status": "supported_within_scope", "id": "MC-002", - "last_reviewed": "2026-08-28", + "last_reviewed": "2026-11-02", "maturity": "experimental", "provenance": "machine_generated_owner_executed", - "review_due": "2026-10-27", + "review_due": "2027-01-01", "review_window_days": 60 }, { "evidential_status": "supported_within_scope", "id": "MC-003", - "last_reviewed": "2026-08-30", + "last_reviewed": "2026-11-02", "maturity": "experimental", "provenance": "machine_generated_owner_executed", - "review_due": "2026-10-29", + "review_due": "2027-01-01", "review_window_days": 60 }, { "evidential_status": "supported_within_scope", "id": "MC-004", - "last_reviewed": "2026-08-31", + "last_reviewed": "2026-11-02", "maturity": "experimental", "provenance": "machine_generated_owner_executed", - "review_due": "2026-10-30", + "review_due": "2027-01-01", "review_window_days": 60 }, { @@ -1291,7 +1291,7 @@ "REGISTRY.last_owner_review": { "kind": "REGISTRY", "source": "claims.yaml:last_owner_review", - "value": "2026-10-24" + "value": "2026-11-02" }, "REGISTRY.version": { "kind": "REGISTRY", diff --git a/films/flagship/source-map.json b/films/flagship/source-map.json index 4b457ba..5c9ffb4 100644 --- a/films/flagship/source-map.json +++ b/films/flagship/source-map.json @@ -4,8 +4,8 @@ "films/flagship/script.json": "db9bb0e5f475bf0dfc05085e71899eab9b0e47d39446e5850f22dd4404d0cca2", "scripts/films/build_flagship.py": "09248adf7cf047ffad802eadacb2122b35c5f9131fe6b6c46358aa6abfcffcfe", "films/flagship/rehearsal-template.html": "1d6e8f858360f6248108a249d62f0f206d1b9bb36e4cbaa21238ea475dc0e614", - "films/data/facts.json": "012eb01c4194078bea88d35cfb1e9ff68741329202ffd941733387835b9763bd", - "claims.yaml": "45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004", + "films/data/facts.json": "08fd6222062db6f50f0a0ebc4ffc88e68a947e43280e7875e55bbea959b1a0f9", + "claims.yaml": "ebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd", "census.yaml": "68b14081b85e18ca1a7b5bce33bbf38fe3b8dafe876e7329057241404037b425", "films/the-stack/manifest.yaml": "f81ae5c045665327da9e9bf8469e56d9f808820f81f422ccab731e87311bebe0", "films/the-multiplication/manifest.yaml": "fd6aeee56e9335eb6658e4e520142eaacfc85049ad178d44f32128cf70b3ba77", diff --git a/index.html b/index.html index dee61f5..4e46d5b 100644 --- a/index.html +++ b/index.html @@ -250,7 +250,7 @@

Bring one results table, claim, or receipt: Evidence ledger

machine-readable document.

@@ -192,7 +192,7 @@

MC-002

Supported within scope · Machine-generated, owner-executed

On the 170-prompt per-item subset publicly released with BELLS's 2025 misuse-detection evaluation (pinned commit below), the five specialized supervisors — Lakera Guard, Prompt Guard, LangKit, NeMo Guardrails, LLM Guard — have an OR-union of released binary verdict columns that flags 73 of the 82 prompts labelled harmful, leaving 9 all-miss (11.0%). The product of their individual miss rates is 3.5%; the release-recomputed all-miss rate is about 3.1× that independence plug-in on this file. The same OR-union flags 19 of the 50 benign prompts.

-
Scope
Exactly the released file at the bound commit: 170 of the evaluation's 990 non-adversarial prompts (the study's ~4,165 adversarial prompts have no per-item release), selected by the authors under an unstated rule. Counting arithmetic only, at the vendors' released binary verdicts; no population estimate, no uncertainty interval, and no claim about any system's true rates is offered or licensed. Computed by scripts/reanalyze_bells_subset.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified file.
Binding
non_adversarial_prompts.csv @ 507566c5
The upstream repository declares no license, so the file is cited and hash-verified (sha256 791dd4b0a168f2eb5831b308083a49 2e83200a9fa82585643c739023b03f57c3, asserted by the reproduction script), never redistributed from this record.
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
Reviewed
2026-08-28 · window 60 days
Triggers
  • executable fires when the released subset on the default branch changes
  • manual the authors release the full per-item dataset, state the subset's selection rule, or correct the released verdicts
Falsifier
Recomputing from the bound, hash-verified file yields any count different from the expected block, or the file at the pinned ref no longer matches the recorded sha256, or the released columns are shown not to be the labelled systems' verdicts.
CONSEQUENCE REJECT
Forbidden rescues
  • do not substitute a different subset, column set, or harm-level filter to preserve the numbers
  • do not fold borderline prompts into either denominator after seeing the results
  • do not recast this subset arithmetic as a population estimate, with or without an interval, if the primary counts are challenged
  • do not cite the release-recomputed-to-plug-in ratio without the selection caveat that scopes it
Non-claims
  • the subset is author-selected with an unstated rule — nothing here estimates any system's true rate, and no confidence interval is offered because the sampled population is undefined
  • not a ranking or endorsement of any vendor; the released verdicts are at unstated default configurations
  • says nothing about adversarial prompts, which dominate the full evaluation and have no per-item release
  • the release-recomputed-to-plug-in ratio describes this subset's arithmetic, not a general law of guardrail dependence
  • not an observed deployed five-guard stack; interpreting this static OR aggregation as a stack requires the separate full-exposure, parallel, fixed-operating-point assumptions
+
Scope
Exactly the released file at the bound commit: 170 of the evaluation's 990 non-adversarial prompts (the study's ~4,165 adversarial prompts have no per-item release), selected by the authors under an unstated rule. Counting arithmetic only, at the vendors' released binary verdicts; no population estimate, no uncertainty interval, and no claim about any system's true rates is offered or licensed. Computed by scripts/reanalyze_bells_subset.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified file.
Binding
non_adversarial_prompts.csv @ 507566c5
The upstream repository declares no license, so the file is cited and hash-verified (sha256 791dd4b0a168f2eb5831b308083a49 2e83200a9fa82585643c739023b03f57c3, asserted by the reproduction script), never redistributed from this record.
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
Reviewed
2026-11-02 · window 60 days
Triggers
  • executable fires when the released subset on the default branch changes
  • manual the authors release the full per-item dataset, state the subset's selection rule, or correct the released verdicts
Falsifier
Recomputing from the bound, hash-verified file yields any count different from the expected block, or the file at the pinned ref no longer matches the recorded sha256, or the released columns are shown not to be the labelled systems' verdicts.
CONSEQUENCE REJECT
Forbidden rescues
  • do not substitute a different subset, column set, or harm-level filter to preserve the numbers
  • do not fold borderline prompts into either denominator after seeing the results
  • do not recast this subset arithmetic as a population estimate, with or without an interval, if the primary counts are challenged
  • do not cite the release-recomputed-to-plug-in ratio without the selection caveat that scopes it
Non-claims
  • the subset is author-selected with an unstated rule — nothing here estimates any system's true rate, and no confidence interval is offered because the sampled population is undefined
  • not a ranking or endorsement of any vendor; the released verdicts are at unstated default configurations
  • says nothing about adversarial prompts, which dominate the full evaluation and have no per-item release
  • the release-recomputed-to-plug-in ratio describes this subset's arithmetic, not a general law of guardrail dependence
  • not an observed deployed five-guard stack; interpreting this static OR aggregation as a stack requires the separate full-exposure, parallel, fixed-operating-point assumptions
@@ -200,7 +200,7 @@

MC-003

Supported within scope · Machine-generated, owner-executed

For k guardrails scored on a common item set at a common operating point under block-on-any composition, the rate at which every guard misses the same item is fixed by the published per-guard miss rates only up to the interval [max(0, Σp − (k−1)), min p], and every point of that interval is attained by some joint law with those marginals. Marginals prove non-degradation of a static OR composition relative to its best member, but, when that identified set is non-degenerate, they do not identify strictly positive incremental benefit. They also identify a positive lower bound on static benign union-flag rate when any member has a positive benign flag rate. On the BELLS 2025 partial per-item release — 82 prompts labelled harmful — the exact finite identified set is {0/82 … 12/82}; MC-002 recomputes 9/82 all-miss from the hash-verified released verdict file. Marginal catch counts plus the union identify only bounds on each guard's exclusive full-stack coverage; the registered leave-one-out unions identify the realized values: NeMo 18, Lakera 3, and LangKit, Prompt Guard, and LLM Guard 0 on this stratum.

-
Scope
The mathematics is classical — Bonferroni for the lower endpoint, monotonicity for the upper, Fréchet for sharpness — and nothing here claims to have derived it. What is contributed is the application: pricing, in probability units, what marginal-only guardrail reporting leaves undetermined, and identifying leave-one-out unions as a compact privacy-preserving disclosure of exclusive full-stack coverage. They do not identify pairwise or higher-order overlap, Shapley values, or causal structure. Every statement is conditional on block-on-any composition, per-item decisions that are functions of the item alone, a fixed population, full exposure, and a fixed operating point; it says nothing about sequential routes where downstream measurement is censored by upstream blocks, agentic systems where an intervention changes the trajectory, or any vendor's product. The empirical half is one stratum of one author-selected subset: 82 harmful prompts, with 50 benign and 38 borderline held as separate strata and never folded in. It is a BELLS-specific reproduction, not an assertion that BELLS is the census's only per-item outcome release. scripts/reanalyze_bells_subset.py re-computes the BELLS counts from the hash-verified upstream file in CI; scripts/identification.py checks the stated identification transformations against that registered envelope.
Binding
identification.py
The file carries the analytic derivation in its docstring and checks explicit endpoint constructions numerically — the upper by nesting the events, the lower by laying the complements end to end around a circle. Its finite formulas are additionally exhaustively checked for n ≤ 4 and k ≤ 3.
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
Reviewed
2026-08-30 · window 60 days
Triggers
  • executable fires when the derivation or its tests change without a re-review; re-reviewed 2026-08-29 after removing a duplicate leave-one-out calculator and rechecking the registered outputs
  • manual a reader identifies an error in the stated Fréchet, finite-grid, or aggregate exclusive-coverage bounds, or a result is used outside the declared static full-exposure assumptions
Falsifier
A joint law with the stated marginals achieves an all-miss rate outside the interval; an exact finite catch-set arrangement lies outside the stated finite grid or exclusive-coverage bounds; or recomputing from the bound BELLS file yields any count different from the expected block.
CONSEQUENCE REJECT
Forbidden rescues
  • do not change the composition rule after the fact to preserve the direction of the result
  • do not restate "identified only up to" as "estimated to be" — a bound is not an estimate
  • do not quote where a release-recomputed value sits inside the interval as a score or a percentage of a gap closed; the endpoints come from adversarial couplings with no detector-behavioural content
  • do not cite the zero-exclusive-coverage results as a vendor ranking, a causal attribution, or evidence about any product outside this stratum
  • do not fold the borderline stratum into either denominator to change any figure
Non-claims
  • no new mathematics is claimed; the bounds are Fréchet's and the lower endpoint is Bonferroni's
  • not a claim that guardrails in general fail together — this stratum cannot separate shared blind spots from prompt-difficulty heterogeneity
  • not a vendor evaluation; the released verdicts are at unstated default configurations, and one supervisor fires exactly once in the 170 released rows
  • the identified interval is what the marginals leave open, not a prediction about any deployed stack
  • leave-one-out unions identify only exclusive full-stack coverage; they do not identify pairwise or higher-order overlap, Shapley values, or causal contribution
  • says nothing about adversarial prompts, which dominate the full evaluation and have no per-item release
+
Scope
The mathematics is classical — Bonferroni for the lower endpoint, monotonicity for the upper, Fréchet for sharpness — and nothing here claims to have derived it. What is contributed is the application: pricing, in probability units, what marginal-only guardrail reporting leaves undetermined, and identifying leave-one-out unions as a compact privacy-preserving disclosure of exclusive full-stack coverage. They do not identify pairwise or higher-order overlap, Shapley values, or causal structure. Every statement is conditional on block-on-any composition, per-item decisions that are functions of the item alone, a fixed population, full exposure, and a fixed operating point; it says nothing about sequential routes where downstream measurement is censored by upstream blocks, agentic systems where an intervention changes the trajectory, or any vendor's product. The empirical half is one stratum of one author-selected subset: 82 harmful prompts, with 50 benign and 38 borderline held as separate strata and never folded in. It is a BELLS-specific reproduction, not an assertion that BELLS is the census's only per-item outcome release. scripts/reanalyze_bells_subset.py re-computes the BELLS counts from the hash-verified upstream file in CI; scripts/identification.py checks the stated identification transformations against that registered envelope.
Binding
identification.py
The file carries the analytic derivation in its docstring and checks explicit endpoint constructions numerically — the upper by nesting the events, the lower by laying the complements end to end around a circle. Its finite formulas are additionally exhaustively checked for n ≤ 4 and k ≤ 3.
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
Reviewed
2026-11-02 · window 60 days
Triggers
  • executable fires when the derivation or its tests change without a re-review; re-reviewed 2026-08-29 after removing a duplicate leave-one-out calculator and rechecking the registered outputs
  • manual a reader identifies an error in the stated Fréchet, finite-grid, or aggregate exclusive-coverage bounds, or a result is used outside the declared static full-exposure assumptions
Falsifier
A joint law with the stated marginals achieves an all-miss rate outside the interval; an exact finite catch-set arrangement lies outside the stated finite grid or exclusive-coverage bounds; or recomputing from the bound BELLS file yields any count different from the expected block.
CONSEQUENCE REJECT
Forbidden rescues
  • do not change the composition rule after the fact to preserve the direction of the result
  • do not restate "identified only up to" as "estimated to be" — a bound is not an estimate
  • do not quote where a release-recomputed value sits inside the interval as a score or a percentage of a gap closed; the endpoints come from adversarial couplings with no detector-behavioural content
  • do not cite the zero-exclusive-coverage results as a vendor ranking, a causal attribution, or evidence about any product outside this stratum
  • do not fold the borderline stratum into either denominator to change any figure
Non-claims
  • no new mathematics is claimed; the bounds are Fréchet's and the lower endpoint is Bonferroni's
  • not a claim that guardrails in general fail together — this stratum cannot separate shared blind spots from prompt-difficulty heterogeneity
  • not a vendor evaluation; the released verdicts are at unstated default configurations, and one supervisor fires exactly once in the 170 released rows
  • the identified interval is what the marginals leave open, not a prediction about any deployed stack
  • leave-one-out unions identify only exclusive full-stack coverage; they do not identify pairwise or higher-order overlap, Shapley values, or causal contribution
  • says nothing about adversarial prompts, which dominate the full evaluation and have no per-item release
@@ -208,7 +208,7 @@

MC-004

Supported within scope · Machine-generated, owner-executed

On the per-item guard-verdict files publicly released with Multimodal Safeguard Bench's full_run (pinned commit below; six verdict files and the release's two printed-metric files hash-verified), the three guards — Llama Guard 4, Llama Guard 3 Vision, ShieldGemma 2 — have harness-normalized native `unsafe` labels whose static Boolean OR is 1 on 192 of 200 harmful-labelled text items (8 all-zero rows, 4.0%) and all 200 harmful-labelled image items (0 all-zero rows), while the same OR is 1 on 32 of 250 benign-labelled text items (12.8%) and every one of 250 benign-labelled image items. Llama Guard 3 Vision alone has a 1-bit on all 250 released benign image items. The complete 2^3 native-label pattern table, leave-one-out bit ORs, and unique bit contributions for each released stratum are bound below. The release prints per-guard metrics and two two-guard compositions; it prints no three-guard bit OR, no all-zero-row count, no leave-one-out table, and no pattern decomposition — and every recomputed quantity that overlaps what it does print is asserted equal to the printed value.

-
Scope
Exactly the six committed verdict files of the full_run directory at the bound commit: 400 harmful and 500 benign items, each stratum half text and half image, one Boolean `blocked` adapter bit per item per guard. Let L be the release's harmful/benign label, A_s each guard's native `unsafe` predicate, B_s the harness mapping of that predicate to `blocked`, and O = OR_s B_s. The bound arithmetic reports B and O conditional on L. Because the harness uses `blocked` items to suppress target generation, O is also a valid counterfactual harness-block decision for these released rows under a fixed block-on-any rule. It does not establish a source-defined translation from every A_s to one shared catch event E: Llama Guard 3 Vision is a multimodal prompt/response classifier, while ShieldGemma 2 is an image-only three-policy classifier, and the pipeline merely maps each native `unsafe` label to `blocked`. Counting arithmetic only, at the guards' released native operating rules (the Llama guards emit labels autoregressively; ShieldGemma 2 blocks above a 0.5 policy-violation probability; the source documents no matched operating-point calibration, so M-ladder threshold comparability is unchanged). ShieldGemma 2's text-item bits are deterministic passes as committed and documented upstream, which is why its text column is zero — an explicit released outcome, not missing data. Strata are computed exactly as released and never folded; file totals are sums of the bound strata and carry no additional content. Computed by scripts/reanalyze_msbench.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified files.
Binding
PatrickKollman/Multimodal-Safeguard-Bench/tree/fb6f32e6b50b6faad833b815ebcd80afd2068bff/results/full_run @ fb6f32e6
Six guard-verdict JSONL files are the data inputs; metrics.json and ensemble.json are additionally pinned because the reproduction script asserts agreement with every overlapping printed value. All eight sha256 hashes are recorded in scripts/reanalyze_msbench.py and verified before a single count is taken.
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
Reviewed
2026-08-31 · window 60 days
Triggers
  • executable fires when a bound verdict file on the default branch changes (6 bound sources)
  • executable fires when the release's printed per-guard metrics change
  • executable fires when the release's printed composition metrics change
  • executable fires when the guards' native adapter semantics change
  • executable fires when the harness label-to-block action changes
  • executable fires when the stated route or operating-point semantics change
  • manual the authors publish a corrected or extended full_run, print their own three-guard joint statistics, document a matched operating-point calibration, or publish a source-defined translation from all native predicates to one shared catch event
Falsifier
Recomputing from the bound, hash-verified files yields any count different from the expected block, or any pinned file no longer matches its recorded sha256, or any recomputed quantity that overlaps the release's printed metrics disagrees with the printed value, or the committed blocked columns are shown not to be the named guards' verdicts.
CONSEQUENCE REJECT
Forbidden rescues
  • do not fold text and image strata — or harmful and benign files — into pooled denominators to move any figure; the strata are bound exactly as released
  • do not substitute verdicts from the release's other run directories (adaptive_run, carrier sweeps, rendering probes) to preserve a number bound to full_run
  • do not reinterpret ShieldGemma 2's deterministic text passes as missing data to shrink a denominator after seeing the results
  • do not cite the benign-image 250/250 union without attributing it to Llama Guard 3 Vision's 250/250 column, and do not cite the harmful strata without the benign strata
  • do not call an OR of the harness-normalized native labels a shared-event catch union, stack safety result, or three-independent- guard finding unless a source-defined event translation is added
  • do not recast these file counts as population estimates, with or without an interval, if the primary counts are challenged
Non-claims
  • not a shared-event catch statistic: the common `blocked` bit is a harness normalization of distinct native predicates, and no source-defined translation to a common event E has been identified
  • not a population estimate — the items derive from HarmBench and XSTest under the release's own construction, and no interval is offered because no sampled population is defined
  • not a ranking, endorsement, or indictment of any guard or vendor; the verdicts are at native, unmatched operating rules
  • not an observed deployed three-guard stack; interpreting the OR aggregation as a stack requires the separate full-exposure, parallel, fixed-operating-point assumptions
  • says nothing about the release's carrier-prompt, adversarial-UAP, or cross-VLM runs, which are separate artifacts with their own contracts
  • the counts are about the committed verdict bytes; the release's own changelog documents that ShieldGemma 2's image scores are sensitive to the text-rendering stack, so nothing here predicts what any guard would do under a different rendering environment
  • the zero three-guard image all-zero-bit count is a counting fact about these 200 released image items, not evidence of general image attack safety
+
Scope
Exactly the six committed verdict files of the full_run directory at the bound commit: 400 harmful and 500 benign items, each stratum half text and half image, one Boolean `blocked` adapter bit per item per guard. Let L be the release's harmful/benign label, A_s each guard's native `unsafe` predicate, B_s the harness mapping of that predicate to `blocked`, and O = OR_s B_s. The bound arithmetic reports B and O conditional on L. Because the harness uses `blocked` items to suppress target generation, O is also a valid counterfactual harness-block decision for these released rows under a fixed block-on-any rule. It does not establish a source-defined translation from every A_s to one shared catch event E: Llama Guard 3 Vision is a multimodal prompt/response classifier, while ShieldGemma 2 is an image-only three-policy classifier, and the pipeline merely maps each native `unsafe` label to `blocked`. Counting arithmetic only, at the guards' released native operating rules (the Llama guards emit labels autoregressively; ShieldGemma 2 blocks above a 0.5 policy-violation probability; the source documents no matched operating-point calibration, so M-ladder threshold comparability is unchanged). ShieldGemma 2's text-item bits are deterministic passes as committed and documented upstream, which is why its text column is zero — an explicit released outcome, not missing data. Strata are computed exactly as released and never folded; file totals are sums of the bound strata and carry no additional content. Computed by scripts/reanalyze_msbench.py from scripts/mjgd_reference.py; the expected block below is what CI re-asserts against the downloaded, hash-verified files.
Binding
PatrickKollman/Multimodal-Safeguard-Bench/tree/fb6f32e6b50b6faad833b815ebcd80afd2068bff/results/full_run @ fb6f32e6
Six guard-verdict JSONL files are the data inputs; metrics.json and ensemble.json are additionally pinned because the reproduction script asserts agreement with every overlapping printed value. All eight sha256 hashes are recorded in scripts/reanalyze_msbench.py and verified before a single count is taken.
Dimensions
visibilityPublicprovenanceMachine-generated, owner-executedsupport roleExecuted outputmaturityExperimental
Reviewed
2026-11-02 · window 60 days
Triggers
  • executable fires when a bound verdict file on the default branch changes (6 bound sources)
  • executable fires when the release's printed per-guard metrics change
  • executable fires when the release's printed composition metrics change
  • executable fires when the guards' native adapter semantics change
  • executable fires when the harness label-to-block action changes
  • executable fires when the stated route or operating-point semantics change
  • manual the authors publish a corrected or extended full_run, print their own three-guard joint statistics, document a matched operating-point calibration, or publish a source-defined translation from all native predicates to one shared catch event
Falsifier
Recomputing from the bound, hash-verified files yields any count different from the expected block, or any pinned file no longer matches its recorded sha256, or any recomputed quantity that overlaps the release's printed metrics disagrees with the printed value, or the committed blocked columns are shown not to be the named guards' verdicts.
CONSEQUENCE REJECT
Forbidden rescues
  • do not fold text and image strata — or harmful and benign files — into pooled denominators to move any figure; the strata are bound exactly as released
  • do not substitute verdicts from the release's other run directories (adaptive_run, carrier sweeps, rendering probes) to preserve a number bound to full_run
  • do not reinterpret ShieldGemma 2's deterministic text passes as missing data to shrink a denominator after seeing the results
  • do not cite the benign-image 250/250 union without attributing it to Llama Guard 3 Vision's 250/250 column, and do not cite the harmful strata without the benign strata
  • do not call an OR of the harness-normalized native labels a shared-event catch union, stack safety result, or three-independent- guard finding unless a source-defined event translation is added
  • do not recast these file counts as population estimates, with or without an interval, if the primary counts are challenged
Non-claims
  • not a shared-event catch statistic: the common `blocked` bit is a harness normalization of distinct native predicates, and no source-defined translation to a common event E has been identified
  • not a population estimate — the items derive from HarmBench and XSTest under the release's own construction, and no interval is offered because no sampled population is defined
  • not a ranking, endorsement, or indictment of any guard or vendor; the verdicts are at native, unmatched operating rules
  • not an observed deployed three-guard stack; interpreting the OR aggregation as a stack requires the separate full-exposure, parallel, fixed-operating-point assumptions
  • says nothing about the release's carrier-prompt, adversarial-UAP, or cross-VLM runs, which are separate artifacts with their own contracts
  • the counts are about the committed verdict bytes; the release's own changelog documents that ShieldGemma 2's image scores are sensitive to the text-rendering stack, so nothing here predicts what any guard would do under a different rendering environment
  • the zero three-guard image all-zero-bit count is a counting fact about these 200 released image items, not evidence of general image attack safety
diff --git a/observatory/index.html b/observatory/index.html index 6f9b5df..0777510 100644 --- a/observatory/index.html +++ b/observatory/index.html @@ -275,7 +275,7 @@

MC-001

4 non-claims
-
+

MC-002

Supported within scope @@ -285,7 +285,7 @@

MC-002

Machine-generated, owner-executed · Experimental - reviewed 2026-08-28 · window 60d + reviewed 2026-11-02 · window 60d falsifier consequence: REJECT · 4 forbidden rescues executablemanual
@@ -295,7 +295,7 @@

MC-002

5 non-claims
-
+

MC-003

Supported within scope @@ -305,7 +305,7 @@

MC-003

Machine-generated, owner-executed · Experimental - reviewed 2026-08-30 · window 60d + reviewed 2026-11-02 · window 60d falsifier consequence: REJECT · 5 forbidden rescues executablemanual
@@ -315,7 +315,7 @@

MC-003

6 non-claims
-
+

MC-004

Supported within scope @@ -325,7 +325,7 @@

MC-004

Machine-generated, owner-executed · Experimental - reviewed 2026-08-31 · window 60d + reviewed 2026-11-02 · window 60d falsifier consequence: REJECT · 6 forbidden rescues executableexecutableexecutableexecutableexecutableexecutableexecutableexecutableexecutableexecutableexecutablemanual
@@ -595,7 +595,7 @@

Replay manifest

Human review

-

Last owner review: 2026-10-24. These review +

Last owner review: 2026-11-02. These review events cannot be executed by CI, and the record says so instead of borrowing the executable triggers' credibility:

  • CC-001 — kernel semantics change outside the bound module file
  • CC-003 — E2 produces empirical results — the synthetic scoping here must be restated
  • CC-004 — witness API change outside the bound module file
  • CC-005 — a conforming dataset is collected — E2's status must be restated
  • CC-006 — real E2 data is collected — this claim is superseded by an E2 result
  • REL-001 — a public protocol revision is bound, a real E2 result activates or prevents activation, or participant recruitment begins
  • MC-001 — a new qualifying evaluation is published or reported, an author corrects a row, or a benchmark publishes a previously absent joint statistic
  • MC-002 — the authors release the full per-item dataset, state the subset's selection rule, or correct the released verdicts
  • MC-003 — a reader identifies an error in the stated Fréchet, finite-grid, or aggregate exclusive-coverage bounds, or a result is used outside the declared static full-exposure assumptions
  • MC-004 — the authors publish a corrected or extended full_run, print their own three-guard joint statistics, document a matched operating-point calibration, or publish a source-defined translation from all native predicates to one shared catch event
  • AF-001 — the report publishes the artifact share, states which comparison group the differences are against, corrects a figure, or expands the analysis to a different sample
  • GV-001 — the hosted build is removed, moves from https://cubits11.github.io/ghost-visualizer/, or stops serving the bound revision's Evidence Mode — the scope sentence is then wrong
  • GCE-001 — GCE repository resumed or reframed
  • SITE-001 — site redesign
  • SITE-002 — workflow triggers or cadence change
  • E3-001 — a third pilot is run on a pool where both guards' miss rates are intermediate
  • E3B-001 — the contamination status of either guard's training corpus becomes checkable
  • E6-001 — Anthropic publishes the model's output at the Table 2 median vector, releases the 3,259 five-vectors, or ships a kernel whose digest differs from the pinned bytes
  • E7B-001 — a pool with intermediate marginals and more than 21 harmful items becomes available, or the upstream repository's scores change at a new commit
  • E9-001 — a frozen multi-pool or new-guard contract (E10) is executed, either pinned revision is withdrawn, or G2's training data is disclosed
  • diff --git a/records/conformance/index.html b/records/conformance/index.html index 1af957f..3965561 100644 --- a/records/conformance/index.html +++ b/records/conformance/index.html @@ -147,7 +147,7 @@

    Sources

    time. Hashes are of the bytes this page was generated from.

    - +
    filesha256
    research/DIRECTION.yaml92ed64289f8dae5bbc577053e7f11d9e4551a0cc415b0c28bb62150291d691d1
    metrics/ledger_snapshot.json5f06783231a669d9b161186b5be18dbb83836ff1e2c57ffe57166042be69f460
    corrections/records/2026-09-10.jsona2cca203fd92ffe17ec09d841e90154cfd9cf966f4aeb1b1a62ad0c223cb2cc0
    corrections/records/2026-09-10-preserved-inputs.jsonb20ad976edcfcccf2a8fa7fa8db609a3b735bb5a2ab1a75c459395fe437a100d
    scripts/verify_prereg.py103f77853671d0041b6621d8a8afd5d868dfc907d41b65fdd5ba043de6ba7dc3
    claims.yaml45790a72bbccd9bafb0d297d5128d59e357e3ef759e18f03733482cef621a004
    research/DIRECTION.yaml92ed64289f8dae5bbc577053e7f11d9e4551a0cc415b0c28bb62150291d691d1
    metrics/ledger_snapshot.json5f06783231a669d9b161186b5be18dbb83836ff1e2c57ffe57166042be69f460
    corrections/records/2026-09-10.jsona2cca203fd92ffe17ec09d841e90154cfd9cf966f4aeb1b1a62ad0c223cb2cc0
    corrections/records/2026-09-10-preserved-inputs.jsonb20ad976edcfcccf2a8fa7fa8db609a3b735bb5a2ab1a75c459395fe437a100d
    scripts/verify_prereg.py103f77853671d0041b6621d8a8afd5d868dfc907d41b65fdd5ba043de6ba7dc3
    claims.yamlebb5887be3009e73cb04e99a12a2d71e4ef9fa8da340b386ad7a1e824211febd
diff --git a/sitemap.xml b/sitemap.xml index 998695b..54a7de0 100644 --- a/sitemap.xml +++ b/sitemap.xml @@ -1,7 +1,7 @@ - https://cubits11.github.io/2026-10-24 + https://cubits11.github.io/2026-11-02 https://cubits11.github.io/answers/how-to-evaluate-guardrails-you-plan-to-stack/2026-09-01 https://cubits11.github.io/answers/what-does-the-second-guardrail-add/2026-09-01 https://cubits11.github.io/answers/why-guardrail-miss-rates-do-not-multiply/2026-09-01 @@ -21,11 +21,11 @@ https://cubits11.github.io/claims/ga-001/2026-10-17 https://cubits11.github.io/claims/gce-001/2026-10-17 https://cubits11.github.io/claims/gv-001/2026-10-24 - https://cubits11.github.io/claims/2026-10-24 + https://cubits11.github.io/claims/2026-11-02 https://cubits11.github.io/claims/mc-001/2026-10-17 - https://cubits11.github.io/claims/mc-002/2026-10-17 - https://cubits11.github.io/claims/mc-003/2026-10-17 - https://cubits11.github.io/claims/mc-004/2026-10-17 + https://cubits11.github.io/claims/mc-002/2026-11-02 + https://cubits11.github.io/claims/mc-003/2026-11-02 + https://cubits11.github.io/claims/mc-004/2026-11-02 https://cubits11.github.io/claims/rel-001/2026-10-17 https://cubits11.github.io/claims/site-001/2026-10-17 https://cubits11.github.io/claims/site-002/2026-10-17 @@ -57,7 +57,7 @@ https://cubits11.github.io/foundations/d-010/2026-09-06 https://cubits11.github.io/foundations/2026-09-06 https://cubits11.github.io/foundations/meshes/2026-09-06 - https://cubits11.github.io/ledger/2026-10-24 + https://cubits11.github.io/ledger/2026-11-02 https://cubits11.github.io/limits/2026-10-17 https://cubits11.github.io/missing-column/disclosure/2026-09-06 https://cubits11.github.io/missing-column/2026-09-15 @@ -73,12 +73,12 @@ https://cubits11.github.io/notes/noetic-log-002/2026-10-17 https://cubits11.github.io/now/2026-10-17 https://cubits11.github.io/ns/falsifiable/v1/2026-10-17 - https://cubits11.github.io/observatory/2026-10-24 + https://cubits11.github.io/observatory/2026-11-02 https://cubits11.github.io/overlap/2026-09-15 - https://cubits11.github.io/records/conformance/2026-10-24 + https://cubits11.github.io/records/conformance/2026-11-02 https://cubits11.github.io/resume/2026-10-17 https://cubits11.github.io/stack-study/2026-09-15 - https://cubits11.github.io/status/2026-10-24 + https://cubits11.github.io/status/2026-11-02 https://cubits11.github.io/status/preview/2026-10-17 https://cubits11.github.io/trials/necromancer/2026-09-02 https://cubits11.github.io/try/2026-10-17 diff --git a/status/index.html b/status/index.html index 685a0b8..a606028 100644 --- a/status/index.html +++ b/status/index.html @@ -107,7 +107,7 @@

Checks, review dates and pending work

Check outcomes and overdue revi

GitHub is read when this page opens. A failed or unavailable request shows UNKNOWN. No workflow result establishes scientific validity.

Exercise the failed and unavailable preview states · Inspect scientific check coverage · What receipts and checks establish

Review queue

Claim review dates are calculated from the registry's last review plus its review window. Recorded renewals retain the earlier deadline, calculated from the pinned registry revision. Other dated promises come from the publication list. Static snapshot: 2026-10-17; JavaScript updates labels against today's date in New York.

- +

Research pipeline

Published result — claim registration pending · observed 2026-10-17

E9 — 1,000-item scoped result

E9’s 1,000-item result merged in PR 26. A post-result claim registration, E9-001, is proposed in draft PR 31 and is not in the registry. The result is not a completion of E2 and does not establish deployment safety.

The result discloses a verdict rule outside readable preregistration prose, selected pools, omitted calibration uncertainty and no independent replication. Its interval excludes zero but does not lie wholly beyond the 0.05 effect-size threshold.

Read the pinned result · Inspect publication record