Conversation
…CAL + 4 IMPORTANT + 3 WATCH + 1 DISMISS) Fourth Aminata adversarial pass this session. Third on the Otto composition stack (Otto-90 oracle-scoring v0 → Otto-94 iteration-1 on multi-Claude → Otto-99 detector → this pass). Reviews Otto-99's provenance-aware bullshit-detector design (PR #282) which explicitly named "Aminata adversarial pass" as dependency #1 of adoption. 3 CRITICAL findings: 1. **Cross-detector collusion** — detector + Aminata + Codex all share training-corpus / repo-access / PR-comment lineage. Reviewer-set lineage-coupling reintroduces carrier-laundering one meta-layer up. Detector's SD-9 operationalisation does not break this. 2. **Min-merging on ordinal bands is Goodhart-bait at G_carrier_overlap** — most adversary-manipulable gate drives whole output. Author controls citation structure = author controls carrier-overlap gate. Sparse citations + dense implicit framing-inheritance = gameable GREEN. 3. **G_evidence fig-leaf** — advisory-only until independent- oracle substrate ships; v0 detector is four-gate not five-gate; `likely confabulated` output type is dead code until dependency #3 lands. Not stated as v0 caveat in design. 4 IMPORTANT: no-signal vs kNN-evasion indistinguishability; Otto-wake second-review insufficient for status-pin independence; DetectorOutputRetracted flood-control absent; G_coverage_plausibility gate missing. 3 WATCH: worst-band masks distribution; self-demo is theatre not validation; composition-stack silent-failure surface absent TLA+ invariants (Soraya-routable). 1 DISMISS: parameter-ADR gate reused from oracle-scoring v0. None block the research-doc land (Otto-99 correctly framed research-grade). **All ten findings would block a v1 implementation-ADR.** Non-fusion disclaimer load-bearing in this pass: Aminata's concordance with prior Aminata passes is same-agent signal NOT independent concordance. The detector applied to THIS review would correctly emit `looks similar but lineage- coupled` per SD-9 — and it would be right. Archive-header format self-applied — 17th aurora/research doc in a row. Lands within-standing-authority per Otto-82/90/93 calibration — advisory research-grade; not a gate. Otto-100 milestone tick. Closes dependency #1 of adoption path named in Otto-99 (PR #282). Next natural step is Otto integrating CRITICAL findings at write-time into a v1 detector-design revision OR scheduling that integration to a future tick. Otto-100 tick primary deliverable.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 4ecadfdba2
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
|
||
| ## Summary | ||
|
|
||
| Three CRITICAL, four IMPORTANT, three WATCH, one DISMISS. |
There was a problem hiding this comment.
The summary line reports Three CRITICAL, four IMPORTANT, three WATCH, one DISMISS, but the document body defines only 10 finding headings (2 CRITICAL + 4 IMPORTANT + 3 WATCH + 1 DISMISS), and later also says All ten findings. This mismatch makes the triage signal ambiguous (it implies an extra critical finding that is not actually enumerated), which can skew downstream prioritization and ADR decisions.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
Pull request overview
Adds a new research-grade adversarial review document capturing Aminata’s 4th-pass findings on the provenance-aware “bullshit-detector” design (PR #282), including CRITICAL/IMPORTANT/WATCH items and concrete asks.
Changes:
- Introduces a structured findings write-up (3 CRITICAL, 4 IMPORTANT, 3 WATCH, 1 DISMISS) with rationale and requested follow-ups.
- Summarizes operational status and non-fusion disclaimer to frame authority/independence boundaries.
- Adds a “Relevant paths” section intended to connect this pass to related design artifacts.
Comment on lines
+41
to
+45
| query that trips the detector's `looks lineage-coupled` | ||
| band will likely also trip Aminata into the same band of | ||
| concern — because Aminata's cone and the detector's cone | ||
| overlap by construction. The design names SD-9 | ||
| operationalisation but does NOT address reviewer-set |
There was a problem hiding this comment.
The output-type label here (looks lineage-coupled) is inconsistent with the earlier/elsewhere label looks similar but lineage-coupled. This kind of terminology drift makes it hard to grep/track outputs and can cause downstream mismatches if the string becomes part of receipts/logs. Pick one canonical name (preferably the one defined in the detector design) and use it consistently throughout this doc.
Suggested change
| query that trips the detector's `looks lineage-coupled` | |
| band will likely also trip Aminata into the same band of | |
| concern — because Aminata's cone and the detector's cone | |
| overlap by construction. The design names SD-9 | |
| operationalisation but does NOT address reviewer-set | |
| query that trips the detector's `looks similar but | |
| lineage-coupled` band will likely also trip Aminata into | |
| the same band of concern — because Aminata's cone and the | |
| detector's cone overlap by construction. The design names | |
| SD-9 operationalisation but does NOT address reviewer-set |
|
|
||
| **Ask.** Require `second-reviewer` to be one of: | ||
| different persona (Aminata / Kenji / other named | ||
| specialist), different model (Codex), or human (Aaron). |
There was a problem hiding this comment.
This line names the human maintainer directly ("human (Aaron)"), but the repo convention is to use role references (e.g., "human maintainer") in docs and code. See docs/AGENT-BEST-PRACTICES.md:284-290 for the policy. Please rewrite to a role-ref while keeping the meaning of the independence requirement.
Suggested change
| specialist), different model (Codex), or human (Aaron). | |
| specialist), different model (Codex), or human maintainer. |
Comment on lines
+291
to
+293
| - [`docs/research/provenance-aware-bullshit-detector-2026-04-23.md`](provenance-aware-bullshit-detector-2026-04-23.md) | ||
| (under review, PR #282). | ||
| - [`docs/research/semantic-canonicalization-and-provenance-aware-retrieval-2026-04-23.md`](semantic-canonicalization-and-provenance-aware-retrieval-2026-04-23.md) |
There was a problem hiding this comment.
These relative links currently point at files that do not exist in the repository (provenance-aware-bullshit-detector-2026-04-23.md and semantic-canonicalization-and-provenance-aware-retrieval-2026-04-23.md). If the intent is to reference work that is only present in other open PRs, consider linking to the PRs (or to permanent commit/PR URLs) instead of a local relative path, so the links don't break when this doc lands independently.
Suggested change
| - [`docs/research/provenance-aware-bullshit-detector-2026-04-23.md`](provenance-aware-bullshit-detector-2026-04-23.md) | |
| (under review, PR #282). | |
| - [`docs/research/semantic-canonicalization-and-provenance-aware-retrieval-2026-04-23.md`](semantic-canonicalization-and-provenance-aware-retrieval-2026-04-23.md) | |
| - `docs/research/provenance-aware-bullshit-detector-2026-04-23.md` | |
| (under review, PR #282). | |
| - `docs/research/semantic-canonicalization-and-provenance-aware-retrieval-2026-04-23.md` |
Comment on lines
+301
to
+303
| - [`docs/DRIFT-TAXONOMY.md`](../DRIFT-TAXONOMY.md) | ||
| pattern 5 — real-time diagnostic the detector aims to | ||
| mechanise. |
There was a problem hiding this comment.
This link points to docs/DRIFT-TAXONOMY.md, which is not present in the repo, so it will be broken in rendered Markdown. Either update the link to an existing drift-taxonomy document (e.g. the research precursor) or add/land the referenced operational doc before linking to it here.
Suggested change
| - [`docs/DRIFT-TAXONOMY.md`](../DRIFT-TAXONOMY.md) | |
| pattern 5 — real-time diagnostic the detector aims to | |
| mechanise. | |
| - Drift taxonomy pattern 5 — real-time diagnostic the | |
| detector aims to mechanise. |
AceHack
added a commit
that referenced
this pull request
Apr 24, 2026
…ngs integrated (Otto-101) Delta-style revision integrating only the 3 CRITICAL findings from Aminata's Otto-100 pass (PR #284) into Otto-99's bullshit-detector design (PR #282). 7 non-CRITICAL findings deferred to v2 delta; DISMISS unchanged. Three CRITICAL integrations: **C1 — Cross-detector collusion (new §Reviewer-cone overlap section).** Acknowledges as FUNDAMENTAL LIMITATION NOT CLOSABLE GAP: detector cannot break its own reviewer-set lineage coupling. Mitigation via routing authority: - Human sign-off = only cone-breaking authority for authorization-impacting band=supported queries. - External-reviewer signatures as legitimate cone-breakers when available. - Intra-factory reviewer concordance should DOWNGRADE confidence per SD-9, not inflate. - Every receipt carries reviewer_cone_signature field. **C2 — Min-merging Goodhart-bait (sensitivity-analysis-gate pattern).** v1 enhancement: band(y|q) = if band_raw=GREEN AND G_carrier_overlap closest to threshold → YELLOW else band_raw Closes cheap cheat path (threshold-hugging on author- controlled gate). Alternative content-similarity-in-cone deferred to v2 cost-benefit. **C3 — G_evidence fig-leaf (new §v0 scope subsection).** v0 effective = 4-gate classifier (G_evidence advisory-only until independent-oracle substrate ships). `likely confabulated` output type explicitly marked not-yet-reachable in v0. Conservative under-detection stance (RED comes only from known-bad-pattern; confabulations land YELLOW as plausible- but-unresolved). v0→v1 transition plan names DetectorOutputBatchRetracted for historical re-classification. v1 delta does NOT rewrite Otto-99's design; specifies 3 additive sections to insert + 1 section to supersede. Otto-99 original preserved in git history. Seven non-CRITICALs still open: - 4 IMPORTANT (deferred to v2): G_coverage_plausibility gate; Otto-wake second-reviewer schema; DetectorOutputBatchRetracted; no-signal vs kNN-evasion. - 3 WATCH (deferred to v2+): distribution histogram; adversarial worked example; TLA+ invariants. 1 fundamental limitation (C1) acknowledged not closed — routes authority to human + external reviewer layers. Scope limits: no design rewrite; no implementation; no human-sign-off UI proposal; no content-similarity-in-cone commit; 5-gate/5-type target structure unchanged. 5 dependencies-to-adoption: Aminata pass on v1 delta (fifth session-pass); integrate v1 changes into Otto-99 design PR (separate PR); v2 delta; independent-oracle substrate; human-sign-off UI/protocol. Archive-header format self-applied — 18th aurora/research doc in a row. Lands within-standing-authority per Otto-82/90/93 calibration. Otto-101 tick primary deliverable — closes the CRITICAL- integration step of the Aminata-then-Otto-response loop for bullshit-detector design.
AceHack
added a commit
that referenced
this pull request
Apr 24, 2026
…session-pattern convergence observable Milestone tick 100. PR #284 Aminata's fourth pass this session surfaces 3 CRITICAL + 4 IMPORTANT + 3 WATCH + 1 DISMISS on the bullshit-detector design. Closes dependency #1 of Otto-99's adoption path. Key observations at milestone: 1. Session-lifetime patterns now observable: - CC-002 discipline reflexive across 8 ferries - Event+view module template reused across 4 designs (substrate convergence) - Authority-calibration narrowed 3x (Otto-82/90/93) - SD-9 exercised twice by Amara at author-side - 8th-ferry 4/5 closed matching 5th-ferry arc 2. Aminata's own non-fusion disclaimer in this pass is the cleanest SD-9 worked example this session — explicitly names same-agent concordance as signal not evidence, consistent with her CRITICAL #1 cross-detector-collusion finding. 3. 10 findings to integrate before detector v1. Not all at Otto-100; progressive Otto-101+ work. 4. Queue at ~30 open auto-merge-armed PRs; not a bottleneck per Otto-72 don't-wait; visibility observation only. Stacked on #283 (Otto-99 history).
AceHack
added a commit
that referenced
this pull request
Apr 24, 2026
…session-pattern convergence observable Milestone tick 100. PR #284 Aminata's fourth pass this session surfaces 3 CRITICAL + 4 IMPORTANT + 3 WATCH + 1 DISMISS on the bullshit-detector design. Closes dependency #1 of Otto-99's adoption path. Key observations at milestone: 1. Session-lifetime patterns now observable: - CC-002 discipline reflexive across 8 ferries - Event+view module template reused across 4 designs (substrate convergence) - Authority-calibration narrowed 3x (Otto-82/90/93) - SD-9 exercised twice by Amara at author-side - 8th-ferry 4/5 closed matching 5th-ferry arc 2. Aminata's own non-fusion disclaimer in this pass is the cleanest SD-9 worked example this session — explicitly names same-agent concordance as signal not evidence, consistent with her CRITICAL #1 cross-detector-collusion finding. 3. 10 findings to integrate before detector v1. Not all at Otto-100; progressive Otto-101+ work. 4. Queue at ~30 open auto-merge-armed PRs; not a bottleneck per Otto-72 don't-wait; visibility observation only. Stacked on #283 (Otto-99 history).
AceHack
added a commit
that referenced
this pull request
Apr 24, 2026
…session-pattern convergence observable Milestone tick 100. PR #284 Aminata's fourth pass this session surfaces 3 CRITICAL + 4 IMPORTANT + 3 WATCH + 1 DISMISS on the bullshit-detector design. Closes dependency #1 of Otto-99's adoption path. Key observations at milestone: 1. Session-lifetime patterns now observable: - CC-002 discipline reflexive across 8 ferries - Event+view module template reused across 4 designs (substrate convergence) - Authority-calibration narrowed 3x (Otto-82/90/93) - SD-9 exercised twice by Amara at author-side - 8th-ferry 4/5 closed matching 5th-ferry arc 2. Aminata's own non-fusion disclaimer in this pass is the cleanest SD-9 worked example this session — explicitly names same-agent concordance as signal not evidence, consistent with her CRITICAL #1 cross-detector-collusion finding. 3. 10 findings to integrate before detector v1. Not all at Otto-100; progressive Otto-101+ work. 4. Queue at ~30 open auto-merge-armed PRs; not a bottleneck per Otto-72 don't-wait; visibility observation only. Stacked on #283 (Otto-99 history).
AceHack
added a commit
that referenced
this pull request
Apr 24, 2026
…ngs integrated (Otto-101) Delta-style revision integrating only the 3 CRITICAL findings from Aminata's Otto-100 pass (PR #284) into Otto-99's bullshit-detector design (PR #282). 7 non-CRITICAL findings deferred to v2 delta; DISMISS unchanged. Three CRITICAL integrations: **C1 — Cross-detector collusion (new §Reviewer-cone overlap section).** Acknowledges as FUNDAMENTAL LIMITATION NOT CLOSABLE GAP: detector cannot break its own reviewer-set lineage coupling. Mitigation via routing authority: - Human sign-off = only cone-breaking authority for authorization-impacting band=supported queries. - External-reviewer signatures as legitimate cone-breakers when available. - Intra-factory reviewer concordance should DOWNGRADE confidence per SD-9, not inflate. - Every receipt carries reviewer_cone_signature field. **C2 — Min-merging Goodhart-bait (sensitivity-analysis-gate pattern).** v1 enhancement: band(y|q) = if band_raw=GREEN AND G_carrier_overlap closest to threshold → YELLOW else band_raw Closes cheap cheat path (threshold-hugging on author- controlled gate). Alternative content-similarity-in-cone deferred to v2 cost-benefit. **C3 — G_evidence fig-leaf (new §v0 scope subsection).** v0 effective = 4-gate classifier (G_evidence advisory-only until independent-oracle substrate ships). `likely confabulated` output type explicitly marked not-yet-reachable in v0. Conservative under-detection stance (RED comes only from known-bad-pattern; confabulations land YELLOW as plausible- but-unresolved). v0→v1 transition plan names DetectorOutputBatchRetracted for historical re-classification. v1 delta does NOT rewrite Otto-99's design; specifies 3 additive sections to insert + 1 section to supersede. Otto-99 original preserved in git history. Seven non-CRITICALs still open: - 4 IMPORTANT (deferred to v2): G_coverage_plausibility gate; Otto-wake second-reviewer schema; DetectorOutputBatchRetracted; no-signal vs kNN-evasion. - 3 WATCH (deferred to v2+): distribution histogram; adversarial worked example; TLA+ invariants. 1 fundamental limitation (C1) acknowledged not closed — routes authority to human + external reviewer layers. Scope limits: no design rewrite; no implementation; no human-sign-off UI proposal; no content-similarity-in-cone commit; 5-gate/5-type target structure unchanged. 5 dependencies-to-adoption: Aminata pass on v1 delta (fifth session-pass); integrate v1 changes into Otto-99 design PR (separate PR); v2 delta; independent-oracle substrate; human-sign-off UI/protocol. Archive-header format self-applied — 18th aurora/research doc in a row. Lands within-standing-authority per Otto-82/90/93 calibration. Otto-101 tick primary deliverable — closes the CRITICAL- integration step of the Aminata-then-Otto-response loop for bullshit-detector design.
AceHack
added a commit
that referenced
this pull request
Apr 24, 2026
…ngs integrated (Otto-101) Delta-style revision integrating only the 3 CRITICAL findings from Aminata's Otto-100 pass (PR #284) into Otto-99's bullshit-detector design (PR #282). 7 non-CRITICAL findings deferred to v2 delta; DISMISS unchanged. Three CRITICAL integrations: **C1 — Cross-detector collusion (new §Reviewer-cone overlap section).** Acknowledges as FUNDAMENTAL LIMITATION NOT CLOSABLE GAP: detector cannot break its own reviewer-set lineage coupling. Mitigation via routing authority: - Human sign-off = only cone-breaking authority for authorization-impacting band=supported queries. - External-reviewer signatures as legitimate cone-breakers when available. - Intra-factory reviewer concordance should DOWNGRADE confidence per SD-9, not inflate. - Every receipt carries reviewer_cone_signature field. **C2 — Min-merging Goodhart-bait (sensitivity-analysis-gate pattern).** v1 enhancement: band(y|q) = if band_raw=GREEN AND G_carrier_overlap closest to threshold → YELLOW else band_raw Closes cheap cheat path (threshold-hugging on author- controlled gate). Alternative content-similarity-in-cone deferred to v2 cost-benefit. **C3 — G_evidence fig-leaf (new §v0 scope subsection).** v0 effective = 4-gate classifier (G_evidence advisory-only until independent-oracle substrate ships). `likely confabulated` output type explicitly marked not-yet-reachable in v0. Conservative under-detection stance (RED comes only from known-bad-pattern; confabulations land YELLOW as plausible- but-unresolved). v0→v1 transition plan names DetectorOutputBatchRetracted for historical re-classification. v1 delta does NOT rewrite Otto-99's design; specifies 3 additive sections to insert + 1 section to supersede. Otto-99 original preserved in git history. Seven non-CRITICALs still open: - 4 IMPORTANT (deferred to v2): G_coverage_plausibility gate; Otto-wake second-reviewer schema; DetectorOutputBatchRetracted; no-signal vs kNN-evasion. - 3 WATCH (deferred to v2+): distribution histogram; adversarial worked example; TLA+ invariants. 1 fundamental limitation (C1) acknowledged not closed — routes authority to human + external reviewer layers. Scope limits: no design rewrite; no implementation; no human-sign-off UI proposal; no content-similarity-in-cone commit; 5-gate/5-type target structure unchanged. 5 dependencies-to-adoption: Aminata pass on v1 delta (fifth session-pass); integrate v1 changes into Otto-99 design PR (separate PR); v2 delta; independent-oracle substrate; human-sign-off UI/protocol. Archive-header format self-applied — 18th aurora/research doc in a row. Lands within-standing-authority per Otto-82/90/93 calibration. Otto-101 tick primary deliverable — closes the CRITICAL- integration step of the Aminata-then-Otto-response loop for bullshit-detector design.
AceHack
added a commit
that referenced
this pull request
Apr 24, 2026
…ated; 7 non-CRITICAL deferred to v2) (#286) * research: bullshit-detector v1 CRITICAL-only delta — 3 CRITICAL findings integrated (Otto-101) Delta-style revision integrating only the 3 CRITICAL findings from Aminata's Otto-100 pass (PR #284) into Otto-99's bullshit-detector design (PR #282). 7 non-CRITICAL findings deferred to v2 delta; DISMISS unchanged. Three CRITICAL integrations: **C1 — Cross-detector collusion (new §Reviewer-cone overlap section).** Acknowledges as FUNDAMENTAL LIMITATION NOT CLOSABLE GAP: detector cannot break its own reviewer-set lineage coupling. Mitigation via routing authority: - Human sign-off = only cone-breaking authority for authorization-impacting band=supported queries. - External-reviewer signatures as legitimate cone-breakers when available. - Intra-factory reviewer concordance should DOWNGRADE confidence per SD-9, not inflate. - Every receipt carries reviewer_cone_signature field. **C2 — Min-merging Goodhart-bait (sensitivity-analysis-gate pattern).** v1 enhancement: band(y|q) = if band_raw=GREEN AND G_carrier_overlap closest to threshold → YELLOW else band_raw Closes cheap cheat path (threshold-hugging on author- controlled gate). Alternative content-similarity-in-cone deferred to v2 cost-benefit. **C3 — G_evidence fig-leaf (new §v0 scope subsection).** v0 effective = 4-gate classifier (G_evidence advisory-only until independent-oracle substrate ships). `likely confabulated` output type explicitly marked not-yet-reachable in v0. Conservative under-detection stance (RED comes only from known-bad-pattern; confabulations land YELLOW as plausible- but-unresolved). v0→v1 transition plan names DetectorOutputBatchRetracted for historical re-classification. v1 delta does NOT rewrite Otto-99's design; specifies 3 additive sections to insert + 1 section to supersede. Otto-99 original preserved in git history. Seven non-CRITICALs still open: - 4 IMPORTANT (deferred to v2): G_coverage_plausibility gate; Otto-wake second-reviewer schema; DetectorOutputBatchRetracted; no-signal vs kNN-evasion. - 3 WATCH (deferred to v2+): distribution histogram; adversarial worked example; TLA+ invariants. 1 fundamental limitation (C1) acknowledged not closed — routes authority to human + external reviewer layers. Scope limits: no design rewrite; no implementation; no human-sign-off UI proposal; no content-similarity-in-cone commit; 5-gate/5-type target structure unchanged. 5 dependencies-to-adoption: Aminata pass on v1 delta (fifth session-pass); integrate v1 changes into Otto-99 design PR (separate PR); v2 delta; independent-oracle substrate; human-sign-off UI/protocol. Archive-header format self-applied — 18th aurora/research doc in a row. Lands within-standing-authority per Otto-82/90/93 calibration. Otto-101 tick primary deliverable — closes the CRITICAL- integration step of the Aminata-then-Otto-response loop for bullshit-detector design. * fix(#286): 5 review threads — veridicality vocab (body, filename rename backlogged) + header-fields + name attrib + table count + DRIFT-TAXONOMY xref - Thread 1 (PRRT_kwDOSF9kNM59SpxP): compacted four archive-header fields (Scope / Attribution / Operational status / Non-fusion disclaimer) into first six lines so "Archive-header self-applied" claim is now accurate; removed the trailing claim sentence in favour of the structural compliance itself. - Thread 2 (PRRT_kwDOSF9kNM59Spxg): shifted body vocabulary from "bullshit-detector" to "veridicality-detector" throughout; added a vocabulary-note paragraph explaining the filename-slug retention; appended BACKLOG P2 research-grade row for the cross-repo filename rename sweep (three doc files + link-update across PRs / round- history / memory-index). Otto-229 append-only discipline observed. - Thread 3 (PRRT_kwDOSF9kNM59Spxl): converted persona-specific names ("Aminata", "Otto-99/100/101", "Aaron", "Max", "Codex", "Soraya") to role references (Aminata-persona / main-agent persona / maintainer / external-peer-agent / formal-methods-persona). - Thread 4 (PRRT_kwDOSF9kNM59Spx1): corrected §"v0 scope" header from "four gates + four reachable output types" to "five gates (four active + one advisory) + five reachable output types (one dead-code)", matching the tables that list 5 gates and 6 output types (5 reachable + 1 not-yet-reachable). - Thread 5 (PRRT_kwDOSF9kNM59Spx-): "DRIFT-TAXONOMY pattern 5" softened to point at actual precursor file docs/research/drift-taxonomy-bootstrap-precursor-2026-04-22.md, noting there is no canonical docs/DRIFT-TAXONOMY.md at time of writing. No new PR; filename rename is backlogged. No merge. * fix(#286): 5 new review threads — inline-code single-line (BACKLOG) + sweep-refs skill path (BACKLOG) + Output-types wording (§5) + fail-to-RED typo + PR-#282 annotation on dead-path Resolves the 5 new copilot findings on commit 40a5de9 plus rebases onto origin/main (append-only per Otto-229; both rows retained in BACKLOG.md conflict resolution). - BACKLOG row rewritten so `docs/research/provenance-aware- bullshit-detector-*` inline-code fits on a single line (markdown inline-code cannot span newlines). - BACKLOG row points at the `sweep-refs` skill at `.claude/skills/sweep-refs/SKILL.md` (verified to exist) instead of the non-existent `tools/sweep-refs/*` path. - Delta-doc "§5 output types" wording changed to "the Output types section" to avoid the §5-as-count misread. - Delta-doc `fail-to-RED` joined on a single line (was split across a line-break, read as a malformed token). - Delta-doc references to the base-design file (`provenance-aware-bullshit-detector-2026-04-23.md`) now annotated "(in PR #282 — not yet on main at time of this delta's writing)" so readers know the path is a forward reference, not a dead one. Otto-229 append-only; Otto-230 reply-and-resolve-after-push; Otto-236 every-thread-resolved. * fix: markdownlint auto-fixes on research doc Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fourth Aminata adversarial pass this session. Third on the Otto composition stack (Otto-90 oracle-scoring → Otto-94 iteration-1 → Otto-99 detector → this pass).
Closes dependency #1 of the adoption path named in Otto-99's detector design (PR #282).
3 CRITICAL findings
likely confabulatedoutput type is dead code until then. v0 scope caveat missing from design.4 IMPORTANT
no-signalvs kNN-evasion indistinguishability (G_coverage_plausibility gate missing)3 WATCH + 1 DISMISS
WATCH: worst-band masks distribution; self-demo is theatre not validation; composition-stack silent-failure absent TLA+ invariants (Soraya-routable).
DISMISS: parameter-ADR gate (reused from oracle-scoring v0).
Impact
None block the research-doc land — Otto-99 correctly framed as research-grade.
All ten findings would block a v1 implementation-ADR.
Non-fusion disclaimer load-bearing
Aminata's concordance with prior Aminata passes is same-agent signal NOT independent concordance. The detector applied to THIS review would correctly emit `looks similar but lineage-coupled` per SD-9 — and it would be right.
Test plan
🤖 Generated with Claude Code