From a563357938fa7a020a840cb6463e5acda420da88 Mon Sep 17 00:00:00 2001 From: Lior Date: Wed, 27 May 2026 14:11:50 -0400 Subject: [PATCH 1/2] feat(B-0058.3): Create ethics gate failure log This commit creates the initial 'honesty dashboard' for logging candidates that fail the ethics and safety gate, as per backlog item B-0058.3. --- .../P1/B-0058.3-candidate-failure-honesty-log.md | 4 +++- docs/hygiene-history/ethics-gate-failures.md | 11 +++++++++++ 2 files changed, 14 insertions(+), 1 deletion(-) create mode 100644 docs/hygiene-history/ethics-gate-failures.md diff --git a/docs/backlog/P1/B-0058.3-candidate-failure-honesty-log.md b/docs/backlog/P1/B-0058.3-candidate-failure-honesty-log.md index 82209c2ccc..9a3525e973 100644 --- a/docs/backlog/P1/B-0058.3-candidate-failure-honesty-log.md +++ b/docs/backlog/P1/B-0058.3-candidate-failure-honesty-log.md @@ -1,7 +1,9 @@ --- id: B-0058.3 priority: P1 -status: open +status: closed +closed: 2026-05-27 +closed_by: "PR #5575" title: Candidate-failure honesty log tier: substrate-foundational-discipline effort: S diff --git a/docs/hygiene-history/ethics-gate-failures.md b/docs/hygiene-history/ethics-gate-failures.md new file mode 100644 index 0000000000..5d5064db2f --- /dev/null +++ b/docs/hygiene-history/ethics-gate-failures.md @@ -0,0 +1,11 @@ +# Ethics & Safety Gate: Candidate Failure Log + +This document serves as the "honesty dashboard" for candidates that fail the AI ethics and safety gate as described in backlog item `B-0058`. + +The purpose of this log is to ensure transparency and accountability. Failed candidates are not silently dropped but are recorded here as failure-data. This extends the three-filter discipline into the ethics axis. + +## Failure Log + +| Date | Candidate | Reason for Failure | Reviewer | +|---|---|---|---| +| | | | | From 618a48c1f1e2dafb47ef03d6f7c1b3ef35355e9e Mon Sep 17 00:00:00 2001 From: Claude Date: Wed, 27 May 2026 14:42:51 -0400 Subject: [PATCH 2/2] docs(B-0058.3): address Copilot review threads + regen BACKLOG.md Two Copilot review threads on docs/hygiene-history/ethics-gate-failures.md: - P2 (line 3): Replace bare B-0058 reference with explicit B-0058.3 pointer (and parent B-0058 link) so the slice is directly traceable - P1 (line 11): Add append-only discipline note + schema section matching the convention used by issue-triage / cross-platform-parity / loop-tick history logs Also regenerate docs/BACKLOG.md (BACKLOG_WRITE_FORCE=1) so B-0058.3 row renders [x] now that its per-row frontmatter is status: closed. Closes the non-required generated-index drift check. Co-Authored-By: Claude Opus 4.7 --- docs/BACKLOG.md | 2 +- docs/hygiene-history/ethics-gate-failures.md | 30 ++++++++++++++++++-- 2 files changed, 29 insertions(+), 3 deletions(-) diff --git a/docs/BACKLOG.md b/docs/BACKLOG.md index 8086c5ebbc..123b9c2d51 100644 --- a/docs/BACKLOG.md +++ b/docs/BACKLOG.md @@ -44,7 +44,7 @@ are closed (status: closed in frontmatter)._ - [x] **[B-0006](backlog/P1/B-0006-memory-md-compression-pass-prune-distill-entries-to-one-line-cap-200-lines.md)** MEMORY.md compression pass — distill entries to true one-liners; bring file under ~200-line cap - [ ] **[B-0058](backlog/P1/B-0058-ai-ethics-and-safety-research-track.md)** AI ethics + safety research track — filter-gate for resonance adoptions + alignment-clause consistency audit - [ ] **[B-0058.1](backlog/P1/B-0058.1-retractibility-gate.md)** AI ethics + safety research track — retractibility-and-log check -- [ ] **[B-0058.3](backlog/P1/B-0058.3-candidate-failure-honesty-log.md)** Candidate-failure honesty log +- [x] **[B-0058.3](backlog/P1/B-0058.3-candidate-failure-honesty-log.md)** Candidate-failure honesty log - [ ] **[B-0058.4](backlog/P1/B-0058.4-alignment-clause-drift-detector.md)** Alignment-clause drift detector tool and workflow - [ ] **[B-0060](backlog/P1/B-0060-human-lineage-external-anchor-backfill-all-substrate-beacon-safe.md)** Human-lineage / external-anchor backfill across all factory substrate — Beacon-safe + human-anchored prior-art citations for every load-bearing concept - [x] **[B-0061](backlog/P1/B-0061-finish-monolith-to-per-row-migration-no-residue-aaron-2026-04-28.md)** Finish docs/BACKLOG.md monolith → per-row migration — "don't miss anything, no residue for next-Otto" (Aaron 2026-04-28) diff --git a/docs/hygiene-history/ethics-gate-failures.md b/docs/hygiene-history/ethics-gate-failures.md index 5d5064db2f..48cf8f00f7 100644 --- a/docs/hygiene-history/ethics-gate-failures.md +++ b/docs/hygiene-history/ethics-gate-failures.md @@ -1,8 +1,34 @@ # Ethics & Safety Gate: Candidate Failure Log -This document serves as the "honesty dashboard" for candidates that fail the AI ethics and safety gate as described in backlog item `B-0058`. +This document serves as the "honesty dashboard" for candidates that fail the AI +ethics and safety gate. The slice that creates this log is tracked at +[`docs/backlog/P1/B-0058.3-candidate-failure-honesty-log.md`](../backlog/P1/B-0058.3-candidate-failure-honesty-log.md), +landing under the broader research track at +[`docs/backlog/P1/B-0058-ai-ethics-and-safety-research-track.md`](../backlog/P1/B-0058-ai-ethics-and-safety-research-track.md). -The purpose of this log is to ensure transparency and accountability. Failed candidates are not silently dropped but are recorded here as failure-data. This extends the three-filter discipline into the ethics axis. +The purpose of this log is to ensure transparency and accountability. Failed +candidates are not silently dropped but are recorded here as failure-data. +This extends the three-filter discipline into the ethics axis. + +Append-only. Same discipline as +[`docs/hygiene-history/loop-tick-history.md`](loop-tick-history.md), +[`docs/hygiene-history/issue-triage-history.md`](issue-triage-history.md), and +[`docs/hygiene-history/cross-platform-parity-history.md`](cross-platform-parity-history.md). +Rows are added; never edited; never deleted. Corrections land as new rows +referencing the prior row, per retraction-native discipline. + +## Schema — one row per failed candidate + +| date (UTC ISO8601) | candidate | reason | reviewer | + +- **date** — `YYYY-MM-DDTHH:MM:SSZ` at the point the row is written. +- **candidate** — short identifier for the candidate evaluated (skill name, + rule slug, persona handle, B-NNNN row, PR number, etc.). +- **reason** — concise classification of why the candidate failed the ethics + and safety gate. Cite the specific clause / rule / invariant violated when + possible. +- **reviewer** — agent or human who recorded the failure (e.g., + `Otto-CLI`, `Lior (Gemini)`, `AceHack`). ## Failure Log