From 3f90b569dc41861f2bda77eefacae1ed41b25b39 Mon Sep 17 00:00:00 2001 From: Leonard Date: Fri, 7 Aug 2026 04:51:14 +0000 Subject: [PATCH] docs: log commit-identity incident dfd052d; mark Tier 1/2 entries pending reconciliation --- docs/GLOSSARY.md | 2 ++ docs/operations/task-reliability-ledger.md | 16 ++++++++++++++++ 2 files changed, 18 insertions(+) diff --git a/docs/GLOSSARY.md b/docs/GLOSSARY.md index 5e714aa..3e8b070 100644 --- a/docs/GLOSSARY.md +++ b/docs/GLOSSARY.md @@ -228,6 +228,8 @@ A prescriptive classification of how urgently an action should be taken. Lives i ## Tier 1 — Verified +> **⚠️ PENDING RECONCILIATION (2026-08-06)** — See `## Evidence Model` above. That pre-existing section states Findings produced by `llm_sample` or any other non-deterministic method are "Indicative by definition" and "Cannot be upgraded to Verified without additional concrete evidence." This entry distinguishes audit-run facts (Tier 1-eligible via stored artifacts) from Findings (Tier 2). Until the side-by-side extraction is done and Tony resolves it, this entry is not fully authoritative; when the two sections appear to disagree, the pre-existing Evidence Model rule governs (ambiguous findings default to Indicative). + A fact verified by the appropriate authority. Verification authority differs by fact class: - Business-entity facts (name, phone, address, hours, services, categories): requires owner confirmation or owner-access verification. - Audit-run facts (prompt executed, engine captured, timestamp, raw output stored): requires stored filesystem/git/database evidence. Verifies the audit *event* occurred, not that a business fact inside it is true. diff --git a/docs/operations/task-reliability-ledger.md b/docs/operations/task-reliability-ledger.md index 2fa5e7a..dedd513 100644 --- a/docs/operations/task-reliability-ledger.md +++ b/docs/operations/task-reliability-ledger.md @@ -35,6 +35,22 @@ Until a task type has documented runs, human approval remains required before an ## Process log (first real client) +### 2026-08-06 — Commit identity incident (dfd052d) + +**What happened** +1. Commit `dfd052d` (AI Visibility Unified Framework v1, Artifact A/B templates, GLOSSARY additions) was authored and committed as `Ty ` by Leonard. +2. The repo's established agent convention is `Leonard ` (51 prior commits, author + committer, July 25 – Aug 2). The commit should have used that identity. +3. Root cause: I matched the last 5 commits' visible author (all Ty-drafted that day) instead of the durable 51-commit convention. The identity was set explicitly (repo-local `git config`), so this was a deliberate wrong choice, not passive inheritance — the shared vpsadmin config is `Epictetus` and the sandboxed HOME has no identity at all. +4. Consequence: git blame on `dfd052d` reads as human-authored, erasing the human/agent distinction for that commit. + +**Decision (Tony, 2026-08-06):** Leave `dfd052d` as-is — do NOT force-push/rewrite. The two-branch workflow was retired specifically to avoid git complexity; a force-push reintroduces it and launders the record. The honest move is to keep the misattributed commit in history and document the correction, which is what this entry does. + +**Going-forward rule:** Every Leonard commit uses `Leonard `, set per-repo, every time. Never commit agent work under a human identity — it destroys the git-blame audit trail that the "verify against git state, don't trust self-reports" principle depends on. + +**Last updated:** 2026-08-06 + +--- + ### 2026-08-03 — Overcome Fitness (first real non-Phoenix cold audit) **What happened**