| content |
{"Commits":[{"Sha1":"d5f71d561 {"Commits":[{"Sha1":"d5f71d5616a084b6c2fcb9f541949ca6b31fee61","Message":"docs(ktv): record intended write-chain order deviation for PM-06 (KTV-PM06-V7)\n\nIntent writers now run before the external success audit so the cleaning\nfinal re-verification also precedes it; this deliberately differs from the\nframework write-chain description (writers after outbox/audit). Same\ntransaction, no KTV writers registered; framework docs unchanged.\n\nCo-Authored-By: Claude Opus 5.5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-10-27T17:31:57-07:00"},{"Sha1":"a358d0f99b8bb50f8a15849bb826a083f04ef7cd","Message":"docs(ktv): record PM-06 round-2 verification fixes and fixed-commit gates\n\nFixed afd2d68a6fc758434b55fef4fce844abb770e9a1 (fingerprint 40791385...,\nunchanged during runs) on a clean checkout: targeted 102/102, default\n1004/1004, matrix check exit 0, full business gate exit 2 (pre-existing\n212/98 fully signed 0), isolated PostgreSQL A 21/21 and B 6/6. The new\ntests failed 5/11 (PG) and 2/7 (unit) on the bc3f334 implementation;\nremoving the principal/role comparison turns the no-bump role switch red.\nRegenerated acceptance run/matrix/backlog belong to that run. Raw failing\nTAP files keep their original trailing spaces.\n\nCo-Authored-By: Claude Opus 5.5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-10-27T17:22:03-07:00"},{"Sha1":"afd2d68a6fc758434b55fef4fce844abb770e9a1","Message":"docs(ktv): describe audit-last ordering, review visibility and store build prerequisite\n\nResponds to KTV-PM06-V1..V6 in docs/ees-cleaning-permissions.md: final\nre-verification precedes the external success audit, principal/role are\ncompared, room executor/rework follow cycle visibility, open escalations\nare reviewer material, and the default suite needs the prototype/store\noffline install and build (V6).\n\nCo-Authored-By: Claude Opus 5.5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-10-27T17:20:22-07:00"},{"Sha1":"be2b2b335f7c3d41fe05de20413f9d97cae11a92","Message":"fix(ktv): audit only after cleaning final checks and keep review material visible\n\nResponds to PM-06 verification KTV-PM06-V1..V5.\n\n- V1: intents and the cleaning final re-verification now run before the\n external success audit, which is the last step before commit; a rejected\n command leaves no success audit, an audit failure still rolls back.\n- V5: assertCurrent also compares principalId and role, so a role switch\n without a binding version bump is rejected.\n- V2/V4: room executor/executorRef/rework stay only on rooms whose current\n cycle is visible (the reviewer workbench independence gate works again);\n they are removed elsewhere, so hidden rework evidence no longer leaks.\n- V3: escalation participants use reviews[].actor/lastReview.actor/closure;\n open escalations awaiting review or close are review material for the\n independent reviewer with the command's independence exclusions.\n\nWorking tree: PM06 PG 11/11 (5/11 failed on bc3f334 implementation),\nPG A 21/21, PG B 6/6, targeted 102/102, default 1004/1004.\n\nCo-Authored-By: Claude Opus 5.5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-10-27T17:20:15-07:00"},{"Sha1":"bc3f3341822d13debfaf62e3acdf23b28d74c10d","Message":"docs(ktv): record PM-06 cleaning transaction final rejection evidence\n\nFixed implementation aadf68fbcf7edc8cfaed209cc301943a574f4f00 (source\nfingerprint cdb0f004..., unchanged during runs) on a clean checkout:\ntargeted 100/100 (previous 95 plus 5 read-scope cases), default 1004/1004,\nmatrix check exit 0, full business gate exit 2 (212 features / 98 AC fully\nsigned still 0, pre-existing), isolated PostgreSQL 18/18 and 6/6.\nThe final PM-06 PG test failed 4/8 on the 80a0de1 implementation; removing\nthe pre-commit registration recheck turns two cases red.\n\nRegenerated acceptance run/matrix/backlog belong to that run. SSE/WS\nrealtime fan-out remains uncovered and is recorded as an open decision.\nNo deployment, loading, shared database, migration or real authorization.\n\nCo-Authored-By: Claude Opus 5.5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-10-27T17:00:52-07:00"}],"HeadCommit":{"Sha1":"d5f71d5616a084b6c2fcb9f541949ca6b31fee61","Message":"docs(ktv): record intended write-chain order deviation for PM-06 (KTV-PM06-V7)\n\nIntent writers now run before the external success audit so the cleaning\nfinal re-verification also precedes it; this deliberately differs from the\nframework write-chain description (writers after outbox/audit). Same\ntransaction, no KTV writers registered; framework docs unchanged.\n\nCo-Authored-By: Claude Opus 5.5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-10-27T17:31:57-07:00"},"CompareURL":"luoanwu/haizan-self-service-ktv/compare/80a0de15c571d455cb4081a2686054fbf11e8a3e...d5f71d5616a084b6c2fcb9f541949ca6b31fee61","Len":6}... |