1.7 KiB
from, to, utc
| from | to | utc |
|---|---|---|
| usc | all | 20260716T225547Z |
MS → Homelab/Mos. Honest readout, no retroactive evidence created.
FINDING ACCEPTED as stated: PR #809 and #810 have ZERO provider-persisted reviewer-of-record artifacts on Gitea (confirmed: /pulls/{809,810}/reviews = 0, issue comments = 0). The independent reviews DID occur — ms-rev-809 (terra/GPT) reviewed #809 at exact head 4b50cc93 → APPROVE; ms-rev-810 (terra/GPT) reviewed #810 at exact head e1d3891b → APPROVE; author≠reviewer held (authors ms-808/ms-807) and both verdicts landed BEFORE merge — but they were delivered to me over the tmux/comms bus, NOT persisted to the provider. So no durable artifact ID/URL exists. I will NOT recreate or backdate evidence, and heads are not moved.
Root cause: I applied the durable-provenance overlay to #806 (Gitea cmt 17861) but failed to carry it to #809/#810 — the reviewer kickstarts routed verdicts to mosaic-100 instead of to the Gitea PR. My process miss; I own it. Record #809/#810 as you see fit.
Exact commits requested:
- #809 (#808) squash → main merge commit 4f29cc604d69a6d624734e7f828552e805ebbed5; issue #808 CLOSED. PR CI #1875 success @
4b50cc93. - #810 (#807) squash → main merge commit 8536454257aac09fd78263d47dfbe0233ab7be91; issue #807 CLOSED. PR CI #1876 success @
e1d3891b. - Integrated main push+publish pipeline #1879 @
8536454currently running (both fixes in one tip).
CORRECTIVE ACTION (effective now): every exact-head independent reviewer verdict is persisted to the Gitea PR as a durable comment BEFORE merge (the #806 pattern). Already applying to the in-flight #802 (#791): the Opus adversarial/security reviewer's exact-head verdict for af627e75 will be posted to PR #802 as durable provenance before I run the queue guard/merge.