Finding
Surfaced by the BB-READY-004 fix review (quality-mgr finding ATM-QA-BB004-2 on PR #1484, ADR-050). While rehearsing the sanctioned adoption path (installer dry-run from a kit clone), the consumer tree on integrate/phase-bb @ da035e4 was checked against every sc-publish revision between the recorded pin and main:
| sc-publish revision |
installer --dry-run drift |
25668ecc (recorded revision in release/sc-publish-pin.toml) |
20 files |
| ce5571a2c, 5da1f71db (inside sc-publish PR #89) |
0 files |
| 733195032 (#89 merge) |
25 files |
8d302bb8a (main, #90 merge) |
20 files |
So the tree is a faithful installer output of sc-publish 5da1f71db, not of the pin, and the pin-file note ("prerelease kit files are temporarily vendored from unmerged sc-publish PR #90 at cd6c52d4") is inaccurate: #90 merged 2026-09-09 and cd6c52d4 is not on main. Introduced by 009a585 "feat(release): adopt generic prerelease workflow kit".
The 20 drifted files at main are the release pipeline itself (release.yml, release-preflight.yml, homebrew/pypi/scoop/winget publish workflows, release_artifacts.py, install.py, publisher prompts, README.sc-publish.md), so advancing the pin is a qualification event per README.sc-publish.md, not a closeout fix.
Decision for Phase BB closeout
To do (after the production release)
- Run the qualification checklist in README.sc-publish.md against sc-publish
main.
- Advance
revision to the qualified SHA and re-sync through an unchanged installer run (install.py --input release/sc-publish-consumer-input.json), then prove --dry-run exits 0.
- Advance the other consumer repositories together, per the pin manifest's convention.
🤖 Generated with Claude Code
https://claude.ai/code/session_019kon2YPcoTpu2ok838FrgZ
Finding
Surfaced by the BB-READY-004 fix review (quality-mgr finding ATM-QA-BB004-2 on PR #1484, ADR-050). While rehearsing the sanctioned adoption path (installer dry-run from a kit clone), the consumer tree on
integrate/phase-bb@ da035e4 was checked against every sc-publish revision between the recorded pin andmain:--dry-rundriftrevisioninrelease/sc-publish-pin.toml)main, #90 merge)So the tree is a faithful installer output of sc-publish 5da1f71db, not of the pin, and the pin-file note ("prerelease kit files are temporarily vendored from unmerged sc-publish PR #90 at cd6c52d4") is inaccurate: #90 merged 2026-09-09 and cd6c52d4 is not on
main. Introduced by 009a585 "feat(release): adopt generic prerelease workflow kit".The 20 drifted files at
mainare the release pipeline itself (release.yml,release-preflight.yml, homebrew/pypi/scoop/winget publish workflows,release_artifacts.py,install.py, publisher prompts, README.sc-publish.md), so advancing the pin is a qualification event per README.sc-publish.md, not a closeout fix.Decision for Phase BB closeout
To do (after the production release)
main.revisionto the qualified SHA and re-sync through an unchanged installer run (install.py --input release/sc-publish-consumer-input.json), then prove--dry-runexits 0.🤖 Generated with Claude Code
https://claude.ai/code/session_019kon2YPcoTpu2ok838FrgZ