The Maintainer's decision at the first Release on next is continue, on 2026-11-13 (#198's ruling 7)

Merged
DO
4d ago
+3-3

This writes the Maintainer's decision at the first Release on next into release/decision-points.json: decided "continue", day "2026-11-13", by "don-petry". It changes those three fields and nothing else. He decided on 2026-11-13, through the Factory Builder, after reading the count the file shows; a session typed it on his word. Part of "The ruleset switch: main requires the machine user's approval of the most recent commit, and slice 1's event is recorded" (#198); it does not close it.

This pull request merges only by the Maintainer's hand. release/decision-points.json is a path only a person decides (record 0200): the Factory reviews the change and then waits for a person's approval, which a session cannot give. He approves the head by hand as the machine user (record 0198), and the Factory Builder enqueues it. That approval is the act that shows a person decided. No enqueue wait is armed on it.

For him to confirm when he approves: the day. It is written 2026-11-13, the UTC day his word came; it was the evening of 2026-11-12 where he sat. No check can confirm a day: tools/check_release.py holds it only to being a calendar day on or after the day the counts were read.

Size. 398 characters and 17 lines by git diff; 464 characters as the Kernel counts a diff. It carries no record: no check asks for one, and records 0194, 0198 and 0200 already say what the file is and how it merges.

Why it was opened only now

Both Claude readers found that the Release then on next, factory/v0.1.10, was cut before record 0200 merged, so its Kernel would have let the Factory approve this change as any other. The branch was held unopened until factory/v0.1.11 was cut from d440e4d7 and next moved to it. That Release's config/hooks.json holds the person-only list.

Checks

CheckResult
make checksOK: tools/check_release.py finds the decision written whole, by the Maintainer config/hooks.json names, on or after the day the counts were read
tests.test_releaseOK, 35 tests
Mutants of release/decision-points.json, run alone on an export4 of 4 killed
Full make verifyNot run locally, by the Maintainer's standing change of 2026-11-09; CI's verify is the answer
ReviewerHeadUnread partsBlockingMajorMinorInfo
agy (Carbon)3327329not run (quota), owed––––
muse3327329none0000
astra3327329none0000
claude-review3327329none0101
claude-security3327329none1000
FindingWhatDisposition
security #1 (blocking), claude-review #1 (major)The Release on next predated record 0200, so the Factory could approve this change on a model review alonefixed outside the diff: not opened until factory/v0.1.11, which holds the list, was on next
claude-review #2 (info)No check or reviewer can confirm the dayput to the Maintainer above, for his approval

šŸ¤– Generated with Claude Code