The fifth cut and the move of next: release/fixtures/cut.v0-1-4.json
This records the v0.1.4 cut as release/fixtures/cut.v0-1-4.json, in the same form as #710's cut.v0-1-3.json. The Factory Builder asked for it on 2026-11-10. It opens labelled factory:hands-off.
What GitHub answered
- Release run 42868490093 (
release.yml, success) cut factory/v0.1.4 at c1edc4a4, the merge of "Each step's token holds every read its code can reach: apply asks for the Pause, and a sweep walks every step (#688's third dry run)" (#714). It movedfactory/v0-nextthere, and recorded Deployment 7689721283, whose status is success. The fixture holds the run, its one job, and the job's log at the step. - Eval run 42865325326 (
eval.yml,workflow_dispatch, success) ran on c1edc4a4. It ended at 17:45:21Z, before the cut was asked at 17:47:14Z, and the eval workflow lists it as its only run on that commit. - Read after the cut (17:49Z):
The reads went through the Maintainer's login and were trimmed to cut.v0-1-3.json's fields and headers, read for read. tools/redact_trial.py redacted them outside the repo, and only the clean output was read. It took out the names of steps[1] and steps[6], the token's mint and its post step, as for v0.1.3. Of the log's inputs, the token's line, which GitHub masked, and the exchanges' folder are left out, since tools/check_release.py keeps the four inputs a cut is asked by. removed says what was taken.
Checks it adds
tests/test_release.py::test_the_next_cut_moved_next_onreads each cut after the first; v0.1.4 joins its list. For v0.1.4 it holds:- the cut's writes;
- the eval run (by id, workflow, event, conclusion and commit), ended before the cut, and the one run on its commit;
- both tags at the cut commit, and v0.1.3 where it was;
- all five Deployments in order, newest first, and both statuses.
- Nine mutants of the new fixture, each killed by that test, as for v0.1.3: another eval workflow, run or commit; an eval that ended after the cut; another run listed on the commit; v0.1.3 moved; the newest Deployment not v0.1.4's; its status not success; v0.1.3's Deployment still reading only success.
release/AGENTS.md's cut row already names the test, and is unchanged.
No record is needed: no rule changes.
Checks
| Check | Result |
|---|---|
make checks | OK |
tests.test_release | OK, 28 tests |
| Mutants of the cuts' fixtures, run alone on an export | 45 of 45 killed |
| gitleaks on the fixture | No leaks |
Full make verify | Not run locally, by the Maintainer's standing change of 2026-11-09; CI's verify is the answer |
| Reviewer | Head | Unread parts | Blocking | Major | Minor | Info |
|---|---|---|---|---|---|---|
| agy (Argon) | 24269a5 | none | 0 | 0 | 0 | 0 |
| muse | 24269a5 | none | 0 | 0 | 0 | 0 |
| astra | 24269a5 | none | 0 | 0 | 0 | 0 |
| claude-review | 24269a5 | none | 0 | 0 | 0 | 0 |
| claude-security | 24269a5 | none | 0 | 0 | 0 | 1 |
| Finding | What | Disposition |
|---|---|---|
| claude-security@24269a5 #1 (info) | Three hand mutants of the shared helper a_later_cut_moved_next_on survive: the tag's object.type, the eval's head_branch and the eval's status | not fixed here: the same holds for the cuts v0.1.1 to v0.1.3, and it is the helper's, not this fixture's |
š¤ Generated with Claude Code