release-end
Skill0 starsChanged 2 months agoMITView on GitHub
End a release that is in flight, with a reason, closing the change with a code that names a fact. Use when asked to stop, abort, cancel, abandon or back out a release, or to clear a stuck change's markers. It writes the closure and clears every marker; it never merges, deploys, rolls back, or re-adds the person's start.
In this repository8 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- change-activateSkill
- change-requestSkill
- change-settleSkill
- change-watchSkill
- release-scheduleSkill
- release-startSkill
- standard-changeSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 24
Read production's x-build-sha first.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/release-end [PR] "<reason>" [--backed-out --to <sha>]
The person's word to stop. Calls
./change/abort.sh <pr> "<reason>" [--backed-out --to <sha>], then
./change/schedule.sh unschedule <pr> "<reason>" for any future window,
because abort closes only the current one and a future reservation would
survive.
Preconditions, checked before calling
- The PR is OPEN and not MERGED. A merged change is settle's, not abort's.
- It is in flight: it carries at least one of the start label,
release:scheduled,deploy:staging,deploy:production. abort on a bare PR clears nothing and writes a false closure. - A reason is given. abort.sh requires one; a closure with no cause is not a record.
./change/lock.sh statusnames this PR or is free. abort.sh callslock.sh release, and a hand run with no run id releases anyone's lock.- Read production's
x-build-shafirst. If it equals this head, refuse the plain form: ending it now leaves production stranded.
The one thing it writes
The closure record comment, then everything abort.sh does in order:
deployment records to failure, the current window closed with the code,
every marker cleared, the closure code, the lock released, release:ended.
What it refuses, and says
- "no reason given; a closure with no cause is not a record."
- "production is serving
<sha>, this change's head; roll back with targets/node/switch.sh, then end with --backed-out --to<sha>." - "--backed-out asserts an estate action; name the sha."
- "#N carries no marker; there is no release to end."
- "the lock is #M's; ending #N would release it."
- "abandoned and superseded are declared closure codes and abort.sh has no flag for them; refused until the script has them." Never write those labels by hand.
What it never does
Merge; deploy; roll production back (recording a rollback is not performing
one); clear app:* or itil:*; rebook; re-add the start label. Ending does
not requeue: the intent is spent, and saying it again is a second word from
the person (024 F7: a machine fault should not spend the person's word).
How the person knows
The "Change closed: code" comment with build, reason and labels at closure;
the PR left with app:*, the class, the closure code and release:ended;
./change/schedule.sh list showing the window closed with the same code;
./change/lock.sh status free.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

