agentleFS
CaduS0uzaCaduS0uza/qa-supreme/skills/release-gate/SKILL.md

release-gate

Skill1 starsChanged 2 months agoMITView on GitHub

Make the ship / no-ship call with an evidenced checklist — every row backed by command output, every exception named and owned. Use before a release, a deploy, or when asked "can we ship this".

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every row is PASS, FAIL, or WAIVED.

    line 12
  2. Distinctive:

    Never report a gate you did not run.

    line 55

“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Release Gate

The gate exists so that shipping is a decision with evidence, not a feeling with a deadline.

The gate table

Every row is PASS, FAIL, or WAIVED. No row may be filled from memory — each carries the evidence that produced it, from this build, not the last one.

#GateEvidence requiredBlocking
1Smoke green on the release buildrunner output + build id servedyes
2T1 suite greenfull output, 0 failed, build idyes
3Every P0 risk has a passing, mutation-proven testrisk register + proof linesyes
4No P0 false positives in changed testsfalse-positive-hunter scanyes
5No uncaught exceptions / 5xx in the exploration runcrawl + console stagesyes
6Authorization matrix greensecurity-regression outputyes
7No critical/serious a11y violationsaxe results per stateyes
8Perf budgets metmetric vs budget per key pageno (debt)
9Flake rate < 1% over the last 20 runsCI flake reportno (debt)
10Rollback tested and documentedthe actual command, and who runs ityes
11Migrations reversible or forward-fix plannedmigration + down path or planyes
12Monitoring/alerts cover the new pathsalert names and thresholdsno (debt)

Waivers

A waiver is not a skip. It requires: who waived it, why, the blast radius if it is wrong, and when it gets fixed. Write it into the report. An unowned waiver is a FAIL.

The verdict

RELEASE GATE  build 8b1d07e -> production
  1 smoke                     PASS   6/6, build 8b1d07e served
  2 T1 suite                  PASS   248 passed, 0 failed, 6m12s
  3 P0 risks covered          PASS   7/7, all mutation-proven
  4 false positives           PASS   0 P0 in 31 changed tests
  5 exploration clean         FAIL   1 uncaught TypeError on /orders/:id (n7)
  6 authz matrix              PASS   42/42
  7 a11y critical/serious     PASS   0
  8 perf budgets              WARN   LCP 2.9s on /products (budget 2.5s)
  10 rollback                 PASS   `vercel rollback 8b1d07e`, owner: on-call
VERDICT: NO-SHIP — gate 5 is blocking. One defect, one owner, one fix.

Rules

  • The gate does not negotiate with a deadline. It reports; humans decide and own the override.
  • A blocking FAIL is one sentence long: what failed, where, who owns it. Not a paragraph.
  • Never report a gate you did not run. "Not run" is a legitimate row value — and it blocks.
  • Re-run the gate after any change to the release build. A gate from a previous build is not evidence about this one.

Automated form

node runner/bin/qa-supreme.mjs run --url https://staging.example.com --out .qa-supreme
# exit code 1 = NO-SHIP. Wire it as the last step of the deploy pipeline.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.