agentleFS
ai-agent-assemblyai-agent-assembly/agent-assembly/.claude/skills/release-assurance/SKILL.md

release-assurance

Skill4 starsChanged 3 months agoApache-2.0View on GitHub

Give confidence that a change is safe to merge/release: real local validation evidence, a genuine (not rubber-stamped) review, and a correct CI_FAILED-vs-CI_UNAVAILABLE_EXTERNAL classification when CI didn't run cleanly. Auto-used. Invoke before merging any PR, and before tagging a release. Any Horonom-owned repo's PR or release, as the company-wide floor. - A product with its own stricter release gate (e.g. a security-sensitive product requiring a second human sign-off, or a formal release-train window) — run this skill's checks and the stricter…

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Don't invent a release gate item this skill doesn't ask for merely to look thorough — check exactly what the repo's own CI/test/lint/build …

    line 27
  2. Distinctive:

    If CI ran and is genuinely green: trust it, don't rerun without a specific reason to doubt it.

    line 37
  3. Very specific:

    If CI is red: classify as CI_FAILED (real defect — fix it, don't bypass) per governance/workspace/ci-classification.md.

    line 39
  4. Very specific:

    If CI could not execute at all, for a verified external reason (billing/quota/outage unrelated to this change): classify as …

    line 43

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md
<!-- horonom:generated --><!-- Source: horonomy/.github agents/skills/release-assurance/SKILL.md. Do not hand-edit — rerun `python3 agents/common/project_skills.py`. -->

SKILL.md — release-assurance

Purpose

Give confidence that a change is safe to merge/release: real local validation evidence, a genuine (not rubber-stamped) review, and a correct CI_FAILED-vs-CI_UNAVAILABLE_EXTERNAL classification when CI didn't run cleanly.

Type

Auto-used. Invoke before merging any PR, and before tagging a release.

When to use

Any Horonom-owned repo's PR or release, as the company-wide floor.

When NOT to use

  • A product with its own stricter release gate (e.g. a security-sensitive product requiring a second human sign-off, or a formal release-train window) — run this skill's checks and the stricter product gate; never substitute this skill for a stricter one already in place.
  • Don't invent a release gate item this skill doesn't ask for merely to look thorough — check exactly what the repo's own CI/test/lint/build commands validate, not a hypothetical superset.

Procedure

  1. Identify what CI would normally run for this change (read the actual workflow files, don't guess).
  2. Reproduce every meaningful gate locally: tests, lint, type check, build — using the repo's own documented commands.
  3. If CI ran and is genuinely green: trust it, don't rerun without a specific reason to doubt it.
  4. If CI is red: classify as CI_FAILED (real defect — fix it, don't bypass) per governance/workspace/ci-classification.md. Never reclassify a genuine test/build failure as externally unavailable to route around it.
  5. If CI could not execute at all, for a verified external reason (billing/quota/outage unrelated to this change): classify as CI_UNAVAILABLE_EXTERNAL, record local-equivalent evidence (commands, exit codes, test counts) in the PR/Jira, and proceed only if every other merge condition already holds.
  6. Independent review: read the diff, compare against the ticket's AC, check correctness/design/security, verify tests exist where warranted. An implementation sub-agent never reviews or merges its own PR.
  7. For a change to anything rendered on a public surface, attach real screenshot evidence of the changed view (see public-release-reconcile for the full release-surface reconciliation, which this skill does not replace).

References

  • governance/workspace/ci-classification.md — the classification rule this skill applies.
  • governance/engineering/testing-review.md — testing and review invariants.
  • governance/engineering/git-pr-merge.md — the pre-merge review checklist this skill runs.

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.