agentleFS
Sign inStart free
kennguyen887kennguyen887/agent-foundation/skills/rules-shipping/SKILL.md

rules-shipping

Skill1 starsChanged 5 months agoMITView on GitHub

Binding personal rules for finishing a change — load after code changes and before calling work done, writing/choosing tests, opening/updating a PR, releasing, merging, post-merge/deploy checks, or sending any Slack/DM message.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

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

    After any change, before done/PR: callers & dependents.

    line 11
  2. Distinctive:

    Follow the release-safety skill.

    line 15
  3. Distinctive:

    Test only what matters: no junk (glue, constants, library behaviour, mock-only, permutation matrices).

    line 19
  4. Very specific:

    No env var without .env.example + schema.

    line 24
  5. Distinctive:

    Slack: never post unprompted. A channel send needs (1) the send instruction + (2) the target → then send without re-confirming wording.

    line 28

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

rules-shipping

These rules are binding (moved verbatim from ~/.claude/CLAUDE.md). Each heading names its full-text file in ~/.claude/rulebook/: Read that file before acting in its area — the one-liners are the trigger, not the whole rule.

Post-Code Impact Check — 13-post-code-impact-check.md

  • After any change, before done/PR: callers & dependents; behaviour deltas; referenced targets exist; before removing a backfill/cron/"one-off" verify the write path populates the data; backward compat with prod client; state what's unverified. Run gates, fix regressions first.
  • Server gate needing a new client input is half a change: enumerate every caller (incl. deployed bundles), open client PRs in the same session, write enable order (clients → server requires → secret/flag last) into both PR bodies.

Release Safety — 19-release-safety-rules.md

  • Follow the release-safety skill; a backend release is valid only if the current prod app keeps working.

Testing — 28-testing-rules.md

  • New endpoint → e2e happy path; critical flows (auth, payments, money/contract) → happy + error e2e; config-mismatch bugs → a test that would have caught it.
  • Test only what matters: no junk (glue, constants, library behaviour, mock-only, permutation matrices); no bystander tests for code the diff didn't touch; smallest set proving the contract; prefer existing route-harness files.
  • HTTP-layer only (supertest / route harness; outermost boundary for non-HTTP). Unreachable through HTTP → no test, prove by running the real thing. Never add cases to an existing forbidden-class (unit/helper) test file.

Release Readiness & Pre-PR — 29-release-readiness.md, 30-pre-pr-checklist.md

  • Env vars declared + validated; migrations run on clean DB; no NODE_ENV branching; smoke/e2e for primary flow; local migrations applied before schema testing.
  • No env var without .env.example + schema; no schema change without migration; no edits to committed migrations; integrations config-driven.

Post-PR / Post-Merge — 31-post-pr-post-merge-follow-up.md

  • Own the change until verified in the target env. Monitor checks; fix CI on the same branch.
  • Slack: never post unprompted. A channel send needs (1) the send instruction + (2) the target → then send without re-confirming wording. DMs always get confirmed first (draft + recipient + wait for yes). Iterating on a draft is not a send order.
  • Slack style (first time): English, <300 chars hard, line breaks always (one item per line, link at line end — every PR/issue a full URL, never a bare #123; URLs may push the cap), Hello <Name>, … Thank you. for reports to a person (not for PR-review pings), first person with explicit "I", scannable bullets, sparse bold on the ask, one line on what was verified, a line for what you deliberately didn't do + why, approval requests name the irreversible bit, link don't re-explain.
  • After merge, confirm the deploy run went GREEN (gh run list --branch <rc|master>); a red shared pipeline is yours to surface + fix even if it predates you. Verify behaviour in RC/prod; blocked by access → ask for the specific account/record; never store prod secrets; report env checked + evidence + limits.

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.

Check your setup, free