agentleFS
joelbarmettlerUZHjoelbarmettlerUZH/ultrainit.sh/.claude/skills/pre-pr-checklist/SKILL.md

pre-pr-checklist

Skill8 starsChanged 7 months agoApache-2.0View on GitHub

This runs bash -n on all .sh files (excluding test-repos/) and jq empty on all schemas/*.json. This is the same syntax job CI runs first. A failure here blocks the test job from running. Requires Docker. Runs unit, scripts, edge, and integration tests. If Docker is not available locally, the minimum is make test-unit && make test-scripts && make test-edge. Important: Integration tests run with || true in CI and show green even when failing. After make test-integration, inspect the…

3 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

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

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 2

    Distinctive

    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:

    If output.json diverged, re-merge it:

    line 54
  2. Distinctive:

    All commits must follow Conventional Commits:

    line 64
  3. Distinctive:

    All of these will silently break on macOS (bash 3.2, BSD sed, stock awk).

    line 95
  4. Very specific:

    If you modified prompts/synthesizer-tooling.md or lib/synthesize.sh, verify postprocess_descriptions() is still stripping angle brackets:

    line 99

“Common” is measured: kept by at least 6 in 10 of the 3 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

Before You Start

  • CLAUDE.md — project conventions (commit format, naming, cross-platform constraints)
  • .github/workflows/test.yml — exactly what CI runs (syntax job + test job)
  • lib/validate.sh — quality gates applied to generated artifacts
  • tests/fixtures/findings/ — fixture files that must stay in sync with schemas

Checklist

1. Shell syntax check

make check

This runs bash -n on all .sh files (excluding test-repos/) and jq empty on all schemas/*.json. This is the same syntax job CI runs first. A failure here blocks the test job from running.

2. Full test suite

make test-all

Requires Docker. Runs unit, scripts, edge, and integration tests. If Docker is not available locally, the minimum is make test-unit && make test-scripts && make test-edge.

Important: Integration tests run with || true in CI and show green even when failing. After make test-integration, inspect the output explicitly — don't trust the exit code alone.

3. Fixture consistency check

# Spot-check: do all 8 fixture files have valid JSON?
for f in tests/fixtures/findings/*.json; do
    jq empty "$f" && echo "OK: $f" || echo "FAIL: $f"
done

# Verify the merged synthesis fixture is up to date
jq -s '.[0] * .[1]' \
    tests/fixtures/synthesis/output-docs.json \
    tests/fixtures/synthesis/output-tooling.json \
    > /tmp/expected-output.json
diff tests/fixtures/synthesis/output.json /tmp/expected-output.json && echo "output.json in sync" || echo "DIVERGED — re-merge"

If output.json diverged, re-merge it:

jq -s '.[0] * .[1]' \
    tests/fixtures/synthesis/output-docs.json \
    tests/fixtures/synthesis/output-tooling.json \
    > tests/fixtures/synthesis/output.json

4. Commit message format

All commits must follow Conventional Commits:

  • feat: — new feature
  • fix: — bug fix
  • test: — test additions/changes
  • docs: — documentation only
  • refactor: — code change without feature/fix
  • chore: — maintenance (CI, deps, Makefile)
git log --oneline -10  # check recent messages follow the pattern

5. No untested code

The project rule from CLAUDE.md (developer answers): "Write untested code" is in the never-do list. For any new function in lib/, there must be a corresponding test in tests/unit/. For changes to scripts/validate-skill.sh or scripts/validate-subagent.sh, there must be updated tests in tests/scripts/.

6. Cross-platform safety check

Scan your changes for forbidden patterns:

# Portable sed check (GNU/BSD incompatible)
grep -n 'sed -i' lib/*.sh scripts/*.sh ultrainit.sh

# gawk-only gensub check
grep -n 'gensub(' lib/*.sh scripts/*.sh

# bash 4+ only features
grep -n 'declare -A\|mapfile\|readarray\|wait -n\|wait -p' lib/*.sh

All of these will silently break on macOS (bash 3.2, BSD sed, stock awk).

7. No angle brackets in generated skill/subagent descriptions

If you modified prompts/synthesizer-tooling.md or lib/synthesize.sh, verify postprocess_descriptions() is still stripping angle brackets:

grep -n 'postprocess_descriptions\|[<>]' lib/synthesize.sh | head -20

Verify

make check && make test-all

Both must complete without error before opening the PR.

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.