agentleFS
jscott3201jscott3201/yet-more-skills/skills/python-ci-release/SKILL.md

python-ci-release

Skill4 starsChanged 2 months agoApache-2.0View on GitHub

Use explicitly for Python CI, test gates, native release matrices, package artifacts, and release readiness. Preserve existing workflows; do not publish or add CI without authorization.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    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:

    Read existing workflows, release policy, package roots, and the requested scope first.

    line 9
  2. Distinctive:

    Read gate decisions before altering a matrix.

    line 31

≈ “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

Python CI and Release

Read existing workflows, release policy, package roots, and the requested scope first. This skill is a delivery specialist, not permission to add workflows, publish packages, change branch protection, or broaden supported platforms.

Separate everyday feedback from release qualification

Keep PR work focused: existing formatting/lint/type checks, affected behavioral tests and consumers, and a relevant native binding smoke when needed. A Rust core change that affects exposed behavior can require Python tests even if no.py file changed. Avoid path filters that silently exclude Cargo files, shared sources, stubs, or build metadata used by the binding.

Reserve expanded interpreter/platform/feature matrices, sdist reconstruction, free-threading stress, exhaustive compatibility, and long benchmarks for the established staged-release lane. This does not excuse skipping the tests necessary for the changed behavior locally. Report the light gate's actual scope instead of calling it full coverage.

Reuse repository scripts and pinned tool versions. Keep check-mode commands non-mutating: for an existing Ruff setup use ruff check --no-fix and ruff format --check; do not enable unsafe fixes or run an autoformatter across unrelated work. Use the selected type checker, not three competing checkers. Preserve exit codes through shell pipelines and distinguish collection errors, no tests, skipped tests, and expected failures.

Build the product that users install

Use an isolated environment with the intended interpreter and backend. For extension changes, select native architecture, Rust build profile, Python ABI, and maturin features intentionally. No QEMU, emulation, cross-compilation, or universal-binary shortcut is a substitute for native qualification in this workflow. Run on native hosts/containers appropriate to each promised platform; Linux containers do not establish macOS behavior.

Install the explicit built wheel outside the checkout and inspect module origin. Verify runtime dependencies, public imports, native symbols through behavior, stubs/type consumers, entry points, and version metadata. Build from the source distribution in a clean directory before claiming source releases work. Keep debug editable builds out of performance claims.

Coordinate agents: one owner modifies an environment or build directory, fixed-port suites are not run concurrently, and performance runs have no competing builds. Use existing caching where safe; interpreter ABI, feature selection, OS/architecture, and toolchain differences must not reuse incompatible native outputs.

Close with bounded evidence

List checks actually run, interpreter/platform/ABI, artifact tested, failures/skips, and release-only gaps. Dependency advisory scans and license checks should use the repo's established tools, not introduce an unsolicited security-tool stack. Keep credentials out of outputs and untrusted PR execution away from publication secrets.

Read gate decisions before altering a matrix.

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.