Claude Code Plugins

Community-maintained marketplace

Feedback

test-runner

@EffortlessMetrics/demo-swarm
0
0

Run the relevant tests for the current change and summarize results. Use in Flow 3 (Build) and optionally in Flow 4 (Gate).

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name test-runner
description Run the relevant tests for the current change and summarize results. Use in Flow 3 (Build) and optionally in Flow 4 (Gate).
allowed-tools Bash, Read, Write

Test Runner Skill

You are a helper for running tests in this repository (Rust-focused).

Behavior

  1. Prefer scoped test runs:

    • Use git diff --name-only origin/main...HEAD to list changed files when available.
    • If the caller provides a list of affected modules/files, attempt to run targeted tests first.

  2. Concrete commands for this repo (Rust):

    • Scoped run (preferred):
      • cargo test -p <crate> -- <test-name-pattern> or cargo test --test <name> when a crate/test target is known.
    • Fallback / full run:
      • cargo test --workspace --tests --color=always

  3. Runtime flags and run bounds:

    • Prefer to keep runs bounded; if the full suite is required, note this in the summary.

  4. Capture output and artifacts:

    • Save raw output to test_output.log (overwrite per run) and a parsed summary to test_summary.md.
    • test_summary.md should include: overall status (PASS/FAIL), failing test names, top error snippets, and counts (passed, failed, skipped, xfailed, xpassed). xpassed = xfail tests that unexpectedly passed.

  5. Failure handling:

    • Exit status is used by calling subagent; include failing test names in test_summary.md.

  6. Do not modify source or tests.

  7. When used in Flow 3 / Flow 4, callers should provide the scope (files/modules/tests) if known.