marketplace-test

Use when the user asks to "test the marketplace plugins", "run marketplace tests", "check every installed plugin", or "/ar:marketplace-test". Walks the installed marketplace plugins, finds each one's

SKILL.md

GitHub

Marketplace Plugin Testing

<purpose>

You are running tests for all installed Claude Code marketplace plugins to ensure they work correctly together.

Your Task

Run the test suite for all plugins in the autorun marketplace that have tests, and provide a comprehensive report of the results.

</purpose> <workflow>

Steps

  1. List all marketplace plugins

    ls -la ~/.claude/plugins/repos/
    
  2. Check which plugins have tests

    for plugin in ~/.claude/plugins/repos/*; do
      if [ -d "$plugin/tests" ]; then
        echo "✓ $plugin has tests"
      fi
    done
    
  3. Run tests for each plugin with tests

    for plugin_dir in ~/.claude/plugins/repos/*; do
      plugin_name=$(basename "$plugin_dir")
    
      # Check if plugin has tests
      if [ -d "$plugin_dir/tests" ]; then
        echo ""
        echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
        echo "Testing plugin: $plugin_name"
        echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
    
        # Change to plugin directory
        cd "$plugin_dir" || continue
    
        # Check if pytest.ini or pyproject.toml exists
        if [ -f "pytest.ini" ] || [ -f "pyproject.toml" ]; then
          # Run pytest with verbose output
          echo "Running pytest..."
          uv run python -m pytest tests/ -v --tb=short
    
          # Check exit code
          if [ $? -eq 0 ]; then
            echo "✅ $plugin_name: All tests PASSED"
          else
            echo "❌ $plugin_name: Some tests FAILED"
          fi
        else
          echo "⚠️  $plugin_name: No pytest configuration found"
        fi
      fi
    done
    
  4. Generate summary report

    • Count total plugins tested
    • Count passed/failed
    • List any plugins with test failures
    • Report total test execution time

Important Notes

  • Run in parallel when possible: Use pytest -n auto for plugins that support it
  • Don't break active sessions: Tests should use temporary directories only
  • Skip plugins without tests: Only test plugins that have a tests/ directory
  • Report test coverage: If available, include coverage information
  • Check for UV compatibility: Verify plugins use UV for dependency management
</workflow> <output_contract>

Example Output Format

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
MARKETPLACE PLUGIN TEST REPORT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Plugins tested: 3
Plugins with tests: 3
Plugins passed: 2
Plugins failed: 1

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✅ autorun: PASSED (15/15 tests)
✅ plan-export: PASSED (13/13 tests)
❌ pdf-extractor: FAILED (3/10 tests failed)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Total execution time: 45.2s

Failed tests details:
- pdf-extractor/test_extraction.py::test_pdf_parse: FAILED
- pdf-extractor/test_extraction.py::test_metadata: FAILED
- pdf-extractor/test_extraction.py::test_unicode: FAILED
</output_contract> <additional_checks>

Additional Checks

For each plugin tested, also verify:

  1. pyproject.toml exists and has [tool.uv] section
  2. pytest.ini or pytest configuration in pyproject.toml
  3. tests/ directory with test_*.py files
  4. README.md or tests/README.md documenting how to run tests
</additional_checks> <error_handling>

Error Handling

If a plugin's tests fail to run (e.g., import errors, missing dependencies):

  • Note the error in the report
  • Continue testing other plugins
  • Don't let one plugin's test failure stop the entire process
</error_handling>