pytest-runner_skill

Overview

This skill runs and manages Python test suites using pytest with support for fixtures, markers, coverage, and parallel execution. It focuses on safe, repeatable test execution: environment checks, discovery, execution, result parsing, and failure recovery. Checkpoints capture collections, results, coverage, and diagnostics for reproducibility.

How this skill works

Before running tests the skill verifies a Python virtual environment and that pytest (and optional plugins like pytest-xdist and pytest-cov) are installed. It discovers tests, runs them with user-specified options (markers, coverage, parallel workers), captures output and coverage, and stores structured checkpoints. On failures it performs automated diagnostics and suggests narrow reruns or escalation.

When to use it

• Automating local or CI test runs for Python projects
• Collecting coverage reports and HTML summaries
• Running large suites in parallel with pytest-xdist
• Filtering tests by marker, keyword, or last-failed behavior
• Diagnosing test collection, fixture, or import errors

Best practices

• Always verify or activate a virtual environment before running tests
• Install and confirm required pytest plugins (pytest-cov, pytest-xdist) before parallel or coverage runs
• Prefer targeted runs (markers, -k, --lf) during debugging to reduce noise
• Save test collections and results to checkpoints for reproducibility and postmortem analysis
• Ask the user when multiple test dirs, unclear coverage thresholds, or parallelization choices arise

Example use cases

• Run full test suite with coverage and HTML report: python -m pytest tests/ -v --cov=src --cov-report=html
• Run unit tests only using marker filtering: python -m pytest tests/ -v -m "unit"
• Execute tests in parallel: python -m pytest tests/ -v -n auto (requires pytest-xdist)
• Collect failing tests and retry narrower scope up to 3 attempts, then escalate with diagnostics
• Generate a short traceback run and save output to test_results.txt for failure parsing