What problem does it solve? Python agent codebases accumulate style drift, type errors, security flaws, and architecture violations when quality checks depend on manual review or expensive LLM passes. This Skill defines a deterministic, token-free quality gate that runs locally via pre-commit and in CI, catching errors at commit time instead of in the pipeline. ## Core Features & Use Cases - Layered tooling stack: Covers formatting/linting (Ruff), static typing (mypy), security (Bandit, pip-audit, gitleaks), architecture enforcement (import-linter), complexity limits (Radon/Xenon), and test coverage (pytest-cov, Hypothesis, mutmut). - Centralized configuration: Provides a complete pyproject.toml template with Ruff rules, mypy overrides, import-linter layer contracts, and commitizen settings for Conventional Commits plus SemVer. - Versioning discipline: Establishes Keep a Changelog, Conventional Commits, and Semantic Versioning as workspace standards, automated through commitizen. - Use Case: When setting up a new Agno agent with a src/ layered layout, apply this Skill to install pre-commit hooks so every commit is automatically checked by Ruff, mypy, Bandit, gitleaks, and import-linter before it lands. ## Quick Start Ask the AI to configure the code quality tooling for your agent by running scripts/setup_code_quality.sh and installing the pre-commit hooks with uv.
