YAGA

Repository conventions you can check locally and enforce in CI.

YAGA helps teams turn recurring review feedback into repeatable checks. Catch inconsistent commit structure, missing companion changes, unwanted files, and risky workflow settings before they become another review conversation. Start with one check and adopt more as your project needs them.

What can YAGA help with?

Checks produce actionable findings and reports for automation. You choose the policies and checks; YAGA does not require adopting the entire CLI at once.

Start here

Install YAGA and run your first check

YAGA requires Python 3.12 or newer. Its PyPI distribution is yaga-cli and its command is yaga; the distribution named yaga is an unrelated project. The core CLI is tested on Python 3.12 through 3.15. Optional model dependencies have their own installation requirements.

• Adoption recipes — add checks to an existing workflow.
• Assess YAGA for CI — match your problem to a check and its prerequisites.
• Documentation — policies, examples, and operational guidance.
• Command reference — find the check for your task.

Optional review tools

For feedback beyond deterministic rules, YAGA offers model-backed commit advice and agent review plans and receipts. These are optional. The experimental GitHub review adapter has separate deployment and security requirements; it is not needed for ordinary local or CI checks.

Project status

YAGA is beta. CLI and configuration contracts may change. Workflow security checks cover specific rules, not a complete security audit; see the security model for boundaries and reporting.

Changelog · Contributing

Licensed under BSD-3-Clause.