SKILL.md Format
Invoking Skills
Skill Learning Workflow
- Task begins — task_tracker begins tracking a task
- Task completes — Agent completes the task successfully
- Evaluation — skill_learner evaluates if the approach is worth saving
- Crystallization — If the task passes worthiness checks, crystallizes it as a new skill
- Reuse — Future tasks search for matching skills
- Feedback loop — Confidence adjusts based on success/failure outcomes
Skill Directory Structure
Every learned skill is version-controlled and inspectable — it’s just a file in your repo, so you can diff, edit, or revert it like any other change.
Tools
Built-in and declarative tools the agent uses to act
Workflows
Chain skills into deterministic, repeatable pipelines
Hooks
Intercept, block, or modify agent behavior at every stage
Plugins
Extend GitAgent with installable tools, skills, and hooks