skill-create

You can extract coding patterns from git history by analyzing commit logs and file changes to identify recurring workflows, then generating SKILL.md files that codify discovered practices. This surfaces architectural conventions automatically from repository data.

Repository commit logs allow detection of commit conventions, file co-changes, recurring workflows, architectural patterns, and testing signals. Parsing these commits identifies established team practices hidden within historical code changes.

Generating SKILL.md files from git history requires single-repo projects with substantial commit histories. Minimal commit data limits pattern detection accuracy, preventing reliable extraction of recurring coding conventions and workflows.

You turn recurring git workflows into valid Claude Skill frontmatter by running the skill-create command within your repository. It collects git log data, detects patterns, and outputs formatted SKILL.md files with proper frontmatter.

The best way to automate SKILL.md generation for repository conventions is analyzing local git history to detect file co-changes and commit conventions. This produces valid markdown files documenting team practices without manual writing.