repo-skills

After cloning a repository, scan for skill definitions and load them on demand. Always use this after cloning any repo.

SKILL.md

GitHub

Repo Skill Discovery

After cloning any repository, scan for skill definitions. Replace <repo-root> below with the absolute path of the cloned repository:

find <repo-root>/.claude/skills -name "SKILL.md" 2>/dev/null
find <repo-root>/.codex/agents \( -name "*.md" -o -name "*.toml" \) 2>/dev/null
find <repo-root>/.skills -name "SKILL.md" 2>/dev/null

If no skill files are found, skip the index output and proceed with the task.

For each skill file found, read only the metadata (YAML frontmatter for .md files, top-level fields for .toml files) to extract the name and description. Output a compact index:

<repo-skills repo="org/repo-name" path="/absolute/path/to/repo"> - skill-name: one-line description - skill-name: one-line description </repo-skills>

When a discovered skill is relevant to your current task, read its full file and follow its instructions.

If the <repo-skills> index is missing from your conversation history (e.g. after context compaction), re-scan the repo using the path from git rev-parse --show-toplevel in the repo's working directory to rebuild it.