The Agent Skills Directory

[SAFE]: The skill uses well-known development tools (git, gh) for its primary functions. Instructions include safety mechanisms to warn users about uncommitted changes and prevent destructive actions.
[INDIRECT_PROMPT_INJECTION]: The skill reads GitHub issue data (titles and labels), which is an external ingestion point. (1) Ingestion points: gh issue view command in SKILL.md. (2) Boundary markers: Not explicitly defined in the prompt text. (3) Capability inventory: Shell command execution via git checkout and git worktree add. (4) Sanitization: Present. The skill mandates converting issue titles into restricted 'slugs' (lowercase, hyphenated, length-limited), which mitigates command injection risks when this data is used in shell commands.

branch