The Agent Skills Directory

[COMMAND_EXECUTION]: The workflow uses git add -A to stage all changes, including new and untracked files. This behavior can lead to the accidental staging and committing of sensitive files (e.g., .env, configuration files, or local keys) that are not excluded via .gitignore.
[PROMPT_INJECTION]: The skill contains instructions that explicitly override typical user-in-the-loop safety patterns by mandating the agent "Pick the most appropriate candidate yourself. Do not ask the user" and "Commit immediately". This removes the human review step for repository modifications.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes the content of file diffs without isolation.
Ingestion points: Reads staged code changes via git diff --cached and git status --short in SKILL.md.
Boundary markers: None. The skill does not instruct the agent to ignore or delimit instructions found within the diff.
Capability inventory: Executes shell commands including git add and git commit as defined in SKILL.md.
Sanitization: None. Data from the diff is directly used by the LLM to generate commit content.

fast-commit