The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could potentially influence the agent's behavior.
Ingestion points: The skill instructions in SKILL.md direct the agent to inspect status, staged and unstaged diffs, untracked files, and recent commit style from the local repository.
Boundary markers: There are no explicit delimiters or specific instructions to disregard instructions that might be embedded within the files, diffs, or commit messages being analyzed.
Capability inventory: The skill authorizes the agent to perform write operations by creating commits (git commit).
Sanitization: No explicit sanitization, validation, or filtering of the ingested repository data is described before it is used to generate commit messages or execute the commit action.

workflow-commit