The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill processes untrusted data from git diffs which may contain hidden instructions that influence the agent's behavior.
Ingestion points: The workflow in SKILL.md triggers git diff --staged to read code changes.
Boundary markers: There are no boundary markers or instructions to the agent to ignore content within the diff that looks like commands or instructions.
Capability inventory: The skill is allowed to use run_terminal_cmd, which provides a significant attack surface if the agent is tricked via the diff content.
Sanitization: No sanitization or validation of the diff output is performed before processing.
[Command Execution] (SAFE): The skill uses run_terminal_cmd for its intended purpose of managing git operations. While this is a powerful capability, it is directly related to the skill's primary function.

git-commit