The Agent Skills Directory

[SAFE]: The skill does not contain any malicious code, obfuscated payloads, or unauthorized data exfiltration patterns. It follows standard development practices for git workflows.
[COMMAND_EXECUTION]: The skill instructions involve executing standard shell commands, specifically git diff --staged and git commit, which are necessary for its primary purpose of managing source control commits.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes untrusted data from git diffs.
Ingestion points: Untrusted data enters the agent context through git diff --staged as described in SKILL.md.
Boundary markers: There are no explicit instructions to the agent to use delimiters or ignore instructions embedded within the processed diff content.
Capability inventory: The skill is capable of executing shell commands via git.
Sanitization: The skill does not perform sanitization or validation of the content provided by the git diff command.

commit