The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill instructions are focused on git best practices and security gating. The skill explicitly instructs the agent to check for and block sensitive information such as API keys and credentials before committing code. The use of git commands is constrained by a requirement for explicit developer approval.- [INDIRECT_PROMPT_INJECTION]: The skill establishes a process for analyzing codebase changes via git commands. While this involves processing untrusted data (code diffs), the risk is mitigated by the skill's mandatory security gates and the requirement for human verification before any action is taken.

git-commits