skills/toy-crane/skills/commit/Gen Agent Trust Hub

commit

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to perform local Git operations (staging and committing). These are confined to the local repository context and explicitly exclude remote operations like 'git push'.
  • [DATA_EXPOSURE]: The skill contains a specific safety rule to protect against accidental data exposure by excluding sensitive files like '.env', credentials, and private keys from commits unless explicitly authorized by the user.
  • [PROMPT_INJECTION]: No malicious prompt injection or behavior override patterns were detected. The instructions focus on logical grouping of changes and adherence to Conventional Commits standards.
  • [REMOTE_CODE_EXECUTION]: No remote code execution patterns, external script downloads, or untrusted package installations were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 07:42 AM
Security Audit — agent-trust-hub — commit