commit

Pass

Audited by Gen Agent Trust Hub on Feb 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell commands (git, rm, ls, echo) to automate repository maintenance and version control tasks. This includes deleting common temporary patterns like test_*.js or temp_*.py to keep the workspace clean.\n- [DATA_EXFILTRATION]: The skill facilitates data transfer via git push. It implements logic to automatically push feature branches while requiring explicit user confirmation for protected branches (e.g., main, master), which is standard behavior for developer tools.\n- [PROMPT_INJECTION]: The skill processes untrusted content through git diff. This creates a surface for indirect prompt injection where malicious code changes could attempt to manipulate the commit message or cleanup process. However, the skill mitigates this by presenting all planned actions to the user for confirmation before proceeding.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 26, 2026, 06:25 AM