skills/leek/agent-skills/commit/Gen Agent Trust Hub

commit

Pass

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a constrained Git workflow focused on session-specific commits. No malicious patterns, obfuscation, or data exfiltration techniques were found.
  • [COMMAND_EXECUTION]: Utilizes standard git command-line tools for local version control tasks. The instructions emphasize specific file paths rather than bulk operations, reducing the risk of accidental staging.
  • [PROMPT_INJECTION]: No prompt injection or safety bypass patterns were detected. The guardrails explicitly restrict the agent from performing potentially destructive history-altering commands like git reset or git stash that could discard uncommitted work.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 7, 2026, 01:27 PM
Security Audit — agent-trust-hub — commit