mode-optimize
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE] (SAFE): The skill is an instructional workflow for code optimization and contains no executable code or malicious logic.\n- [DATA_EXFILTRATION] (SAFE): No access to sensitive files (e.g., credentials, SSH keys) or unauthorized network exfiltration patterns were detected.\n- [REMOTE_CODE_EXECUTION] (SAFE): No instructions for downloading and executing remote scripts (e.g., curl to bash) were found.\n- [PROMPT_INJECTION] (SAFE): The content consists of legitimate technical guidelines and does not contain patterns intended to bypass AI safety filters or override system instructions.\n- [INDIRECT_PROMPT_INJECTION] (SAFE): While the skill processes user-provided code for refactoring, its strict adherence to behavioral preservation, baseline comparisons, and mandatory testing serves as a primary defense against embedded instructions in processed data.
Audit Metadata