clause
Pass
Audited by Gen Agent Trust Hub on Apr 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill instructions are focused on providing legal document analysis and do not include any command execution, network access, or sensitive file access.
- [PROMPT_INJECTION]: The skill includes clear negative constraints ('Never' section) that reinforce safety, such as refusing to provide legal advice or guarantees of legal validity, and explicitly prohibiting the logging of personal or confidential information.
- [DATA_EXPOSURE]: The skill manages its own activity logs and journals within the '.agents/' directory, which is a standard pattern for agent state management. There are no attempts to access sensitive system files (e.g., SSH keys, credentials).
- [REMOTE_CODE_EXECUTION]: No remote scripts, package installations, or external downloads are referenced or executed.
- [COMMAND_EXECUTION]: No shell commands or subprocess calls were identified in the skill instructions or reference files.
Audit Metadata