ai-md

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [SAFE]: The skill provides a methodology for converting natural language instructions into a structured-label format to improve LLM compliance and efficiency. The content is purely instructional and contains no malicious logic.
  • [COMMAND_EXECUTION]: The documentation includes bash script snippets for measuring file size and creating backups in the local ~/.claude/ directory. These commands use standard system utilities (wc, cp) and are intended for project maintenance.
  • [REMOTE_CODE_EXECUTION]: No remote execution or external downloads were identified. References to curl in the guide are part of semantic labeling examples (e.g., curl→數據) and do not represent actual network operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 07:38 AM
Security Audit — agent-trust-hub — ai-md