learn

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses ls and grep to list and search for files within the .claude/skills/ directory for management purposes.\n- [PROMPT_INJECTION]: The skill provides an indirect prompt injection surface (Category 8) by extracting and persisting session-derived information.\n
  • Ingestion points: Processes discoveries, workflows, and session history into a new SKILL.md file.\n
  • Boundary markers: Absent; the template for skill extraction does not use specific delimiters to isolate potentially untrusted session data.\n
  • Capability inventory: File listing, searching, and writing within the project's local directory structure.\n
  • Sanitization: Includes a 'Quality Gate' check that explicitly instructs the agent to verify no sensitive information (credentials or personal data) is included in the output.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 09:17 AM