diagnose

Pass

Audited by Gen Agent Trust Hub on Jul 19, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security risks were identified during the analysis of the provided files.
  • [COMMAND_EXECUTION]: The skill implements an internal action router to process user-supplied arguments for songwriting tasks (diagnose, demo, audit, etc.). It does not provide a mechanism for arbitrary shell command execution.
  • [DATA_EXFILTRATION]: The skill mentions writing generated files to specific paths as defined in a configuration file (artifact-persistence.md). This behavior is consistent with the intended purpose of maintaining songwriting drafts and project artifacts.
  • [DATA_EXPOSURE]: The skill references the environment variable ${CLAUDE_PROJECT_DIR} to locate project-level template overrides. This is a standard configuration pattern for development tools and does not constitute unauthorized data access.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 19, 2026, 04:25 PM
Security Audit — agent-trust-hub — diagnose