co-write

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: All file operations and path references in the skill are scoped to the project environment (e.g., CLAUDE_PROJECT_DIR) and local context directories, ensuring no unauthorized access to sensitive system files.
  • [SAFE]: The skill does not perform any network requests, external downloads, or execution of remote scripts.
  • [SAFE]: There are no indicators of prompt injection, obfuscation, or hidden instructions designed to bypass safety protocols. The instructions focus entirely on songwriting methodology and session facilitation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 07:55 PM
Security Audit — agent-trust-hub — co-write