script-writer
Pass
Audited by Gen Agent Trust Hub on Feb 23, 2026
Risk Level: SAFE
Full Analysis
- Prompt Injection (SAFE): No instructions aimed at overriding agent behavior, bypassing safety protocols, or extracting system prompts were found in the scripts or the reference documentation.
- Data Exposure & Exfiltration (SAFE): The skill creates its own hidden directory (~/.claude/) for storing its JSON database; it does not access sensitive system paths such as SSH keys, cloud credentials, or environment files. No network operations were detected.
- Unverifiable Dependencies & Remote Code Execution (SAFE): The skill relies solely on the Python standard library (json, pathlib, etc.) and does not download or execute any remote scripts or untrusted packages.
- Dynamic Execution (SAFE): Neither the JavaScript nor Python components use unsafe functions like eval() or exec(), and no runtime code generation logic is present.
- Persistence Mechanisms (SAFE): The skill maintains local state for user preferences but does not attempt to achieve persistence via shell profiles, cron jobs, or system services.
Audit Metadata