memory
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill facilitates reading historical context from local memory files (
errors.md,context.md,files.md) stored within the user's home directory (~/.claude/projects/). Access is restricted to these specific files via a hardcoded list inmemory.py, preventing arbitrary file reads. - [SAFE]: No external network activity, data exfiltration, or remote code downloads are present. The script uses standard Python libraries and does not define any external dependencies.
- [SAFE]: The script implements a safe path construction logic using environment variables (
MAX_PROJECT_ID,MAX_SESSION_ID) and hardcoded filenames, ensuring operations remain within the intended session memory scope. - [SAFE]: Content processing includes visual boundary markers in the output (e.g., `📁 filename
- description`) which helps the agent distinguish between skill output and historical content, mitigating potential confusion from re-processed instructions.
Audit Metadata