persistent-memory

This README outlines a local persistent-memory feature that reads/writes markdown files under ~/.persistent-memory and notes the agent can execute commands. The document itself contains no explicit malware or malicious code, but describes behaviors (automatic persistence, command execution, plaintext storage in ~) that, if implemented without safeguards, present moderate security risk: disclosure of secrets, accidental exfiltration via backups/sync, and potential for arbitrary command execution. Treat as a high-priority security design review: add explicit user consent, encryption, access controls, command execution restrictions, and sensitive-data detection before adopting.