The Agent Skills Directory

[SAFE]: The skill manages personal memory by reading and writing to the "~/.codex" directory and project-local folders. This activity is restricted to the local filesystem and aligns with the skill's stated purpose of providing persistent context for the agent.- [SAFE]: The Python script "memory_cli.py" implements the memory logic using standard Python library modules. It performs file system operations such as reading, writing, and creating directories, but does not invoke external shell commands, download remote content, or use dynamic execution functions like eval().- [SAFE]: While the skill scans conversation history to identify memory candidates, it uses predefined regex patterns and stores the findings as text data. This processes untrusted data safely without creating an execution path for indirect prompt injection.

personal-memory