The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess module to run git commands for repository analysis. These calls are secured by using shell=False, enforcing strict timeouts, and limiting execution to a specific whitelist of read-only subcommands.
[EXTERNAL_DOWNLOADS]: The skill coordinates with a companion script (recall.py) located on the local filesystem. This dependency is used to export session data and is resolved through standard agent skill paths.
[DATA_EXPOSURE]: To prevent the accidental leak of sensitive information, the skill includes a specialized privacy module (privacy.py) that scans for and redacts various forms of credentials (API keys, tokens, private keys) from the transcripts before they are processed.
[SAFE]: The implementation follows best practices for security and privacy, such as explicitly blocking symlink traversal during file operations and ensuring all generated artifacts are stored with restricted file permissions (mode 0600).

recap