The Agent Skills Directory

[SAFE]: No malicious behaviors or security risks were identified. The skill is designed as a passive router that delegates tasks to other components without performing high-risk operations itself.
[COMMAND_EXECUTION]: The skill uses a shallow directory listing (ls) to identify the presence of specific project files. This is a restricted, read-only operation used for state discovery and does not involve user-controlled arguments or shell injection risks.
[DATA_EXPOSURE]: The skill reads project metadata files such as pyproject.toml and JOURNAL.md. These reads are targeted and used solely to classify the project's environment and development status for routing purposes.

data-scientist-companion