The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard local development tools like pytest, ruff, and black to validate, lint, and format code as part of its core functionality.
[CREDENTIALS_UNSAFE]: Includes a dedicated robustness-and-safety.md reference that explicitly prohibits storing secrets in source code and mandates the use of environment variables.
[DYNAMIC_EXECUTION]: Generates and runs temporary scripts for data quality checks (schema verification and null ratio analysis), which is consistent with the stated data engineering purpose.
[PROMPT_INJECTION]: Uses a structured routing mechanism via an INDEX.csv file to determine which sub-skill should handle a request, reducing the surface area for direct instruction overrides.
[DATA_EXPOSURE]: The skill focus is on code generation within the user's workspace; no unauthorized network operations or exfiltration patterns were identified.

code-builder-colony