The Agent Skills Directory

[SAFE]: The skill implements a structured project management system that prioritizes user authorization and clear documentation layers.
[COMMAND_EXECUTION]: The skill includes a local utility script (scripts/init_codestable.py) for workspace setup. This script uses standard Python libraries (pathlib, argparse) to perform legitimate filesystem operations, such as creating directories and writing template files within the project root. It does not invoke external processes or perform unauthorized modifications.
[DATA_EXPOSURE]: No evidence of sensitive data harvesting or unauthorized network transmission was found. The skill emphasizes local project state management and explicitly instructs the agent to avoid accessing sensitive paths like SSH or AWS credentials outside of specific dynamic context warnings provided in the general developer guidelines.
[REMOTE_CODE_EXECUTION]: The skill does not download or execute code from external or untrusted sources. All logic is contained within the provided skill files.
[PROMPT_INJECTION]: The instructions do not contain safety bypass patterns or attempts to override system guidelines. It incorporates clear boundaries and requires explicit user confirmation for significant operations like pushing code, deleting files, or initializing workspaces.

cs