The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes internal Python scripts to perform audits and initialize documentation. The shared CLI utility in _skill_cli.py implements a run_command function that uses subprocess.run with shell=False and resolves the full path of executables via shutil.which. This is a secure implementation that prevents shell injection attacks.
[DATA_EXPOSURE]: The skill reads project source code files (Vue, TypeScript, Python) and documentation to perform its audit functions. All file operations are restricted to the project workspace provided by the user via the --project-root argument. Reports are written to a dedicated .heyeddi directory within the project, following standard local development practices.
[INDIRECT_PROMPT_INJECTION]: The check_engineering_plan.py and audit_engineering.py scripts ingest untrusted data from the project codebase and user-supplied plans. The scripts use regex-based pattern matching to identify 'engineering smells' and generate reports. Since the scripts do not execute the content of the analyzed files and the agent primarily uses the resulting JSON/Markdown for reporting, the risk of indirect prompt injection influencing the agent's behavior is low.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were detected. The skill does not perform network operations or download external scripts; all logic is contained within the local Python files provided in the skill folder.

engineering-excellence