The Agent Skills Directory

[SAFE]: The analysis logic is entirely contained within local scripts (codebase_analyzer.py and prd_scaffolder.py) that utilize only the Python standard library, eliminating risks related to untrusted external dependencies.
[SAFE]: The codebase analyzer implements a strict directory ignore list (e.g., .git, node_modules, venv) and an extension-based file whitelist to prevent the accidental ingestion of sensitive metadata, binary files, or environment secrets.
[SAFE]: No network exfiltration or remote code execution patterns were detected. The scripts operate locally and output data to the user's terminal or a specified local file.
[SAFE]: The instructions provided to the agent focus on objective technical extraction (routes, models, enums) and qualitative business analysis without attempting to bypass safety filters or execute arbitrary shell commands.

code-to-prd