The Agent Skills Directory

[SAFE]: The skill provides a structured framework for analyzing Python architecture and includes scripts to maintain a pattern database. No malicious patterns, persistence mechanisms, or credential exposures were identified.
[COMMAND_EXECUTION]: Scripts such as pipeline.py and merge_findings.py utilize subprocess.run to orchestrate internal validation tasks and query git metadata. These operations are scoped to the local environment and are used for routine maintenance without accepting untrusted shell input.
[DYNAMIC_EXECUTION]: The skill's documentation (flow.md, CONTEXT.md) describes design patterns involving eval() and exec() for code generation. These are included as architectural examples from mainstream libraries and are not used for runtime execution within the skill's own utility scripts.
[INDIRECT_PROMPT_INJECTION]: The skill processes JSON data produced by subagent explorations. It mitigates potential injection risks by employing strict JSON schema validation (patterns-db.schema.json) and specific extraction logic to ensure data integrity before merging findings into the central database.

python-design