The Agent Skills Directory

[SAFE]: No security issues or malicious patterns detected. The skill follows established patterns for A-share market data analysis and adheres to the vendor's documented architecture.
[COMMAND_EXECUTION]: The script validation/smoke.py uses subprocess.run to execute local scripts within the repository (build_tradability_panel.py, audit_trades.py, validate_audit.py). This is a legitimate use of subprocess for automated end-to-end testing of the command-line interface.
[INDIRECT_PROMPT_INJECTION]: The skill processes external trade and market data (CSV/Parquet). Risk is mitigated by strict type coercion using pandas and numpy, and the output is restricted to structured JSON reports and audit ledgers.
[EXTERNAL_DOWNLOADS]: The fetch_bar_panel.py script is designed to interact with a sibling skill skill-pandadata-api for data acquisition. This is a modular design consistent with the vendor's ecosystem and does not involve unauthorized remote code execution.

a-share-tradability-auditor