ashare-pre-st-filter

Pass

Audited by Gen Agent Trust Hub on Nov 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local Python script scripts/fetch_sina_penalties.py using subprocess.run to retrieve regulatory penalty data. The script includes input validation for the stock code (6-digit check) and uses the subprocess list-argument format, which mitigates shell injection risks.
  • [EXTERNAL_DOWNLOADS]: The skill fetches public regulatory data from Sina Finance (vip.stock.finance.sina.com.cn). The scraper script implements an ALLOWED_HOST whitelist to prevent Server-Side Request Forgery (SSRF) and follows throttled retry logic.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes financial reports and regulatory penalty text from third-party sources. It mitigates risk by applying strict quantitative logic, keyword-based normalization, and a specific reporting schema, rather than allowing raw external text to directly control agent behavior.
  • [SAFE]: Data dependencies on Tushare and Akshare are standard for Chinese financial analysis. No hardcoded credentials, obfuscation, or persistence mechanisms were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 21, 2026, 05:36 PM