a-share-data

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external financial news and company announcements, which introduces a surface for indirect prompt injection attacks where malicious instructions could be embedded in the fetched data.\n
  • Ingestion points: scripts/fetch_danginvest.py (via --news) and scripts/fetch_stock_events.py (via news keyword filtering) ingest text content from external domains.\n
  • Boundary markers: The skill instructions do not specify any delimiters or warnings to ignore instructions found within the processed news data.\n
  • Capability inventory: The skill allows network requests (GET) and local file writing for caching purposes. It does not have access to sensitive system files or administrative shell commands.\n
  • Sanitization: Content from external news APIs is processed without explicit sanitization or filtering for adversarial prompts.\n- [EXTERNAL_DOWNLOADS]: Fetches market data and news from external financial platforms.\n
  • Evidence: Scripts connect to gtimg.cn (Tencent), sina.com.cn (Sina), eastmoney.com (EastMoney), xueqiu.com, and dang-invest.com to retrieve structured financial data and news.\n- [COMMAND_EXECUTION]: The skill facilitates the execution of local Python scripts to process financial data and generate technical indicators.\n
  • Evidence: SKILL.md defines command routes for various scripts within the scripts/ directory using the python3 interpreter.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 03:05 PM
Security Audit — agent-trust-hub — a-share-data