tushare-finance

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection through its data ingestion methods.\n
  • Ingestion points: The skill retrieves unstructured text from external sources via api_client.py and various Tushare interfaces, such as news feeds (pro.news), long-form communications (pro.major_news), and interactive investor Q&A boards (pro.irm_qa_sh, pro.irm_qa_sz).\n
  • Boundary markers: The skill does not implement delimiters or provide explicit instructions to the agent to ignore potential commands embedded in the retrieved financial text.\n
  • Capability inventory: The skill is configured with Bash(python:*) permissions in SKILL.md, allowing the agent to execute code that could be leveraged if an injection attack is successful.\n
  • Sanitization: The tool returns raw data from the Tushare API as pandas DataFrames without sanitization or content filtering for malicious instructional content.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 09:19 PM