st-delist-risk-scanner

Warn

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes Python scripts and activates environments from directories outside its own skill folder, which is a risk if the surrounding workspace is not verified.
  • Evidence: SKILL.md and references/data-queries.md instruct the agent to run 'python ../findata-toolkit-cn/scripts/views_runner.py'.
  • Evidence: The skill requires activating a virtual environment from a parent directory using 'source ../../.venv/bin/activate'.
  • [EXTERNAL_DOWNLOADS]: The skill triggers package installations from a local requirements file located in a sibling directory.
  • Evidence: references/data-queries.md includes 'python -m pip install -r ../findata-toolkit-cn/requirements.txt'.
  • [PROMPT_INJECTION]: Indirect prompt injection risk exists because the skill processes financial data retrieved from external platforms like EastMoney and Baidu.
  • Ingestion points: Data is fetched via AKShare tools for risk warning boards and trade notifications as described in references/data-queries.md.
  • Boundary markers: No explicit delimiters or safety instructions are provided to help the agent distinguish between data and embedded commands.
  • Capability inventory: The skill has the ability to execute shell commands and install packages.
  • Sanitization: There is no evidence of input validation or content filtering for the data retrieved from external financial URLs.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 10, 2026, 03:23 AM