limit-up-pool-analyzer
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands to activate a virtual environment and run data views, specifically
source ../../.venv/bin/activateandpython ../findata-toolkit-cn/scripts/views_runner.pyas seen inreferences/data-queries.md. - [EXTERNAL_DOWNLOADS]: The agent is instructed to install packages from a local requirements file via
python -m pip install -r ../findata-toolkit-cn/requirements.txt, which results in the acquisition of dependencies from the Python Package Index (PyPI). - [DYNAMIC_EXECUTION]: The skill uses a local script (
views_runner.py) that accepts dynamic inputs via command-line arguments (e.g.,--set symbol=000001), creating a pattern where data or user inputs are passed directly to a shell-executed Python process. - [INDIRECT_PROMPT_INJECTION]: The skill processes financial market data and user-provided stock lists to generate analysis reports.
- Ingestion points: Market data from external providers and user-specified stock pools defined in
SKILL.md. - Boundary markers: The instructions do not define explicit delimiters to isolate external data from the analytical prompts.
- Capability inventory: The skill environment provides access to shell execution and local script running capabilities.
- Sanitization: There is no evidence of input validation or sanitization for external data before it is incorporated into the prompt context.
Audit Metadata