wind-alice-equity-research-expert
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a background update script (update-check.mjs) that executes shell commands to manage its lifecycle.
- The script uses child_process.spawnSync to run git, npx, and node for update verification and installation.
- These commands are directed towards official vendor repositories and standard developer tools.
- [REMOTE_CODE_EXECUTION]: The automated update process uses npx skills to download and install newer versions of the skill.
- Update sources include official GitHub (Wind-Information-Co-Ltd/wind-skills) and Gitee (wind_info/wind-skills) repositories.
- This mechanism is a standard feature for maintaining the skill's compatibility with the Wind Alice platform.
- [EXTERNAL_DOWNLOADS]: The skill facilitates the download of financial research reports and data attachments.
- Attachments are fetched from the alice.wind.com.cn domain, which serves as the official artifact storage for the Wind service.
- All downloads are performed using the user's provided Wind API credentials.
- [INDIRECT_PROMPT_INJECTION]: The skill processes responses from the Wind Alice API, which may contain external research data.
- Data is ingested through the platform's SSE (Server-Sent Events) endpoint in scripts/request.js.
- The skill has capabilities to write files to the local .agents/download/ directory and run background maintenance tasks.
- Given the specialized financial nature of the ingested data and the provided usage constraints, the exploitation surface is considered typical for this category of tool.
Audit Metadata