wind-find-finance-skill
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the AI to run a local Node.js script (
scripts/check-updates.mjs) to check for updates to the skill catalog or the skill itself. - [EXTERNAL_DOWNLOADS]: The AI is instructed to provide the user with installation commands that fetch and install code from the vendor's repositories on GitHub (
github.com/Wind-Information-Co-Ltd/wind-skills) and Gitee. - [COMMAND_EXECUTION]: The skill recommends global installation and update commands (using the
-gflag) which may affect the user's broader system environment or other AI agents. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes skill descriptions from a local catalog file to generate recommendations.
- Ingestion points: The AI reads
references/skills-catalog.mdto identify relevant tools for the user. - Boundary markers: Absent; the instructions do not specify delimiters or warnings to ignore potential instructions embedded in the catalog data.
- Capability inventory: The skill can execute the update check script and recommends the execution of package installation commands to the user.
- Sanitization: None; the content from the catalog is used directly in the AI's response generation process.
Audit Metadata