uniwind
Pass
Audited by Gen Agent Trust Hub on Feb 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were found in the skill's code or documentation. The behavior is consistent with its intended purpose as a developer tool.\n- [EXTERNAL_DOWNLOADS]: The skill fetches documentation from 'https://docs.uniwind.dev/llms.txt'. This is the official project domain and is used to discover available guides and resources.\n- [REMOTE_CODE_EXECUTION]: The documentation references standard installation procedures using Node.js tools, such as 'npm install' and 'npx uniwind-pro'. These commands are expected for the library's setup and configuration.\n- [PROMPT_INJECTION]: The skill features an ingestion point for external data, representing a surface for indirect prompt injection.\n
- Ingestion points: 'https://docs.uniwind.dev/llms.txt' is used to discover documentation pages (referenced in SKILL.md).\n
- Boundary markers: Absent; instructions for fetching the index do not include specific delimiters or 'ignore' warnings.\n
- Capability inventory: The skill directs the agent to provide code fixes and setup guidance, which includes suggesting shell commands for environment configuration.\n
- Sanitization: Absent; external content is processed as documentation context.
Audit Metadata