build-cognitive-source-driven
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a purely defensive and methodological framework designed to prevent AI hallucinations by mandating the verification of technical implementations against authoritative sources.
- [SAFE]: All external references and documentation URLs target well-known and reputable technology providers, including official domains for React, Node.js, Next.js, and Prisma.
- [SAFE]: No instances of prompt injection, obfuscation, or unauthorized data access were found; the skill's instructions focus exclusively on improving code reliability through standard developer best practices.
- [SAFE]: The workflow incorporates a citation phase (CITE) that encourages transparency by requiring the agent to provide URLs for the documentation it uses, which facilitates human review of the agent's logic.
Audit Metadata