workers-best-practices
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides comprehensive guidelines for secure Cloudflare Workers development, such as using cryptographically secure random number generators (crypto.randomUUID) and constant-time comparisons for sensitive tokens.
- [CREDENTIALS_UNSAFE]: The instructions explicitly warn against hardcoding secrets in source code or configuration files, advising the use of platform-specific secret management tools like Wrangler secrets to prevent credential leakage through version control.
- [INDIRECT_PROMPT_INJECTION]: The skill reviews user-provided Workers code and configuration, which constitutes a data ingestion surface. It mitigates potential risks by directing the agent to verify technical claims against official Cloudflare documentation.
- Ingestion points: User-provided source code and Wrangler configuration files (SKILL.md).
- Boundary markers: None explicitly defined for user code analysis.
- Capability inventory: The skill performs code analysis and retrieves documentation from official Cloudflare developer domains (references/configuration.md).
- Sanitization: Directs the agent to prioritize authoritative Cloudflare documentation over pre-training or analyzed code context when verifying API and runtime behavior.
Audit Metadata