cloudflare-agents
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides clear instructions on preventing SQL injection by using tagged template literals with the built-in SQLite database support (
this.sql). - [SAFE]: Security best practices are emphasized throughout the documentation, specifically regarding authenticating requests in the Worker entry point before delegating to specific agent instances to prevent unauthorized access.
- [SAFE]: The templates demonstrate proper secret management by instructing users to use environment variables and Cloudflare Worker secrets rather than hardcoding sensitive API keys.
- [INDIRECT_PROMPT_INJECTION]: The skill includes patterns for processing external data via web scraping and retrieval-augmented generation (RAG). While these represent common attack surfaces for indirect prompt injection, the skill is a developer resource for these specific features and provides foundational implementation examples.
Audit Metadata