ai-persistence/build-cloudflare-adapter
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructs the agent on implementing application-level state persistence using Cloudflare D1 (SQLite) and Durable Objects. The provided patterns for database interactions and state management follow industry standards and do not introduce inherent vulnerabilities.
- [INDIRECT_PROMPT_INJECTION]: The persistence layer defined in this skill serves as a storage mechanism for untrusted user inputs (chat messages). This establishes a surface for indirect prompt injection.
- Ingestion points: External data enters the system through the request handler, specifically in the
chatParamsFromRequest(request)call within thefetchfunction inSKILL.md. - Boundary markers: The skill does not define specific prompt-level delimiters but emphasizes application-level security boundaries.
- Capability inventory: The skill facilitates D1 database writes (
.run()), Durable Object lock management, and network interaction with LLM providers (e.g.,openaiText). - Sanitization: The skill mitigates risks by explicitly instructing the implementation of server-side authorization: "Authorize thread access at the route — derive the user from the session, never trust a client-supplied id."
Audit Metadata