The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing official Cloudflare development tools, such as the Wrangler CLI (npm install -g wrangler), and various platform integration packages (e.g., @cloudflare/next-on-pages, @cloudflare/agents, and framework-specific adapters). These resources originate from a well-known and trusted technology organization.
[PROMPT_INJECTION]: The skill describes execution patterns for Cloudflare Workers that ingest and process external web requests (Request objects). While this constitutes a potential surface for indirect prompt injection—particularly when data is passed to AI models via Workers AI or Agents—the skill explicitly instructs developers to perform input validation and sanitization as a best practice.
Ingestion points: HTTP requests handled by the fetch entry point and Pages Functions within the functions/ directory.
Boundary markers: The skill includes a dedicated 'Security' section that recommends sanitizing all user input.
Capability inventory: The documented patterns include interactions with relational databases (D1), object storage (R2), and AI inference services (Workers AI).
Sanitization: The guide provides code examples demonstrating the use of parameterized queries (.bind()) to mitigate injection risks in database operations.

cloudflare