The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references standard platform initialization tools (e.g., npm create cloudflare, npm create hono) and official deployment tools (e.g., cloudflare/wrangler-action@v3) from well-known and trusted organizations.
[COMMAND_EXECUTION]: Instructions include standard use of the wrangler CLI for project initialization, local development simulation, secret management, and production deployment.
[SAFE]: The skill demonstrates high security awareness by explicitly instructing the agent to follow best practices:
SQL Injection Prevention: Advocates for parameterized queries using prepare().bind() for D1 database operations.
Credential Safety: Recommends using wrangler secret put instead of hardcoding secrets in configuration files or environment variables.
Side-Channel Mitigation: Advises using crypto.subtle.timingSafeEqual() for comparing secrets to prevent timing attacks.
Data Leak Prevention: Warns against using module-level mutable state which can leak sensitive data across requests due to V8 isolate reuse.
Memory Safety: Enforces streaming for large payloads to avoid exceeding the 128 MB memory limit of the Workers runtime.

infra-platform-cloudflare-workers