The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches official type definitions and best practices documentation from established developer portals and package registries. These downloads are used to provide the agent with accurate API signatures and configuration schemas.- [COMMAND_EXECUTION]: Employs standard development utilities such as npm, tar, and npx (to run tsc and eslint) for the purpose of verifying code correctness and identifying anti-patterns.- [CREDENTIALS_UNSAFE]: The skill is designed to detect security vulnerabilities, specifically flagging hardcoded secrets or credentials found in source code or wrangler configuration files.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code provided by users for review (ingestion point). While it lacks explicit delimiters or sanitization for this external content, the risk is mitigated by the skill's instructions to prioritize authoritative documentation and follow a strict set of predefined rules (capability inventory includes linters and type checkers).

workers-best-practices