The Agent Skills Directory

[SAFE]: The skill adheres to security best practices by instructing users to use platform-native secret management (ctx.secrets and ctx.auth_connections) instead of hardcoding credentials.
[DYNAMIC_EXECUTION]: The provided local-executor.mjs script uses dynamic import() to execute generated JavaScript modules locally. This is a functional requirement for the local testing utility and operates on local files provided by the user.
[INDIRECT_PROMPT_INJECTION]: The skill instructions and examples specifically address data leakage risks by mandating the projection of return values, ensuring that only necessary, sanitized information is returned to the agent's context. This serves as a proactive defense against downstream indirect injection attacks.

code-tools