The Agent Skills Directory

[SAFE]: The skill provides documentation and architectural guidance for using Alchemy 2 with the Effect framework. It does not contain executable malicious code or instructions to bypass safety guardrails.
[SAFE]: The instructions explicitly include security warnings against hardcoding provider tokens in .env files or source code, recommending the use of secure Alchemy profiles for authentication.
[SAFE]: The documentation describes a secure architectural pattern using sandbox-worker-loader, which loads workers into fresh isolates with globalOutbound: null to prevent unauthorized network access during code execution.
[SAFE]: The workflow emphasizes a human-in-the-loop 'plan-then-deploy' process, encouraging users to review infrastructure changes generated by the agent before execution.

learn-alchemy