The Agent Skills Directory

[SAFE]: The skill serves as a development guide and does not contain executable malicious code. It explicitly instructs the agent to follow security best practices, such as using Redacted types for secrets, avoiding raw shell command execution in favor of safe argument arrays, and validating untrusted data at boundaries using schemas. The documentation also provides specific guidance on secure observability, ensuring that integration with external services like Sentry includes context scrubbing to prevent accidental data exposure. No evidence of prompt injection, obfuscation, or unauthorized data exfiltration was found.

effect-program-design