The Agent Skills Directory

[SAFE]: The skill provides best-practice architectural patterns and TypeScript code snippets for enhancing AI agent reliability and cost management. No malicious behaviors, data exfiltration, or prompt injection vectors were detected.
[SAFE]: The observability section demonstrates security awareness by explicitly advising to log only previews of input and output data rather than full production content, helping to prevent accidental exposure of sensitive user data.
[SAFE]: All provided code is for local logic implementation (such as retry backoff, circuit breakers, and token buckets) and does not include external downloads or unauthorized command execution.

agent-reliability