The Agent Skills Directory

[SAFE]: The skill consists of instructional content and evaluation data designed to guide an AI agent through disciplined engineering workflows. No executable scripts or configuration files are present in the skill.
[SAFE]: Analysis of all text content across the files shows no evidence of prompt injection, obfuscation, or attempts to exfiltrate data. The instructions actively promote secure practices such as using sandboxed environments for code execution and verifying outputs against objective metrics.
[SAFE]: No external network requests, unauthorized file access, or privilege escalation patterns were identified. References to tools like curl or git are used conceptually within the documentation to illustrate how to build verification loops.

iterative-refinement