00-blog-writing-guardrails
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions in
SKILL.mdfocus exclusively on content quality, accuracy, and accessibility guidelines for blog posts. No prompt injection attempts or bypass instructions were found. - [SAFE]: The Python script
eval/contract_check.pyperforms a local validation of theSKILL.mdfile to ensure required phrases are present. It does not use dynamic code execution, perform network operations, or access sensitive files. - [SAFE]: The evaluation script
eval/run-eval.shexecutes local Python scripts and does not involve any external downloads or piped remote command execution. - [SAFE]: There is no obfuscated content, hidden metadata, or hardcoded credentials detected across any of the files.
Audit Metadata