cold-email
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is purely instructional and does not contain any executable scripts, shell commands, or binary files. All files are static text documentation.
- [NO_CODE]: The skill consists exclusively of Markdown and JSON files containing copywriting frameworks, benchmarks, and evaluation prompts. There is no code implementation that could be exploited.
- [DATA_EXFILTRATION]: No network activity or sensitive data access patterns were detected. The skill only references reading local documentation for context (e.g., product-marketing.md), which is standard for its intended use.
- [PROMPT_INJECTION]: The instructions focus on persona and tone calibration (e.g., 'write like a peer') without attempting to override safety guidelines or bypass system constraints.
- [EXTERNAL_DOWNLOADS]: No external packages, remote scripts, or third-party dependencies are requested or used.
Audit Metadata