cold-email-copy-grader
Pass
Audited by Gen Agent Trust Hub on Mar 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates entirely through natural language processing and pattern matching against provided email text. It does not perform any network operations, file system access, or code execution.
- [DATA_EXFILTRATION]: No hardcoded credentials, sensitive file paths, or network exfiltration patterns were detected. The skill explicitly states that it requires no database or API access.
- [PROMPT_INJECTION]: The instructions focus on grading and rewriting email copy based on specific marketing frameworks and do not contain any patterns attempting to override safety guidelines or extract system prompts.
- [REMOTE_CODE_EXECUTION]: There are no external dependencies, package installations, or remote script execution patterns found in the skill files.
Audit Metadata