newsletter-writer
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or safety bypass attempts were detected in the skill instructions.
- [NO_CODE]: The skill is purely descriptive and informational. It provides templates and persona instructions for the AI agent to follow without invoking any scripts, binaries, or command-line tools.
- [DATA_EXPOSURE]: No sensitive file paths, hardcoded credentials, or network exfiltration patterns were found.
- [PROMPT_INJECTION]: The instructions focus on structured writing and do not include commands to ignore previous safety guidelines or reveal system prompts.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided information to generate reports (Ingestion Point: chat input), it has no underlying capabilities such as file writing, shell execution, or network access (Capability Inventory: none), rendering the indirect injection surface technically inert for system compromise.
Audit Metadata