Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instructional and content-driven. It guides the AI agent on how to mimic a user's writing style based on provided samples and a voice profile.
- [NO_CODE]: There are no scripts, shell commands, or executable binaries included in the skill. All files consist of markdown text.
- [PROMPT_INJECTION]: The instructions are focused on tone, vocabulary, and structural formatting for email drafting. There are no attempts to bypass safety filters or override the agent's core behavioral constraints.
- [EXTERNAL_DOWNLOADS]: The skill does not reference or download any external resources, packages, or scripts.
- [DATA_EXFILTRATION]: No network operations or data transmission patterns were identified. The skill operates purely within the conversational context to generate text drafts.
- [CREDENTIALS_UNSAFE]: No hardcoded secrets or API keys are present. The skill correctly uses placeholders for user-supplied data in
references/samples.md.
Audit Metadata