email-response-simulation
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes untrusted data, specifically user-provided draft emails and external research results from LinkedIn and other public sources.
- Ingestion points: SKILL.md (Draft email text input; Research query outputs for professional identity, public voice, and company context).
- Boundary markers: The skill does not implement explicit boundary markers or instructions to differentiate between the data being analyzed and the analysis logic itself.
- Capability inventory: The skill is primarily focused on text generation and analysis; no dangerous capabilities like arbitrary shell execution, file system writes, or sensitive network operations were detected.
- Sanitization: No content sanitization or validation is mentioned for the ingested data.
Audit Metadata