Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read context from external files to influence the tone and content of generated emails, creating a surface for indirect instructions to bypass intended behavior.
- Ingestion points: The skill reads
ABOUT-ME.mdandMY-ICP.mdfrom the project root to inform communication style and target audience details. - Boundary markers: The instructions lack specific delimiters or "ignore embedded instructions" warnings when processing these files.
- Capability inventory: The skill's primary capability is text generation for various email sequences (welcome, trial expiry, re-engagement).
- Sanitization: There is no evidence of sanitization, filtering, or validation of the content retrieved from the ingested files.
Audit Metadata