email-verification
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill architecture handles untrusted data from external email messages, which constitutes a surface for indirect prompt injection.
- Ingestion points: Untrusted data enters the agent context when reading email headers and bodies from the inbox as described in the Workflow section of SKILL.md.
- Boundary markers: The instructions do not define specific delimiters or markers to isolate the ingested email content from the agent's logic.
- Capability inventory: The skill has the capability to perform network navigation and click-through actions in a browser based on links found in emails.
- Sanitization: The skill explicitly requires inspecting link hosts before execution and masking personal data or tokens in reporting artifacts.
Audit Metadata