complaint-de-escalation
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted customer messages and email threads which could contain instructions aimed at influencing the agent's behavior. * Ingestion points: Processes external customer emails and chat history within the provided message argument (SKILL.md). * Boundary markers: Uses bracketed placeholders for data interpolation but does not include explicit instructions to the model to ignore potential commands embedded in the customer's text. * Capability inventory: The skill is restricted to text generation; it does not request tool access, file system modifications, or network operations. * Sanitization: No input validation or sanitization is performed on the provided message history.
- [NO_CODE]: The skill consists entirely of markdown instructions and prompts without any accompanying scripts, executables, or automated logic.
Audit Metadata