channel-preference-detection
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides conceptual frameworks, scoring logic, and example data structures for improving sales bot interactions. No security issues were identified across the 11 threat categories.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process raw prospect messages to update communication preferences (e.g., "Please email me instead"). While this is the core intended functionality, it represents a standard input surface for LLM-based systems.
- Ingestion points: Prospect messages described in the 'Explicit Signals' section of SKILL.md.
- Boundary markers: None specified in the implementation examples.
- Capability inventory: The skill does not define dangerous capabilities such as shell execution, file writes, or network operations.
- Sanitization: Not explicitly defined in the provided scoring or profile logic.
Audit Metadata