attribution
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it is designed to ingest and act upon data from external, untrusted sources.
- Ingestion points: The skill instructs the agent to analyze self-reported survey data (SKILL.md) and webhook payloads from third-party tools (references/first-party-tracking.md), which typically contain raw user-generated text.
- Boundary markers: There are no instructions provided to the agent to use delimiters or ignore embedded instructions within these data inputs, increasing the risk of the agent following malicious commands hidden in the data.
- Capability inventory: The skill leverages tools to interact with production CRM systems (HubSpot, Salesforce) and analytics platforms, creating a potential impact path if an injection occurs.
- Sanitization: While the documentation suggests data normalization (e.g., email lowercasing), it does not include patterns for sanitizing or escaping content to mitigate prompt injection.
Audit Metadata