b2b-outreach-suite
Pass
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Verified Source Processing: The skill includes specific instructions to verify findings from prospect research and explicitly prohibits the fabrication of facts or numbers. This promotes accuracy and reduces the risk of generating misleading content.
- Human-in-the-loop Design: The instructions clearly state that the agent produces text for the user to review and send, rather than executing automated communications. This maintains user control over all external interactions.
- Indirect Prompt Injection Consideration: The skill processes information from external sources such as web research and CRM data. 1. Ingestion points: Research involves web search and public websites (references/prospect-research.md). 2. Boundary markers: The playbooks define specific block structures for output but do not explicitly delimit external content. 3. Capability inventory: The skill produces text for user review and does not possess active tools for sending messages. 4. Sanitization: The instructions mitigate risks by requiring sourced claims and distinguishing between facts and hypotheses.
Audit Metadata