prospect-research-to-cadence

Pass

Audited by Gen Agent Trust Hub on Mar 22, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill fetches data from untrusted external sources, including company firmographics and web search results, and interpolates this content into personalized email templates and sequence parameters.
  • Ingestion points: Data enters the context via apollo_organizations_enrich, apollo_mixed_people_api_search, and generic web search tools as defined in the workflow of SKILL.md.
  • Boundary markers: Absent; the prompt instructions do not specify delimiters or instructions to ignore embedded commands within the enriched data points.
  • Capability inventory: The skill possesses the ability to create contacts, enroll users in automated email sequences, and create Gmail drafts via its configured MCP tools.
  • Sanitization: No explicit sanitization or validation of the fetched external data is performed before it is used for drafting or sequence loading.
  • Mitigation: The skill incorporates a mandatory human-review checkpoint (Stage 3: Approve) using the AskUserQuestion tool, which prevents automated execution of potentially malicious instructions embedded in the processed data.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 22, 2026, 06:50 PM