enrich-lead

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection Surface: The skill processes external identifiers provided via user arguments and retrieves information from the Apollo MCP service without explicit boundary markers or input validation. If a malicious profile or input contains embedded instructions, it could potentially influence subsequent tool calls or the presentation flow.
  • Ingestion points: User-supplied data in $ARGUMENTS and profile data retrieved via mcp__claude_ai_Apollo_MCP__apollo_mixed_people_api_search in SKILL.md.
  • Boundary markers: Absent; there are no specific instructions or delimiters guiding the agent to disregard potential instructions within the retrieved profile data.
  • Capability inventory: The skill utilizes tools to search people, enrich organizations, and create contacts (mcp__claude_ai_Apollo_MCP__apollo_contacts_create) in SKILL.md.
  • Sanitization: There is no data validation or escaping phase outlined before processing or storing the enriched contact details.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 11:26 PM
Security Audit — agent-trust-hub — enrich-lead