crm-automation

Pass

Audited by Gen Agent Trust Hub on Jul 8, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [DATA_EXFILTRATION]: The skill transmits CRM data to external services such as Clearbit, Slack, and Google Sheets. These actions are aligned with the skill's primary purpose of CRM lead management and reporting.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection by processing untrusted external data (lead details, social profiles) and interpolating it directly into LLM prompts.\n
  • Ingestion points: Lead data is ingested from external webhooks (form submissions) and enriched via LinkedIn/Clearbit profile lookups in SKILL.md.\n
  • Boundary markers: No explicit delimiters or instructions are used in the ai_scoring_model or Deduplicate with OpenAI prompt templates to distinguish between system instructions and untrusted data.\n
  • Capability inventory: The skill utilizes tools for contact creation/updates in HubSpot, Salesforce, and Pipedrive, and performs automated email outreach and Slack notifications (SKILL.md).\n
  • Sanitization: There is no evidence of filtering or sanitization of the external lead data before it is included in model prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 8, 2026, 07:46 PM
Security Audit — agent-trust-hub — crm-automation