close

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses official Apideck domains (unify.apideck.com, specs.apideck.com) and the official @apideck/unify Node.js package to interact with the Close service.
  • [DATA_EXPOSURE]: The code examples correctly demonstrate using environment variables (e.g., APIDECK_API_KEY) for API keys and application IDs rather than hardcoding credentials.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines patterns for fetching and processing data from Close CRM records, which creates an ingestion point for untrusted data.
  • Ingestion points: CRM data such as contacts, leads, and opportunities fetched via apideck.crm.contacts.list as documented in SKILL.md.
  • Boundary markers: Absent; there are no specific instructions or delimiters provided to the agent to ignore potentially malicious instructions embedded within CRM record content.
  • Capability inventory: The skill enables network operations via the Apideck SDK and curl to interact with external unified APIs.
  • Sanitization: No evidence of data sanitization, schema validation, or escaping of content retrieved from the CRM is present in the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 07:01 PM
Security Audit — agent-trust-hub — close