hubspot-lifecycle-and-lists

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONPERSISTENCEDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill connects to the HubSpot API at api.hubapi.com for CRM operations. HubSpot is a well-known and trusted service for marketing and sales operations.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes contact properties and list memberships from HubSpot, representing a potential indirect prompt injection surface.
  • Ingestion points: Contact data is retrieved via REST API calls in SKILL.md and references/implementation-guide.md.
  • Boundary markers: The implementation handles data using structured JSON, though specific instruction-ignoring delimiters for the agent are not included in the examples.
  • Capability inventory: The skill possesses network access (urllib.request), file system access for event storage (SQLite), and shell command execution (Bash).
  • Sanitization: All API interactions use standard JSON serialization, and webhook processing includes HMAC-SHA256 signature verification.
  • [COMMAND_EXECUTION]: Provides documentation and examples for using curl and jq to inspect CRM data and python3 to execute management scripts.
  • [PERSISTENCE]: The webhook consumer implementation in references/implementation-guide.md uses a local SQLite database to store event IDs and payloads, ensuring delivery durability and enabling reconciliation loops.
  • [DATA_EXFILTRATION]: Includes a cross-portal sync implementation that transfers contact emails between authorized HubSpot instances for agency or multi-portal environments.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:44 AM
Security Audit — agent-trust-hub — hubspot-lifecycle-and-lists