hubspot-product-event-sync

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is architected to ingest product event data from backend event streams, which represents a vulnerability surface for indirect instructions.
  • Ingestion points: Untrusted event data is ingested via the push method in SKILL.md and the HubSpotEventSyncPipeline class in references/implementation-guide.md.
  • Boundary markers: There are no explicit boundary markers or instructions implemented to prevent the agent from being influenced by malicious content embedded within event property values.
  • Capability inventory: The skill possesses the capability to perform network operations (fetch, requests.post), database queries (psql), and execute shell commands (curl, jq).
  • Sanitization: The implementation includes property type validation (e.g., verifying number or bool types), but it does not perform sanitization of string values to remove potentially malicious instruction content.
  • [COMMAND_EXECUTION]: The provided documentation includes examples of shell commands used for data maintenance and API verification.
  • Evidence: The skill provides patterns for using psql to query a local dead-letter queue and curl to interact with the HubSpot CRM API for verification purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:44 AM
Security Audit — agent-trust-hub — hubspot-product-event-sync