segment-cdp

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill includes a Stripe webhook integration in SKILL.md at app/api/webhooks/stripe/route.ts that consumes untrusted JSON data from incoming HTTP requests. This creates a surface for indirect prompt injection.
  • Ingestion points: req.json() in the Stripe webhook handler.
  • Boundary markers: Absent.
  • Capability inventory: Network operations via Segment SDK and fetch.
  • Sanitization: Absent.
  • [DATA_EXFILTRATION]: The skill demonstrates patterns for transmitting user traits, including email and phone numbers, to external platforms such as HubSpot and Salesforce via the Segment API.
  • [EXTERNAL_DOWNLOADS]: The skill references the use of external packages @segment/analytics-next and @segment/analytics-node for implementing tracking functionality.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 09:20 PM
Security Audit — agent-trust-hub — segment-cdp