attribution
Pass
Audited by Gen Agent Trust Hub on Aug 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill describes an architecture for first-party attribution that ingests data from external sources, including URL decorators and webhooks in references/first-party-tracking.md. This ingestion surface could be targeted for indirect injection or data poisoning. \n- Ingestion points: Webhook endpoints for booking tools and URL parameters (metadata[ph_distinct_id]) used for identity stitching. \n- Boundary markers: The skill provides instructions to reject PII (email-shaped values), cap ID lengths, and verify webhook signatures. \n- Capability inventory: The provided implementation logic uses network requests (fetch) to transmit event data to analytics platforms. \n- Sanitization: The skill recommends strict validation of identifiers, signature verification for external webhooks, and failing closed if identity is ambiguous to prevent data corruption or leakage.
Audit Metadata