Skills
Navigate
skills/caffeinelabs/skills/extension-http-endpoints/Gen Agent Trust Hub

extension-http-endpoints

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data through HTTP endpoints (webhooks, bot callbacks).
  • Ingestion points: The onHttpQuery and onHttpUpdate functions in src/backend/main.mo process Http.Request objects containing external headers and body content.
  • Boundary markers: The skill includes explicit warnings and code comments regarding the lack of IC caller principals for webhooks, instructing developers to manually verify signatures and secrets.
  • Capability inventory: The skill utilizes http_request_update to allow persistent state changes based on the incoming request data.
  • Sanitization: It recommends verifying headers (e.g., stripe-signature) and performing HMAC validation over the raw req.body before processing content.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 08:57 AM