vercel-reliability-patterns

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides serverless function templates that ingest untrusted external data, which is a common surface for indirect prompt injection.
  • Ingestion points: The api/orders/route.ts template reads data from request.headers.get('idempotency-key') and request.json().
  • Boundary markers: The provided code templates do not include specific delimiters or warnings for the agent to ignore embedded instructions in the ingested data.
  • Capability inventory: The skill's scripts and instructions involve database operations (db.order.create) and network requests (fetch).
  • Sanitization: The template code does not include explicit validation or sanitization for the header or JSON body content before processing.
  • [EXTERNAL_DOWNLOADS]: The skill references resources from the author's official domains and uses well-known Node.js packages.
  • The documentation links to tonsofskills.com, intentsolutions.io, and jeremylongshore.com, which are associated with the skill's author.
  • The code templates use standard, established packages such as opossum, uuid, and @vercel/kv.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:47 AM
Security Audit — agent-trust-hub — vercel-reliability-patterns