vercel-rate-limits

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides code templates for API handlers and middleware that ingest untrusted data from request headers like x-forwarded-for.
  • Ingestion points: Identified in SKILL.md within the middleware and endpoint handler examples.
  • Boundary markers: None are implemented in the provided snippets.
  • Capability inventory: The provided code is limited to returning HTTP status codes (200, 429) and does not interact with high-risk capabilities like the file system or shell execution.
  • Sanitization: Untrusted header values are used directly as Map keys without validation or filtering.
  • [EXTERNAL_DOWNLOADS]: The skill references and suggests the installation of official Vercel libraries and standard open-source utilities for implementation.
  • Resources: Specifically recommends @vercel/firewall, @vercel/functions, p-queue, and uuid.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:48 AM
Security Audit — agent-trust-hub — vercel-rate-limits