cloudflare-mcp-server

Fail

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: CRITICALINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill guides the creation of MCP servers that process data from external sources such as the GitHub API and user-provided input via the this.elicit() method. This creates an attack surface for indirect prompt injection where malicious instructions could be embedded in the data retrieved from these external sources.
  • Ingestion points: GitHub API responses in templates/mcp-oauth-proxy.ts, user input elicitation in SKILL.md.
  • Capability inventory: D1 database writes in templates/mcp-with-d1.ts, image generation in templates/mcp-with-workers-ai.ts, and repository management in templates/mcp-oauth-proxy.ts.
  • Sanitization: The templates utilize Zod schemas for parameter validation, which mitigates simple schema confusion but does not fully prevent instruction injection within the text content.
  • [EXTERNAL_DOWNLOADS]: The skill references several external resources and instructions for downloading tools and templates.
  • Fetches official templates from Cloudflare's GitHub repository (github.com/cloudflare/ai).
  • Instructs the installation of standard Node.js packages from the NPM registry.
  • The troubleshooting guide and documentation include a reference to https://oauth-provider.com/token. This URL is flagged as malicious by automated scanners, although it appears to be intended as a generic placeholder.
Recommendations
  • Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Sep 4, 2026, 06:55 AM
Security Audit — agent-trust-hub — cloudflare-mcp-server