A Cloudflare Workers-based MCP server for domain security scanning, providing tools for creating scans, managing domain watches, configuring providers and tunnels, and generating security reports.
EdgeIntel Worker exposes 26 tools with basic naming and descriptions, but critical gaps undermine production readiness. All tools have descriptions (10-100 chars, mostly 50-80), but 18 tools lack complete input schemas, parameters are listed but lack type definitions beyond 'string' and 'integer'. No output schemas are documented. Error handling is absent; no recovery guidance or categorization. Tool composition is reasonable (single responsibility per tool), but parameter descriptions are minimal and do not explain constraints, formats, or dependencies. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. Security: no evidence of secret injection, permission gates, or audit logging in the visible code.
Create or update a domain watch for periodic scanning
Create a new provider configuration
Create a new security scan job for one or more domains
Create a new tunnel configuration
Delete a domain watch
Delete a provider configuration
Delete a tunnel configuration
Export scan results in various formats
Generate an AI-powered security brief for scan results
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields (e.g., jobId, scanRunId) for chaining.
18 tools lack complete input schemas. Parameters are listed with minimal type info (string/integer only); no constraints, enums, ranges, or format specifications. LLMs cannot validate inputs before calling.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
Retrieve scan history for a domain
Retrieve available AI inference capabilities
Retrieve the status and details of a scan job
Retrieve the latest scan results for a domain
Retrieve the provider control plane UI application
Retrieve a provider configuration
Retrieve detailed scan results for a specific scan run
Retrieve the tunnel control plane UI application
Send or retrieve tunnel heartbeat status
Retrieve a tunnel configuration
List domain watches that are due for scanning
List all provider configurations
List all tunnel configurations
Test connection to a provider
Test connection to a tunnel
Update a provider configuration
Update a tunnel configuration
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. deleteDomainWatch, deleteProviderSetting, deleteTunnelRecord are marked DESTRUCTIVE but lack annotations to guide agent caution.
No error handling or recovery guidance. Tools return no documented error cases, status codes, or actionable messages. LLMs cannot self-correct or retry intelligently.
Parameter descriptions are minimal (10-30 chars) and lack constraint details. E.g., 'intervalHours' has no min/max; 'format' in exportScan lists options but not as enum; 'domain' has no validation rules.