The probe engine scans markdown link patterns, executes concurrent, SSRF-protected requests using range headers to fetch status codes, and filters responses through a 1-hour in-memory cache to prevent duplicate HTTP calls.

flowchart TD
    A[Draft Markdown] --> B[Extract URLs]
    B --> C[Parallel SSRF Validation]
    C --> D{In-Memory Cache?}
    D -- Hit --> H[Categorise Status]
    D -- Miss --> E[HTTP Probe Request]
    E --> F[Store in Cache]
    F --> H
    H --> I[Live: 2xx/3xx]
    H --> J[Gone: 4xx]
    H --> K[Server Error: 5xx]
    H --> L[Dead: Timeout/DNS]