Agent-ready workflow · server
Blind SSRF Callback Evidence
Confirm one outbound server-side request and map protocol, parser, and redirect behavior without touching internal resources.
Entry contract
Enter with artifacts, leave with evidence.
Use when
- URL field
- webhook
- preview
- import
- remote media
- PDF renderer
Prerequisites
- current scope
- controlled listener
- known URL-consuming feature
MCP retrieval
Agents can search this workflow, retrieve the complete graph, or request one stage through the read-only Vulns.co MCP connector.
Classify the URL consumer
Separate validation, save, preview, fetch, delivery, retry, and rendering actions.
Inputs
- observed feature traffic
Outputs
- consumer state map
- Evidence gate
- One request field and one trigger action are identified.
- Negative control
- A valid benign URL establishes intended behavior.
- Stop condition
- Do not infer a fetch from URL acceptance alone.
Establish a direct callback
Submit one unique HTTPS listener URL and preserve DNS and HTTP events separately.
Inputs
- fresh callback token
Outputs
- raw callback record
- Evidence gate
- The callback token maps to one request and one field.
- Negative control
- An unsubmitted token and inert non-URL input remain quiet.
- Stop condition
- Do not open the token in your own browser.
Test one URL component at a time
Compare safe variations that still point only to the controlled listener.
Inputs
- direct baseline
Outputs
- accepted parser matrix
- Evidence gate
- Direct, userinfo, port, scheme, and fragment tokens are unique and independently attributable.
- Negative control
- The same variation with the URL-sensitive delimiter removed is rejected or quiet.
- Stop condition
- No loopback, link-local, private, metadata, or third-party destinations.
Bound the proven primitive
Report resolution, HTTP request, method, headers, redirects, retries, and triggerability as separate facts.
Inputs
- callback matrix
Outputs
- minimal SSRF evidence packet
- Evidence gate
- The conclusion does not claim internal reachability or response access without proof.
- Negative control
- Browser and security-scanner explanations are checked.
- Stop condition
- Stop after sufficient controlled callback evidence.
Related methods
Continue with the right depth.
Sources