Work with HTTP evidence

Capture a useful baseline, redact secrets and compare requests without losing the context that explains a response.

HTTP evidence is easiest to interpret when one variable changes at a time. A response body, status code, redirect and cache header can tell different parts of the story. Preserve them together.

Never paste a live session token into a report example. Replace secrets consistently so another reader can still tell which request belongs to which test identity.

Before you start

  • The method, origin, path, relevant headers and body for a minimal request.
  • Status, relevant response headers and the smallest response excerpt that supports the observation.
  • The test identity’s role, the object’s owner and whether redirects or caching were involved.

Follow the steps

  1. Collect a baseline request with expected behavior. Record the time, account role and relevant application state.
  2. Replace cookies, bearer tokens, API keys and personal data with clear labels. Preserve relationships between synthetic object IDs and owners.
  3. Change one input or identity and capture the comparison. Distinguish the initial response from any followed redirect.
  4. Ask the model to compare observable fields before proposing explanations. Treat missing evidence as a question to resolve.
  5. If using the product’s HTTP tool, review its documented GET/HEAD, target, gateway and engagement constraints. Do not infer broader scanning from a single request.

Use a precise prompt

Adapt to your evidence
Compare these baseline and non-owner HTTP exchanges. First list differences in status, redirects, headers and protected body fields. Then explain which differences support an access-control finding and which have benign alternatives. Identify one minimal follow-up check. All credentials in the examples are redacted labels.

Verify your work

  • The evidence retains enough context to explain the response without exposing usable credentials.
  • A claimed data disclosure appears in the protected body field, not just the status or length.
  • Redirect, cache and role differences are accounted for before the result becomes a finding.

Bring your own evidence into a conversation and work through the next question.

Open chat