#064Technical3 min
Identify exposed endpoints
Distinguish a reachable address from a response that reveals something private.
Hackaiz team · updated 21/11/2026

01Exposure is about the response and its audience
An endpoint is an address that serves a page, file or API response. Reachability is not automatically a vulnerability: public search and product pages must answer visitors. The concern is an address that returns information or capabilities intended for a restricted audience. A database dump, environment file or internal metrics response is a more meaningful signal than an HTTP success code alone.
Hackaiz includes checks for known sensitive file paths, infrastructure exposure and certain API responses. Use these as focused observations about the deployed service. They are not an inventory of every endpoint your application could generate, and a route that was not discovered should not be described as having passed an access-control test.
02Verify the host you want to inspect
Prepare the website or API on /connect. Domain ownership proof is required before any analysis, including checks of publicly reachable addresses. Use the DNS TXT record _hackaiz, the HTTPS file /.well-known/hackaiz-verification.txt, or Hackaiz's instruction to the project's AI to create that file. Follow the value and hostname shown by the verification panel rather than reusing a value from another target.
Identify whether the sensitive service belongs to the same deployed host. An internal dashboard on a different domain is not covered merely because you checked the home page of your application. Select the authorized checks you need and review the scope before launch. Each selected analysis uses credits; an unticked check is not executed or billed.
03Look for content that confirms the exposure
The sensitive-file checks use response signatures, not just filenames. Examples include an environment-style assignment in /.env.production, a ZIP signature in /backup.zip, or SQL statements in /database.sql. These distinctions matter because a single-page application can return its home page for almost any path, even one whose name looks like a backup. A page with a misleading address is not equivalent to a downloadable archive.
Other exposure checks look for infrastructure responses such as internal metrics or administrative service information. Read the reported address and masked evidence carefully. Do not forward an unredacted response containing credentials to a public ticket. The report is a starting point for restricting the exposed resource and deciding whether secrets or records may already have been disclosed.
04Choose a response appropriate to the resource
For a publicly downloadable backup, removing navigation links is insufficient: the file itself must no longer be served to an unauthorized caller. Move backups to private storage and review the deployment process that put them in the web root. For an exposed configuration file, also determine whether its credentials need rotation. Removing the file does not make a copied password unusable.
For metrics or administrative endpoints, consider whether access should require authentication or stay on a private network. A deliberately public API response needs a documented reason and a review of the fields it returns. Use Danger, To fix and To watch to organize attention, but let the sensitivity of the actual content and the intended audience guide the corrective work.
05Check the deployed fix without widening the claim
Assign the issue to the person responsible for hosting, routing or the API, depending on where access is enforced. A GitHub proposal from Hackaiz is a draft pull request for human review and approval, not an automatically merged deployment. Some exposures require hosting changes as well as code changes; make sure the deployed server, not just the repository, has been corrected.
Repeat the relevant authorized check once the change is live and record which address was retested. This uses credits and verifies a particular correction, not every possible address on the host. Optional monitoring applies only to connected systems. It can be part of a follow-up plan, but do not assume it watches an unrelated dashboard or a newly added subdomain without the appropriate connection.
Frequently asked questions
Take the next step
Was this article helpful?