AI Codebase Scanning

Analyze repositories for exploitable vulnerabilities and verify findings with a working proof of concept.

OpenHack reads source code, routes, dependencies, authentication, data flows, and business logic before it starts testing for vulnerabilities. This gives the agent the application context needed to reason about issues that pattern-based rules can miss.

Suspected vulnerabilities are tested in a controlled environment. Findings that can be demonstrated include the evidence needed to reproduce the issue, understand its impact, and begin remediation.

How it works

  1. 01

    Build application context

    Map the repository structure, routes, dependencies, roles, sensitive data, and trust boundaries across the application.

  2. 02

    Generate testable hypotheses

    Reason about where access control, authentication, data handling, business logic, or concurrency could fail.

  3. 03

    Validate suspected issues

    Attempt a controlled exploit and retain findings that can be demonstrated with evidence.

  4. 04

    Report and remediate

    Show the affected code path, proof of concept, impact, and remediation. The managed platform can also prepare a fix pull request.

What it covers

Authentication and access control

Test for authentication bypasses, insecure direct object references, missing authorization checks, and role boundary failures.

Business logic

Evaluate multi-step application behavior, state changes, and trust assumptions that cannot be reduced to a single code pattern.

APIs and data flows

Follow input and sensitive data through routes, services, storage, and external integrations to identify exploitable paths.

Concurrency and timing

Look for race conditions and timing-dependent behavior where the application state can be changed in an unintended order.

What you get

  • Validated findings with affected code paths
  • A working proof of concept or reproduction steps
  • Application context and impact for each issue
  • Remediation guidance and managed fix pull requests

Get started

Run the open-source agent locally for a repository scan, or connect the managed platform for repository-wide scans and pull request reviews.