SonarQube Hunter Agent

The AI agent that hunts logic flaws

An AI-powered agent that runs a deep analysis of your SonarQube projects, hunting broken access control, business logic, and authentication flaws that traditional analysis was never built to catch.

Issues 3 new
Raised by Hunter Agent

Scanning Scan complete
HIGH
Authorization bypass through user-controlled key
InvoiceController.java:141 CWE-639 Broken access control
HIGH
Session fixation on login
SessionManager.java:88 CWE-384 Auth & sessions
MED
No rate limit on upload endpoint
UploadService.java:52 CWE-770 Business logic
The hard part of verification

Not all vulnerabilities are bugs in your code

Attackers don't need to break your code, they just need to use it the way it wasn't meant to be used.

Privilege escalation hides in plain sight

Nothing in the code looks wrong. A regular user can still end up with admin rights.

Business logic gets abused

Stacked coupons, skipped workflow steps, missing rate limits. None of it looks like a vulnerability to a scanner.

Reviews can't keep pace

AI-generated code ships faster than any human review cycle can cover.

Pentests happen twice a year

Everything shipped in between goes live unchecked until the next engagement.

Complete coverage

SAST and Hunter Agent, one platform.

Algorithmic analysis and agentic reasoning find different classes of vulnerability. Running both, closes the gap.

SonarQube analysis (SAST + SCA)

Known & pattern-based vulnerabilities

  • Injection and unsafe data flows
  • Known CVEs in first-party and open-source dependency code
  • Secrets detection
  • Runs on every commit, consistent and repeatable
SonarQube Hunter Agent

Intent-based vulnerabilities

  • Broken access control and IDOR
  • Business logic flaws
  • Authentication and session gaps
  • Deep, full-codebase reasoning about intent
How it works

It hunts through playbooks, like a security researcher.

A playbook is a curated, multi-step sequence of specialized security prompts that encodes Sonar's application security expertise. The agent runs them across your whole codebase and works a lead until it can prove an issue is real.

InvoiceController.java
139 @GetMapping("/invoices/{id}")
140 public Invoice get(@PathVariable Long id) {
141 return invoiceRepo.findById(id);// no tenant check
142 }

HIGH

Authorization bypass through user-controlled key

InvoiceController.java:141  CWE-639 · Broken access control

Raised by Hunter Agent

The code is valid. The intent is not: any user can read another tenant's invoice.

1. Hunt

The agent explores the full codebase and follows data and identity flows across files, building hypotheses about where intent and implementation diverge, the way a researcher does on a code audit.

2. Confirm

Every suspected vulnerability is investigated and validated before it surfaces. You get high-precision findings, not a queue of false positives to triage.

3. Explain

Each finding arrives with a severity, a plain account of why it matters, and the exact location, so a developer can act in minutes instead of reverse-engineering it.

4. Integrate

Confirmed findings land in your SonarQube issue list, tagged and ready to triage, assign, and fix in the workflow your team already uses.

What it detects

One playbook. Three classes of flaw.

CWE lists are representative, not exhaustive. Works with any language supported by SonarQube Cloud.

Broken access control

IDOR, missing or incorrect authorization checks, privilege escalation, sensitive data exposure, and cross-site request forgery.

CWE-639 CWE-285 CWE-862 CWE-863 CWE-269 CWE-200 CWE-352

Business logic flaws

Skipping required workflow steps, abusing repeatable actions, and missing rate limits.

CWE-840 CWE-841 CWE-799 CWE-837 CWE-770

Authentication & sessions

Broken authentication, session fixation, non-expiring sessions, weak password recovery, missing MFA, and brute-force gaps.

CWE-287 CWE-384 CWE-613 CWE-640 CWE-308 CWE-294 CWE-307
Why it works

Layered security by design.

SAST, SCA, and Hunter Agent win together for the same reason any layered defense does: each catches what the other cannot. Sonar combines them across the Verify stage, so no single check has to carry the whole codebase's trust.

Multilayered verification

Algorithmic verification, a reasoning agent, and your CI pipeline each catch what the others cannot. Layered coverage is dramatically better than any single tool on its own.

Independent and zero trust

The agent that reviews the code is not the one that wrote it. A different method, clear segregation of duties, and findings that are auditable and repeatable.

Inside your workflow

Verification that scales with how fast your team and its agents ship, without adding review headcount. Findings land as SonarQube issues, with nothing new to install.

Be the first to hunt and request early access today

Unsubscribe

Request access

Choosing to proceed means that you agree to the storing and processing of your personal data as described in SonarSource’s Cookie Policy. You can opt out of SonarSource communications at anytime.