Capability

Find What Humans Miss

AI explores every page, every interaction, every edge case. Your QA team handles the judgment calls. CBrowser handles the exhaustive sweep.

200 pages. 10 minutes. No test scripts to write. No selectors to maintain.

Scan Your Site

What It Catches

Four categories of issues that manual testing consistently misses at scale. AI vision and DOM analysis work together on every page.

Visual Bugs

Overlapping elements, broken layouts, responsive failures, z-index issues, and missing images. AI vision inspects every viewport.

Functional Bugs

Broken links, console errors, failed API calls, form submission failures, and dead buttons. Catches the issues users report in support tickets.

Accessibility Bugs

Missing labels, focus traps, color contrast violations, keyboard navigation failures, and missing ARIA attributes. Beyond what Lighthouse flags.

Security Issues

XSS vectors, exposed data in the DOM, insecure form actions, mixed content, and MCP tool injection patterns.

How hunt_bugs Works

No configuration files. No test scripts. No CSS selectors. Point it at a URL and let AI do what it does best: systematically explore everything.

Give it a URL

Point hunt_bugs at any page or your entire domain. Set the crawl depth and max pages.

AI explores everything

Every link, every interaction, every form. Vision analysis and DOM inspection work together to find what static analysis misses.

Get a prioritized bug list

Each bug includes severity, screenshot, reproduction steps, and a suggested fix. Critical issues surface first.

Chaos Testing

Your users don't always have perfect conditions. What happens when the API takes 8 seconds? When JavaScript fails to load? When the network drops to 3G?

chaos_test introduces controlled failures and measures how your application degrades. Graceful degradation is a feature. Silent failure is a bug.

API timeout simulation (2s, 5s, 10s, 30s)

JavaScript failure injection

Network throttling (3G, slow 4G, offline)

chaos_test({
  url: "https://yoursite.com",
  scenarios: [
    "api_timeout_5s",
    "js_disabled",
    "network_3g"
  ],
  measure: [
    "error_handling",
    "loading_states",
    "graceful_degradation"
  ]
})

// Capture baseline after deploy
visual_baseline({
  url: "https://yoursite.com",
  pages: ["/", "/pricing", "/signup"]
})

// Compare after next deploy
visual_regression({
  url: "https://yoursite.com",
  pages: ["/", "/pricing", "/signup"]
})
// Result: 2 pages match,
// /pricing has 3 pixel-level diffs

Visual Regression

Compare screenshots across deploys. Detect unintended visual changes before users do. A CSS change that fixes one page and breaks three others gets caught in seconds.

Capture baselines with visual_baseline, then run visual_regression after every deploy. Pixel-level diffs with highlighted change regions.