argus-testing 0.3.0

AI-powered exploratory testing agent that discovers bugs like a real user

Argus

AI-powered exploratory QA agent. Give it a URL, it explores your app like a real user — clicking buttons, filling forms, trying edge cases — and finds bugs that scripted tests miss.

Unlike Playwright or Cypress, you don't write test scripts. Argus discovers bugs you didn't think to test for.

What It Catches

Argus runs 16 types of detection across every page it visits:

Category	What it finds
Runtime Errors	Console exceptions, HTTP 4xx/5xx, crashes
Logic Bugs	Fake delete/edit (says "Saved!" but data didn't persist), misleading success toasts
Data Issues	Count mismatches, broken dates ("1.52 days ago"), NaN, eternal "Loading..."
Dead Links	Crawls all internal links, finds 404s and 5xx
Broken Images	Images that failed to load
SEO	Missing meta description, OG tags, heading hierarchy issues
Accessibility	Missing alt text, unlabeled form inputs, no lang attribute
Performance	Slow page loads (>3s), large resources (>500KB), too many requests
Security	Mixed content (HTTP resources on HTTPS pages), XSS reflection

Quick Start (MCP Server for Claude Code)

The recommended way to use Argus. Claude Code becomes the AI brain — no API key needed.

pip install argus-testing
playwright install chromium
claude mcp add argus -- argus-mcp

Then in Claude Code:

"Test my app at http://localhost:3000, focus on the checkout flow"

Claude Code will explore your app, try edge cases, verify that actions persist, and generate an HTML bug report.

MCP Tools (15)

Tool	What it does
start_session(url)	Launch browser, navigate to URL
get_page_state()	See interactive elements + page text + counts + toasts + meta tags + a11y issues
click(index)	Click an element
type_text(index, text)	Type into an input field
select_option(index, value)	Select from dropdown
navigate(url)	Go to a URL
go_back()	Browser back
scroll_down()	Scroll the page
screenshot(name)	Capture the current page
get_errors()	Run all passive detectors (console, network, text, images, SEO, a11y, mixed content...)
verify_action(type, text, url)	Verify a delete/edit actually persisted
check_links()	Crawl all internal links, find dead ones
check_performance()	Measure load time, find large resources
crawl_site(max_pages)	Auto-crawl entire site: visit all internal pages, run all detectors, one command
end_session()	Close browser, generate HTML report

Alternative: Standalone CLI

Bring your own LLM API key. Argus has a built-in AI planner that decides what to explore.

pip install argus-testing
playwright install chromium

export DEEPSEEK_API_KEY=sk-...   # or OPENAI_API_KEY, ANTHROPIC_API_KEY

argus http://localhost:3000 --model deepseek/deepseek-chat -n 50
argus http://localhost:3000 -f "test login with edge cases" --headed

Supports 100+ models via LiteLLM: OpenAI, Anthropic, DeepSeek, Gemini, Ollama (free/local), etc.

Tested on Real Sites

Site	Bugs found	Examples
vanlifeyvr.com	1	Missing og:image
nalifex.com	3	Unlabeled search input, 1.5MB uncompressed image, missing og:image
BuggyTasks (test app)	15+	Fake delete, fake edit, broken dates, count mismatch, XSS, auth bypass

Bug Report

Each session generates a self-contained HTML report with:

• Bug cards with severity, type, description, and reproduction steps
• Embedded screenshots (base64 — no external files needed)
• Testing timeline showing every page visited
• Console and network error logs

How It Works

You give a URL
  -> Argus opens a real browser (Playwright)
  -> AI explores: clicks, types, navigates, tries edge cases
  -> 12 passive detectors analyze every page automatically
  -> On-demand: link crawling + performance metrics
  -> Generates HTML report with all bugs found

Requirements

• Python 3.10+
• Chromium (auto-installed via playwright install chromium)

License

MIT