General · EdgepediaSeptember 17, 20266 min read

Axe

Axe (also styled AXE) is the name of several unrelated software products and research systems. The best known is axe-core, an open-source accessibility testing engine maintained by Deque Systems; other current uses include AXE.AI, a commercial offensive-security platform, Axe, a terminal-based command-line framework for running AI agents, and AXE, a research framework for automated exploitation of web vulnerabilities.

Key fact	Detail
axe-core	Open-source accessibility testing engine for websites and other HTML-based user interfaces, testing against WCAG 2.0, 2.1, and 2.2 at levels A, AA, and AAA¹
Axe tool suite	Deque's Axe DevTools, Axe Auditor, and Axe Monitor are built on the axe-core engine; the platform itself is closed source while axe-core and its test-runner integrations are open¹^•²
AXE.AI	Commercial AI-driven cloud native offensive security platform for penetration testing, with engagements priced from $1,500 for professional services to $15,000+ for managed service³
Axe (CLI)	Open-source terminal tool for running AI agents defined in TOML configuration files with SKILL.md skill files; agents accept stdin and write stdout⁴^•⁵
AXE (research)	Agentic eXploit Engine, a multi-agent framework for web application exploitation that reached a 30% exploitation success rate on the CVE-Bench dataset, a 3x improvement over state-of-the-art black-box baselines⁶

axe-core, the accessibility engine

Axe-core is the automated testing rules library that Deque Systems maintains for testing websites and other HTML-based user interfaces against the Web Content Accessibility Guidelines (WCAG), versions 2.0, 2.1, and 2.2, at conformance levels A, AA, and AAA¹. Deque describes it as the engine that powers its other Axe products, including Axe DevTools, Axe Auditor, and Axe Monitor¹. The company's documentation positions Axe DevTools for Web as a way to integrate automated accessibility testing into development processes and track results over time, built on top of axe-core⁷.

The engine's scope is HTML-based interfaces. Deque's mobile testing rulesets for iOS and Android are written separately, to evaluate native app views and the accessibility tree provided by the Android and iOS platforms rather than HTML¹. The axe tools on GitHub are built on the axe-core engine and list more than 100 contributors from around the world⁸. Deque also offers free, open-source linter packages, such as @axe-core/react, that surface accessibility issues directly in the code editor⁹.

Naming history. In October 2020 Deque announced that the tool suite formerly known as WorldSpace would unite under the axe name, a change intended to emphasize the company's reliance on the axe-core engine¹⁰. A November 2024 accessibility conformance report for the axe Developer Hub records the use of axe-core version 4.10.0 in the evaluation tooling¹¹.

Automated accessibility checking has a stated limit that Deque's own guidance and third-party guidance both emphasize: users need to check or verify that the issues a tool finds are real issues, rather than false positives or false assurances¹².

AXE.AI, the security testing platform

AXE.AI is a commercial cybersecurity platform that describes itself as an AI-driven cloud native offensive platform providing penetration testing capabilities³. Its stated services include AI-driven discovery and exploration of potential findings, real-time vulnerability assessment, reconnaissance, and automated reporting³. The company positions the product as combining dozens of single-use security tools, such as Burp Suite for web application testing and Faraday for vulnerability management, into a single dashboard for scanning, exploitation, and reporting¹³^•¹⁴.

Published pricing lists professional services at $1,500, an Enterprise Managed tier at $5,000 to $10,000+, and an Axe.ai Managed tier at $15,000+³. The company's LinkedIn presence states that AXE.AI and Strong Crypto are becoming one company, AXE AI Corp¹⁵.

Axe, the AI agent CLI

A separate product named Axe is an open-source command-line tool for running AI agents, marketed with the tagline "Built to cut. Not to chat."⁴. An agent is defined by a TOML configuration file placed in a configuration directory, with a name, a model, and a pointer to a SKILL.md file describing the agent's task; the axe run command accepts input on stdin and writes output to stdout, so agents can be wired into cron jobs, git hooks, or continuous integration pipelines and chained with pipes⁴^•⁵.

The tool supports multiple model providers, including Anthropic, OpenAI, Ollama, and Gemini, using the models.dev format so that users supply their own API keys⁴. Its built-in file tools (list_directory, read_file, write_file, and edit_file) are sandboxed to the agent's working directory, rejecting absolute paths, .. traversal, and symlink escapes; the documentation notes that the shell command tool's path validation is a heuristic guard rather than airtight sandboxing and recommends running axe inside Docker for full sandboxing¹⁶. An independent review scored the tool 70 out of 100, noting its multi-provider support, sub-agent delegation with depth limiting, persistent markdown memory, and SSRF protection on its web-fetching tools¹⁷. The open-source CLI is free, while the hosted platform at getaxe.dev, which runs agents in isolated containers with persistent memory and run history, has no free tier⁴.

AXE, the exploit research framework

In security research, AXE (Agentic eXploit Engine) is a multi-agent framework for web application exploitation that maps lightweight detection metadata, specifically a CWE classification and a vulnerable code location, to concrete exploits through decoupled planning, code exploration, and dynamic execution feedback⁶. Evaluated on the CVE-Bench dataset, it achieved a 30% exploitation success rate, a 3x improvement over state-of-the-art black-box baselines, and produced actionable, reproducible proof-of-concept artifacts for successful exploits⁶.

Distinguishing the uses

The shared name covers four distinct things with different purposes and audiences: axe-core is a browser-side accessibility rules engine embedded in testing tools and CI pipelines; AXE.AI sells human-guided offensive security testing on a commercial platform; the Axe CLI is a developer utility for composing and running AI agents locally; and the AXE research framework automates the confirmation of reported vulnerabilities. Context usually disambiguates them: accessibility content refers to axe-core and Deque, cybersecurity content to AXE.AI or the research paper, and AI-agent tooling to the CLI.

References

Topic: Encyclopedia › Technology and the built world › Communications and everyday technology › Titles 3 to Lm

Initially written Sep 17, 2026 · Reviewed: Sep 30, 2026; Oct 15, 2026 · Edited: Sep 30, 2026; Oct 15, 2026 · Last review: Oct 15, 2026

Notice something wrong?

Axe