h5i

h5i (pronounced "high-five") is a web security testing workspace for AI agents, with the goal of securing web applications in two complementary ways: using AI agents to discover vulnerabilities, and using formal verification to prove correctness. It emphasizes "build with agents, find bugs with red teaming, prove properties with formal verification."

Purpose h5i is aimed at security work for web applications, divided into two major directions:

1. Discover vulnerabilities (red team testing): provide AI agents with a unified toolset for browser automation, HTTP traffic manipulation, and reconnaissance, in order to test existing web applications for vulnerabilities, reproduce attacks, and run security regression checks in CI/CD.
2. Prove correctness (formal verification): build applications using h5i-app (an Axum-based Rust framework) and verify them against Lean 4, proving properties such as authorization, data isolation, and business logic directly against the Rust implementation.

Main capabilities h5i's agent interface is a skill, carried by the binary. Its red team testing capabilities include:

• Browser automation: navigate applications, interact with forms, and test authenticated workflows.
• HTTP security testing: capture, modify, replay, and compare HTTP requests to investigate vulnerabilities.
• Reconnaissance: discover endpoints, parameters, and hidden attack surfaces through crawlers and JavaScript analysis.
• CI/CD: save confirmed attack flows and replay them in CI/CD to catch security regressions.
• Sandboxed workflows: run agents in isolated environments with configurable filesystem and network restrictions. h5i can be used with any web service, regardless of the framework it uses, without migrating to h5i-app.

How to use Open your usual agent runtime environment and enter an instruction like "Audit https://<your-target>.com with h5i." to begin the audit. Agents access these capabilities through a unified CLI, and you can also monitor their activity through a dashboard. The repository materials also mention that man pages can be installed and the full command reference can be viewed.

Formal verification and h5i-app Red team testing finds bugs through testing; formal verification takes a complementary approach: proving that specified properties hold for all possible inputs and behaviors. h5i-app is an Axum-based Rust web framework that makes application logic suitable for handling with Lean 4. You can use Rust to write web applications, convert them to Lean 4 through Aeneas, and then prove various properties such as authorization, isolation, and business logic.

Applicable scenarios

• Red team security testing, vulnerability discovery, and attack reproduction for existing web applications.
• Performing security regression checks in CI/CD pipelines.
• Formal verification of authorization, data isolation, business logic, and more for applications built with Rust and h5i-app.
• Running security testing agents in scenarios that require isolated environments.

Limitations and information not yet provided The repository materials do not provide specific pricing methods, prices, a platform compatibility list, performance results, or detailed usage steps; this content is not yet provided. To learn the full commands and project overview, refer to the official documentation: the official website provides a project overview and Slides, and MANUAL.md or man h5i provides the full command reference. The project is licensed under the Apache License 2.0.