cyber-harness - AI-driven pi-like agent for cyber security — singl

项目说明

<p align="center">
<img src="web/assets/logo.svg" width="180" alt="cyber logo">
<h1 align="center">cyber-harness</h1>
<p align="center">An en agent harness for cybersecurity</p>
</p>

<p align="center">
<a href=" src=" alt="Release"></a>
<a href=" src=" alt="CI"></a>
<a href=" src=" alt="Downloads"></a>
<a href=" src=" alt="AGPL-3.0"></a>
<a href=" src=" alt="Stars"></a>
</p>

<p align="center">
<a href="README_CN.md">中文文档</a>
</p>

---

cyber-harness is an agent runtime for cybersecurity. Model reasoning, tool execution, knowledge, obser, and collaboration are assembled through the same extension lifecycle. Use the reference distribution or compose the capabilities your own Go application needs.

aiscan packages the agent, core scanners, proxy routing, skills, and IOA collaboration. aiscan-full adds the Web workbench, browser automation, passi, and deep crawling.

cyber-audit is the source-code audit distribution in an independent Go module. It combines model-led in with ripgrep, ast-grep, OS and proton, and supports interacti, one-shot audits and e reports.

> Use only on explicitly authorized targets.

Start here

Download the matching aiscan or aiscan-full archive from GitHub Releases, extract it, and add the executable to PATH. See Getting started (中文) for installation, PowerShell configuration, and a complete first run.

The agent lets the model choose tool calls. The scan pipeline chooses work through rules and scan events, with optional AI stages. Both use the same underlying tool infrastructure.

Configure a model

Run aiscan init to configure your model in ~/.cyber/cyber.yaml. For project o, use aiscan init --project. These are shared cyber-harness commands, also a in the minimal agent CLI. Existing files are preserved. Configuration guide.

Example cyber.yaml:

Set CYBER_API_KEY to your credential. For Anthropic-compatible services, use provider: anthropic and the corresponding settings. See the configuration reference for protocols, profiles, and precedence.

Documentation

Start at the documentation home. Detailed documentation is maintained in Chinese and describes the current source; use the matching Git tag for a release.

Concepts introduces the harness, models, tools, sessions, and knowledge, without requiring Go experience.

The user guide progresses from installation and a first task to sessions, tools, Skills, scanning, and collaboration.

The de guide builds an application in Go, starting with a runnable tool and then an embedded session, extensions, and host integration.

Architecture explains composition and lifetime, the Agent loop, execution, context, and data flow. Configuration and API details remain in the reference.

Build and embed

Use the Go declared in go.mod. Full builds also need Node.js/npm; standard and full builds use CGO_ENABLED=0. Build tags are defined in editions.env; nati requires CGO and is a separate record build.

Arsenal maintains tool definitions and default in the harness arsenal.yaml. Each distribution selects names in cmd/aiscan/bundle.yaml or audit/cmd/cyber-audit/bundle.yaml. See Arsenal bundles for cross compilation, offline distribution, and tool upgrade behavior.

For custom distributions, call harness.BaseExtensions(config), append your extensions, and pass them to extension.New. The host owns Load/Close. The reference distribution is assembled in cmd/aiscan. Embedders can use pkg/harness for a generic host, or compose their own extensions; it does not reproduce the complete aiscan distribution. See the runnable custom example and the extension de guide.

Contributing

Read the de guide and documentation standards. Describe the behavior change, affected entry points, and in your PR. Update the relevant guide or reference with behavior changes.

Disclaimer

1. This tool is intended for authorized security testing and research purposes only. If you need to test its capabilities, please set up your own lab en.
2. Before using this tool for any scanning, you must ensure compliance with local laws and regulations and obtain sufficient authorization. Do not scan unauthorized targets.
3. If you engage in any illegal activity while using this tool, you shall bear all consequences yourself. We assume no legal or joint liability.
4. Before installing and using this tool, please carefully read and fully understand all terms. Limitation and disclaimer clauses may be highlighted for your attention.
5. Unless you ha read, understood, and accepted all terms of this agreement, please do not install or use this tool. Your use or any other express or implied acceptance constitutes your agreement to be bound by these terms.

License

This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).

Links

---

<p align="center">
<a href="
<img src=" alt="Star History" width="600">
</a>
</p>

项目信息

开发语言：Go

Star 数：314

技术要点

该项目属于红蓝攻防方向的开源资源，代码结构清晰，适合安全研究人员学习参考。项目围绕AI-driven pi-like agent for cyber security — single binary for pentest, red team, bug bounty展开，提供了可运行的实现，对理解相关安全机制、工具原理与攻防技术具有较高的研究价值。

从技术实现角度看，此类项目的核心在于对底层原理的深入理解与工程化落地。开发者在研究过程中可以结合官方文档与社区讨论，梳理项目的初始化流程、核心模块职责与数据流转，逐步掌握其中的技术细节与设计思路。

使用与学习建议

1. 下载源码压缩包，解压到本地环境。

2. 根据项目类型准备对应的运行环境与依赖。

3. 按照 README 中的步骤完成配置与部署。

4. 在隔离的测试环境中运行验证，避免影响生产系统。

免责声明

本资源来源于开源社区，仅供安全研究、技术学习与合法授权范围内的攻防演练使用，请勿用于任何非法或侵害他人权益的用途。使用本项目产生的一切后果由使用者自行承担。