# CyberStrikeAI — AI Security Testing Platform (MCP)

> CyberStrikeAI is a Go-based AI security testing platform with MCP transports and an auditable web UI; verified 3,783★ and starts via `./run.sh`.

## Install

Save the content below to `.claude/skills/` or append to your `CLAUDE.md`:

## Quick Use

```bash
git clone https://github.com/Ed1s0nZ/CyberStrikeAI.git
cd CyberStrikeAI
chmod +x run.sh && ./run.sh
# Open http://localhost:8080 and configure your model credentials in Settings (README).
```

## Intro

CyberStrikeAI is a Go-based AI security testing platform with MCP transports and an auditable web UI; verified 3,783★ and starts via `./run.sh`.

**Best for:** Authorized security teams who want a tool-orchestrated, logged workflow with MCP integration

**Works with:** Go 1.21+, Python 3.10+, and an OpenAI-compatible model endpoint (per README quick start)

**Setup time:** 15-35 minutes

### Key facts (verified)

- GitHub: 3783 stars · 634 forks · pushed 2026-05-13.
- License: Apache-2.0 · owner avatar + repo URL verified via GitHub API.
- README-backed entrypoint: `./run.sh`.

## Main

- Use it only for authorized engagements: validate your scope and keep the platform on a private network; treat logs/audit as part of the deliverable.

- Start with the built-in one-command deploy, then configure an OpenAI-compatible endpoint in Settings before running any orchestration workflows.

- Keep integrations minimal at first: prove MCP stdio mode works with your client, then add plugins (e.g., Burp extension) only when needed.

### Source-backed notes

- README provides a one-command deployment: `chmod +x run.sh && ./run.sh`, and lists Go 1.21+ and Python 3.10+ as prerequisites.
- README describes native MCP support with multiple transports (HTTP/stdio/SSE) and a password-protected web UI with audit logs.
- README notes first-time configuration requires setting an OpenAI-compatible API key/base URL/model before use.

### FAQ

- **Is it safe to run on a public host?**: Treat it as an internal tool: run on private networks and enable auth (README mentions password-protected UI).
- **Do I need to install 100+ tools first?**: No — start with the core platform; README treats tool installs as optional and incremental.
- **Can I use it with my MCP client?**: Yes — README lists MCP transports including stdio; test with a minimal workflow first.

## Source & Thanks

> Source: https://github.com/Ed1s0nZ/CyberStrikeAI
> License: Apache-2.0
> GitHub stars: 3783 · forks: 634

---

<!-- ZH -->

## Quick Use

```bash
git clone https://github.com/Ed1s0nZ/CyberStrikeAI.git
cd CyberStrikeAI
chmod +x run.sh && ./run.sh
# Open http://localhost:8080 and configure your model credentials in Settings (README).
```

## Intro

CyberStrikeAI 是基于 Go 的 AI 原生安全测试平台，提供 MCP 传输（HTTP/stdio/SSE）与可审计 Web UI/日志；已验证 3,783★，并支持一键 `./run.sh` 部署。

**Best for:** 需要“授权测试”前提下，用 MCP + 审计日志来编排工具链的安全团队

**Works with:** Go 1.21+、Python 3.10+，以及 OpenAI-compatible 模型接口（README Quick Start）

**Setup time:** 15-35 minutes

### Key facts (verified)

- GitHub：3783 stars · 634 forks；最近更新 2026-05-13。
- 许可证：Apache-2.0；作者头像与仓库链接均已通过 GitHub API 复核。
- README 中可对照的入口命令：`./run.sh`。

## Main

- 只用于授权测试：先确认测试范围，把平台放在私有网络内运行，并把日志/审计作为交付的一部分。

- 先用 README 的一键部署跑通，再在 Settings 配置 OpenAI-compatible endpoint，最后再开启编排流程。

- 集成从简：先验证 MCP stdio 能与客户端稳定工作，再按需启用插件（如 Burp 扩展）。

### Source-backed notes

- README 给出一键部署：`chmod +x run.sh && ./run.sh`，并列出 Go 1.21+ 与 Python 3.10+ 作为前置条件。
- README 描述原生 MCP 支持（HTTP/stdio/SSE 等传输）以及带密码的 Web UI 与审计日志。
- README 说明首次使用前需要配置 OpenAI-compatible 的 API key/base_url/model。

### FAQ

- **适合直接暴露到公网吗？**：不建议。把它当作内网工具：私网运行并启用认证（README 提到密码保护 UI）。
- **必须先装齐 100+ 工具吗？**：不必。先跑通核心平台；工具安装在 README 里属于可选、按需增量。
- **能接入我的 MCP 客户端吗？**：可以。README 提到 stdio 等传输；建议从最小流程开始验证。

## Source & Thanks

> Source: https://github.com/Ed1s0nZ/CyberStrikeAI
> License: Apache-2.0
> GitHub stars: 3783 · forks: 634


---
Source: https://tokrepo.com/en/workflows/cyberstrikeai-ai-security-testing-platform-mcp
Author: AI Open Source