# CyberStrike

> For the full, detailed version of this file see:
> [llms-full.txt](https://cyberstrike.io/llms-full.txt)

> CyberStrike is an open-source, autonomous AI agent for offensive security
> testing. It connects to 150+ AI providers (Claude, GPT, Gemini, DeepSeek,
> Ollama and more), ships 7,600+ security skills and 176+ MCP tools, and
> runs entirely self-hosted with bring-your-own-key (BYOK) support.
> Licensed AGPL-3.0.

## Key Features

- **150+ AI Providers**: OpenAI, Anthropic, Google, DeepSeek, Ollama, vLLM, AWS Bedrock, Azure, and more
- **7,600+ Security Skills**: Recon, web app testing, cloud security, network analysis, post-exploitation
- **176+ MCP Tools**: HackBrowser, Cloud Audit, CVE Database, OSINT, GitHub Security, AI Knowledge Graph
- **Bolt Server**: Distributed scanning infrastructure for parallel target assessment
- **Multi-Agent Cluster**: Coordinated agent orchestration for complex engagements
- **Agentic Validation**: Three-stage pipeline (baseline → attack → compare) with PoC generation
- **Web UI**: Real-time dashboard with terminal, chat, findings, and agent monitoring
- **5,300+ Models**: Works with any OpenAI-compatible API endpoint

## Docs

- [Getting Started](https://docs.cyberstrike.io/): Installation, configuration, and first scan
- [Platform Overview](https://cyberstrike.io/): Core capabilities — AI providers, security skills, Bolt MCP tools, HackBrowser
- [Blog](https://cyberstrike.io/blog): Threat research, product updates, and security analysis

## Code

- [GitHub Repository](https://github.com/CyberStrikeus/CyberStrike): Full source code, issues, and contributions
- [Install Script](https://cyberstrike.io/install.sh): One-line installer for Linux and macOS

## Optional

- [About](https://cyberstrike.io/about): Team and mission
- [FAQ](https://cyberstrike.io/faq): Frequently asked questions
- [RSS Feed](https://cyberstrike.io/rss.xml): Blog feed for content syndication