AIPentest/CyberStrikeAI

Project overview

It combines governed execution, evidence retention, and attack-chain replay in one auditable workspace, with 100+ curated YAML tool recipes and MCP-native tool integration for authorized security testing.

Project type

MCP · AI Agent

Use cases

Automation

Deployment

Refer to project documentation

License

Apache-2.0

Best for

• Operations and security teams that need a unified, auditable workspace for authorized security testing.

Key capabilities

• Translates natural-language intent into governed, auditable security actions with multi-agent orchestration modes.
• Ships 100+ curated YAML tool recipes plus MCP-native tool integration with dynamic discovery.
• Combines query rewriting, vector retrieval, reranking, and result post-processing.

Limitations and risks

• The project is for educational and authorized testing purposes only and must not be used on systems without clear written authorization.

Getting started

• Clone the repository, run ./run.sh (which automates environment checks, dependency install, build, and server start), then configure an AI channel and log in via the web GUI. No coding is required.

Evidence and sources

• README: cmd/ # Server, MCP stdio entrypoints, tooling
• README: CyberStrikeAI connects planning, execution, human oversight, evidence, and replay in one auditable workspace.
• README: 🤖 **Agentic execution** translates natural-language intent into governed, auditable security actions.
• README: CyberStrikeAI ships with 100+ curated tools covering the whole kill chain:
• README: 🔌 **MCP integration** supports HTTP, stdio, SSE, external federation, and dynamic tool discovery.

Source updated at: 2026-10-08T18:32:04Z

Open original GitHub project