HexStrike AI MCP Agents
An offensive security MCP server: the agent runs 150+ pentest and recon tools and drives automated target assessments
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- Runs offensive tools against real targets
- Actions may be illegal without the target owner's authorization
- Executes system commands with your privileges
Install
Manual install
git clone https://github.com/0x4m4/hexstrike-ai.git
cd hexstrike-ai
python3 -m venv hexstrike-env
source hexstrike-env/bin/activate
pip3 install -r requirements.txt
python3 hexstrike_server.pyStarts the local server on port 8888. Install the required security tools on the system separately.
This is third-party code. Review the repository files before installing.
What it does
The server bridges an MCP-compatible agent to a set of 150+ security tools: network scanners (nmap, rustscan, masscan), web tools (ffuf, nuclei, sqlmap, wpscan), password cracking (hydra, john, hashcat), reverse engineering and binaries (radare2, ghidra, binwalk), plus cloud and OSINT tools. A decision engine selects tools and parameters for a target, and separate agents help with bug bounty, CTF, CVE analysis and attack-chain discovery. It runs as a local server the agent connects to through the hexstrike_mcp.py bridge. The tools themselves must be installed on the system.
Who it is for. For security professionals, pentesters, bug bounty and CTF participants who are authorized to test their targets.
Good fit when
- You need to automate recon and scanning in an authorized pentest
- You want an agent that selects tools for a target itself
- You want help with bug bounty or solving CTF tasks
Not a fit when
- You have no written authorization to test the target
- You cannot install dozens of external security tools on the system
Example request
Run recon on a domain I am authorized to test and propose an assessment planLimitations
The server only orchestrates utilities; the tools themselves (nmap, sqlmap, ghidra and others) must be installed separately, and some need Kali Linux or Chrome. Tools run with your privileges and can affect real targets, so use them only where you are authorized. No published package: install from a clone and run hexstrike_server.py.
How to disable. Remove the hexstrike-ai entry from the MCP client config, stop the hexstrike_server.py process and delete the cloned repository.
MCP
- Transport
- stdio
- Authentication
- not required
Security check
- Runs offensive tools against real targets
- Actions may be illegal without the target owner's authorization
- Executes system commands with your privileges
README in short
The README describes HexStrike AI as an offensive security MCP platform with 150+ tools and several autonomous agents. The architecture includes a decision engine, tool and parameter selection, caching and error recovery. Setup: clone the repo, venv, dependencies, run hexstrike_server.py and connect hexstrike_mcp.py to Claude Desktop, Cursor, VS Code Copilot, Roo Code and other MCP clients. It lists the network, web, password, binary, cloud and OSINT tools you install separately. MIT licensed.
FAQ
Do the tools come bundled?
No. The server calls utilities installed on the system; you install them separately, and some assume Kali Linux.
How does the agent connect to the server?
First hexstrike_server.py runs on a local port, then the client connects the hexstrike_mcp.py bridge pointing at that server.
Related
A code security audit skill by Cloudflare: the agent runs recon, coverage-led hunting and independent verification of findings, then produces a structured repor
NVIDIA's open stack for running OpenClaw, Hermes and LangChain Deep Agents in OpenShell sandboxes with network policy and managed inference
Security scanner for agent skills and MCP servers: finds prompt injection, data exfiltration and supply chain risks before install
Static code analysis with rules that look like source code, plus a built-in MCP server for AI agents