KadeMCP
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
KadeMCP - CTF Solver MCP Server

A Model Context Protocol server that turns Claude into a hands-on CTF solver.
Connect this server to Claude Desktop and Claude can disassemble binaries, crack hashes, fuzz web apps, extract steganography, and run full pwntools exploits — all from natural language prompts.
Categories
Category | Tools |
Reversing / Pwn |
|
Crypto |
|
Web |
|
Forensics / Stego |
|
Related MCP server: CTF MCP Server
Quick start
Option A — Local (Python)
git clone https://github.com/orgito1015/KadeMCP
cd KadeMCP
pip install -r requirements.txt
# System tools (Debian / Ubuntu)
sudo apt install binutils gdb binwalk tshark exiftool steghide hashcatAdd to Claude Desktop config:
macOS —
~/Library/Application Support/Claude/claude_desktop_config.jsonWindows —
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"ctf-solver": {
"command": "python3",
"args": ["/absolute/path/to/KadeMCP/server.py"]
}
}
}Option B — Docker (recommended for pwn/reversing)
git clone https://github.com/orgito1015/KadeMCP
cd KadeMCP
docker build -t kademcp .{
"mcpServers": {
"ctf-solver": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"--network", "host",
"-v", "/path/to/challenges:/challenges",
"kademcp"
]
}
}
}Tool reference
Reversing / Pwn
Tool | Description |
| objdump disassembly — pass |
|
|
| Check NX, PIE, RELRO, stack canary |
| Execute binary with optional stdin, capture output |
| Run a full Python pwntools exploit — Claude writes it, this runs it |
Crypto
Tool | Description |
| base64, hex, rot13, url, binary, morse, ascii — encode or decode |
| Guess hash type from length and pattern |
| hashcat + rockyou for md5/sha1/sha256/sha512/ntlm |
| Single-byte XOR brute-force ranked by English frequency, or known-key XOR |
| Letter frequency for classical cipher cracking |
| Arbitrary Python — pycryptodome, gmpy2, sympy available |
Web
Tool | Description |
| HTTP GET/POST/PUT/DELETE with headers, cookies, JSON body |
| Fuzz a |
| Automated SQLi detection (error, boolean, time-based) |
| Path traversal / LFI payload testing |
| Arbitrary Python — requests, BeautifulSoup4, httpx available |
Forensics / Stego
Tool | Description |
|
|
| Scan and optionally extract embedded files |
|
|
| Tries steghide, zsteg (PNG/BMP), outguess in sequence |
| Full exiftool dump — flags often hide in EXIF fields |
| tshark modes: summary, http_objects, dns, credentials, follow_tcp |
|
|
Example prompts
"Here's a binary (base64): <...> — find the flag. Start with checksec and strings."
"Crack this hash: 5f4dcc3b5aa765d61d8327deb882cf99"
"The site at http://ctf.local/search?q=FUZZ is probably injectable. Run sqli_test."
"This JPEG might have a hidden message. Try all stego methods."
"Analyze challenge.pcap — look for credentials in HTTP and FTP streams."
"The ciphertext looks like single-byte XOR: 1a3f2b0e... brute-force it."Project structure
KadeMCP/
├── server.py # MCP entry point — registers all tools
├── tools/
│ ├── __init__.py
│ ├── utils.py # Shared: run_cmd, run_python, helpers
│ ├── reversing.py # Reversing + pwn tools
│ ├── crypto.py # Crypto tools
│ ├── web.py # Web exploitation tools
│ └── forensics.py # Forensics + stego tools
├── .github/
│ └── workflows/
│ └── ci.yml # Lint + import check on push/PR
├── Dockerfile
├── requirements.txt
├── pyproject.toml
├── LICENSE
└── README.mdExtending
Each category is self-contained. To add a tool:
Open
tools/<category>.pyAdd a
Tool(name=..., description=..., inputSchema=...)to the listAdd an
if name == "..."branch in thehandle_*function
No changes to server.py needed — it auto-registers everything.
Security
These tools execute arbitrary code and binaries by design. run_python,
run_crypto, run_webscript, and run_pwnscript run arbitrary Python;
run_binary runs uploaded/target binaries; other tools shell out to system
utilities. There is no in-process sandbox — the isolation boundary is the
Docker container (which runs as the non-root ctf user).
Prefer the Docker option; don't run the server directly on a host you care about.
Never expose the server on a network port — anyone who can reach it runs code as the server user.
Treat the connected MCP client as fully trusted.
Optional hardening: set
KADEMCP_WORKSPACE=/some/dirto confine user-suppliedfilepatharguments to that directory (does not sandbox code execution).
See SECURITY.md for the full threat model.
Development
pip install -e ".[dev]" # installs pinned mcp (<2), pytest, ruff, build
ruff check . --select E,F,W,I --ignore E501
pytest -q # offline tests; no external CTF tools requiredNote:
server.pytargets the mcp 1.x decorator API, so the dependency is pinned tomcp>=1.0.0,<2. Installing an unpinnedmcpwill pull 2.x and break startup.
License
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Hunt zero-days by talking to binaries. 40+ tools. Hosted, OAuth + SSO, invite: hi@byteray.ai
34 security tools for AI agents: phishing links, exposure maps, DNS, SSL, PQC, history, pivots.
AI Agent Utility Layer for agent safety, web monitoring, crypto research and tool discovery.
Train, evaluate, red-team and deploy AI agents: prompt tuning, LLM-judged evals, katas.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceConnects AI assistants to 55+ Kali Linux security tools for automated CTF solving, penetration testing, and security analysis across 7 categories including cryptography, forensics, web security, and binary exploitation.80-
- AlicenseNot gradedqualityNot gradedmaintenanceExposes common CTF and cybersecurity tools (crypto, forensics, malware analysis, steganography, reverse engineering, pwn, OSINT) so LLMs can help solve capture-the-flag challenges in a controlled lab environment.-
- AlicenseAqualityDmaintenanceGives AI agents a browser that bypasses bot detection, a MITM proxy for traffic interception, and a Python sandbox to autonomously write and execute security exploits.393MIT
- AlicenseNot gradedqualityBmaintenanceEnables LLM clients to conduct authorized penetration testing and CTF challenges through a multi-mode agent with evidence-chain anti-hallucination, skill evolution, and MCP tool orchestration.MIT