pm3-mcp-server
Enables AI assistants to control Proxmark3 RFID hardware for MIFARE Classic card reading, writing, and cloning via magic-safe write workflows, with automatic card generation detection, verification, and fallback handling.
README
pm3-mcp-server
Windows-first Model Context Protocol server that wraps the Proxmark3 Iceman/RRG CLI for MIFARE Classic workflows.
Not a raw 1:1 CLI passthrough: the core path is magic-safe write — detect card generation → route the right write command → verify → fallback on silent failure.
Authorized use only. Use on cards and systems you own or have explicit permission to test. Magic-card writes and UID changes have irreversible failure modes (e.g. Gen3 freeze).
Requirements
- Windows + Proxmark3 Iceman client (
proxmark3.exe) already installed and flashed - Python 3.10+
- Serial access to the device (do not run the official Proxmark GUI on the same COM port at the same time)
Configuration
| Variable | Meaning |
|---|---|
PM3_EXE_PATH |
Absolute path to proxmark3.exe (required) |
PM3_CWD |
Working directory for dictionaries / relative paths (defaults to exe parent) |
PM3_PORT |
Serial port (default COM3) |
PM3_TIMEOUT_SEC |
Command timeout (default 30, hard cap 120) |
PM3_KEYS_FILE |
Optional path to keys.local.json |
Copy keys.local.json.example → keys.local.json for site keys (gitignored). The server only reports key counts to the LLM, never the key list.
Vendor bundle folder tags (e.g. V9.0.6.7) are labels only. Real client version comes from pm3_info / proxmark3.exe --version (e.g. Iceman v4.20142). This server does not flash firmware.
Setup
git clone https://github.com/wocessade/pm3-mcp-server.git
cd pm3-mcp-server
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
pytest -q
Smoke-check the client (no card required):
& $env:PM3_EXE_PATH --version
Run
$env:PM3_EXE_PATH = "C:\path\to\proxmark3.exe"
$env:PM3_CWD = "C:\path\to\proxmark3"
$env:PM3_PORT = "COM3"
python server.py
Cursor / Claude MCP snippet
{
"mcpServers": {
"pm3": {
"command": "python",
"args": ["C:\\path\\to\\pm3-mcp-server\\server.py"],
"env": {
"PM3_EXE_PATH": "C:\\path\\to\\proxmark3.exe",
"PM3_CWD": "C:\\path\\to\\proxmark3",
"PM3_PORT": "COM3"
}
}
}
}
Prefer the venv interpreter if you installed deps there:
"command": "C:\\path\\to\\pm3-mcp-server\\.venv\\Scripts\\python.exe"
Tool surface
Primary (use these first)
| Tool | Role |
|---|---|
pm3_info |
hf mf info (+ fallback hf 14a info); UID/ATQA/SAK/card_type + client version |
pm3_dump |
Sector dump + key summary |
pm3_write_magic_safe |
Auto write route + post-write magic verify |
pm3_access_decode |
Local trailer / FF078069 access-bits decode |
pm3_get_card_knowledge |
Magic taxonomy snippets for context |
Advanced (only when primary is insufficient): pm3_detect_magic, pm3_try_keys, pm3_read_block, pm3_read_sector, pm3_write_block, pm3_set_uid.
Magic write taxonomy
| Generation | Write | Verify | Never auto |
|---|---|---|---|
| Gen1a/1b | csetblk / csetuid (optional csetblk --gdm = Gen1a wakeup only) |
cgetblk |
gen3freeze, gdmsetcfg |
| Gen2 / CUID | wrbl (--force for block0) |
rdbl |
same |
| Gen3 | gen3uid / gen3blk |
gen3/info paths | gen3freeze; not csetblk --gdm |
| Gen4 GDM | gdmsetblk (then csetblk --gdm variant) |
magic read | gdmsetcfg |
Device-facing responses include _raw (truncated) for graceful parse fallback.
Tests
pytest -q
Covers parser fixtures, access bits (FF078069), write router (Gen3 ≠ --gdm), and client validation (injection / whitelist / batch size).
Live hardware notes
- Place a card on the antenna before
pm3_info/ write tools. - No card →
ok: false,reason: "no_card"(no retry loop). - Gen4 GDM:
wrblon trailers can report success while swallowing data —pm3_write_magic_safeverifies and falls back. - Firmware flashing is out of scope.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.