Am I Exposed MCP
A Bitcoin privacy scanner with CLI and MCP that scores transactions, addresses and wallets on 34 heuristics from 0 to 100.
- 109 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Requests to `mempool.space` reveal your IP and addresses, and an `xpub` reveals every wallet address.
Install
- Works in
Manual install
npm install -g am-i-exposedCommand from the CLI and agent integration section. MCP server: `npx -y am-i-exposed mcp`.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Requests to `mempool.space` reveal your IP and addresses, and an `xpub` reveals every wallet address.
This is third-party code. Review the repository files before installing.
What it does
Am I Exposed shows what the blockchain reveals about a transaction or wallet owner. The web version runs in a browser, and for agents there is a console utility and an MCP server. Analysis runs on the device: 34 heuristics, multi-hop tracing, entity matching and Boltzmann entropy via Rust and WebAssembly. The utility takes a transaction ID, address, `xpub`, PSBT or raw transaction. Commands: `scan tx`, `scan address`, `scan xpub`, `scan psbt`, `boltzmann` and `chain-trace`. The MCP server starts with `am-i-exposed mcp` and exposes five tools: `scan_transaction`, `scan_address`, `scan_psbt`, `scan_wallet` and `compute_boltzmann`.
Who it is for. Bitcoin wallet owners and developers who want to check a transaction privacy before broadcasting.
Good fit when
- A PSBT is reviewed before sending to see which links it reveals.
- A wallet audit by `xpub` shows address reuse.
- The agent runs `chain-trace` at depth 3 to examine a payment chain.
- A report needs JSON with a score and a list of findings.
Not a fit when
- You need to pay or sign: the scanner only analyzes.
- Requests to a public API are against security policy.
- A legal conclusion about fund origin is required: scores are heuristic.
Example request
Analyze the privacy of this transaction and list findings starting with the most serious.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
Analysis runs on the device, but the utility fetches chain data through the `mempool.space` API, whose servers see your IP and requested addresses. For better privacy the README recommends Tor Browser, which routes requests to the `.onion` endpoint, or your own API instance via `--api`. Passing an `xpub` reveals the wallet addresses to the service. The 0-100 score is heuristic and no substitute for legal expertise. Responses are cached in SQLite, and `--no-cache` turns that off.
How to disable. Remove the MCP entry from the client settings and run `npm uninstall -g am-i-exposed`. Delete the SQLite cache by hand.
FAQ
Which tools does the MCP expose?
Five: `scan_transaction`, `scan_address`, `scan_psbt`, `scan_wallet` and `compute_boltzmann`. Full docs are in `cli/skill.md`.
How do I run it from a terminal?
With `npx am-i-exposed scan tx <txid> --json`. The `--fast` flag skips parent transaction context.
Can it work offline?
Only with PSBT: `scan psbt <file>` makes no network calls. Other commands need chain data.
Are test networks supported?
Yes, the `--network` flag accepts `mainnet`, `testnet4` and `signet`.
How do I use my own node?
With the `--api <url>` flag pointing to your own `mempool` instance, for example on Umbrel.
Discussion
No questions or reviews yet.
Related
MCP serversSecurity scanner for MCP servers that looks for hardcoded keys, tool poisoning and injection flaws in repository code.
A scanner that finds leaked keys, dangerous access and connected models in MCP configs of Claude Desktop, Cursor and VS Code.
Security scanner for coding agents: checks code, MCP servers, prompts and model-invented packages, and gives an A to F grade.
A local dashboard for Claude Code, Codex CLI and OpenCode: what the agent loads, its token cost and an MCP server security scan
MCP-Shield finds your MCP configs, connects to servers and scans tool descriptions for hidden commands, exfiltration channels and tool shadowing
An agentic web scanner: terminal workspace, six stages from recon to report, browser validation of findings, REST and MCP
README in short
The project grew from the closure of the OXT and KYCP services. It computes heuristics and a privacy score and shows recommendations, all client-side with no backend.