Skip to main content
Glama
muaviamohammed

io.github.muaviamohammed/mizan

mizan-mcp — verify any AI trading agent's track record (zero-knowledge proof of performance)

mcp-name: io.github.muaviamohammed/mizan

Use it before you allocate to, copy, or pay any trading bot or AI agent that claims returns.

An MCP server that lets any agent verify a trading or AI-agent track-record credential, or mint one for its own record, without ever revealing the strategy.

MIZAN credentials are zero-knowledge (STARK) proofs under the open VTR-1 standard: the record is recomputed inside the proof with the trial count, cost schedule and out-of-sample split bound in-circuit. Anyone verifies a credential offline in ~81 ms. FAIL verdicts are published alongside passes.

Prove the edge. Never reveal the strategy.

Remote (no install)

https://studio.mizan.market/mcp        # streamable HTTP, stateless

Claude Code: claude mcp add --transport http mizan https://studio.mizan.market/mcp

Related MCP server: agentkeychain

Local (stdio)

pip install "mcp>=2"
python3 mizan_mcp_server.py
# Claude Code: claude mcp add mizan -- python3 /path/to/mizan_mcp_server.py

Tools

Tool

What it does

check_agent(agent)

Start here. Is this agent proven? Look up by name, wallet, ERC-8004 id, agent-card URL, credential id or sha256. PROVEN, REFUSED, NOT SIGNIFICANT or UNVERIFIED; unverified answers carry a message to forward to the counterparty.

list_proven_agents(limit)

PASS credentials, Bitcoin-confirmed first; issuer shown per row.

verify_credential(credential_id | sha256)

Registry lookup, proof-bundle download, sha256 check against the registry anchor, optional offline STARK re-verification. Returns PASS, FAIL or NOT SIGNIFICANT exactly as issued.

list_datasets()

Canonical public markets a credential can be minted on.

preview_credential(dataset, positions, model_id)

Free, private, no-account gate test of your own bot or agent. Nothing is published.

mint_credential(dataset, positions, model_id, email, password, …)

Sealed path: sends one position per bar-move only. Preview free; real STARK minted on an H100.

Rules the server enforces

  • Never send a strategy, model, weights, code or prompt anywhere. Positions only.

  • Never claim a PASS without a bundle whose hash matches the registry.

  • A FAIL is a result, not a secret. A credential is a backtest seal, not a forecast.

  • HFT and fill-dependent strategies are refused, permanently.

Also available as

  • AgentKit action provider (mizan): verify_credential + list_datasets — PR to coinbase/agentkit.

  • A2A agent: https://studio.mizan.market/a2a (JSON-RPC message/send → "verify "); card at https://mizan.market/.well-known/agent-card.json.

  • ERC-8004 validator on Base: registration file at https://mizan.market/.well-known/agent-registration.json.

Skill for agents (SKILL.md, tool schemas, A2A card): https://mizan.market/skill/ · Public wall: https://mizan.market/registry/ · Verify by hand: https://mizan.market/verify/ · ERC-8004: https://mizan.market/.well-known/agent-registration.json · Contact: muavia@mizan.market

MIT. © 2026 Mizan Verification, Inc.

Keywords

AI trading agent verification · backtest overfitting · deflated Sharpe ratio · proof of performance · verifiable track record · zero-knowledge proof · zkVM · RISC Zero · STARK · ERC-8004 · A2A · x402 · Hyperliquid vault due diligence · trading bot audit

Related MCP Connectors

Related MCP Servers

  • A
    license
    A
    quality
    C
    maintenance
    Enables AI agents to sign decisions with post-quantum cryptographic proofs and maintain secure audit trails for compliance. It provides tools for stamping events, verifying chain integrity, and exporting audit data across industries like finance and healthcare.
    4
    39 npm
    MIT
  • A
    license
    Not graded
    quality
    A
    maintenance
    Provides a zero-knowledge credential vault for AI agents, allowing secure storage, retrieval, and management of secrets with cross-agent delegation and tamper-evident audit.
    MIT
  • A
    license
    A
    quality
    B
    maintenance
    Enables AI agents to access verifiable DePIN supply-side telemetry, browse and purchase data products using credits, query datasets, and verify data provenance with cryptographic and zero-knowledge tools.
    10
    386 npm
    MIT