valta-mcp

An MCP (Model Context Protocol) server for Valta — financial governance for AI agents. Exposes Valta's spend gate and audit trail as MCP tools, so any MCP-compatible client (Claude Desktop, Claude Code, Cursor, and others) can call them directly.

What this is — and isn't

This server is a thin wrapper over Valta's real, hosted REST API. It does not implement any spend logic, limits, or hashing itself — every tool call here is a pass-through to https://valta.co/api/v1/..., where the actual enforcement (spend gate, audit chain) happens. This repo is the protocol adapter, not the enforcement engine.

Important, read before relying on this for anything security-sensitive: a tool description telling a model "call this before spending" is guidance, not enforcement. This server governs spend that routes through its valta_request_spend tool. It does not intercept or prevent spending that happens through other tools, APIs, or webhooks that don't call it first. If you need a hard guarantee that no spend can happen without authorization, the money-moving action itself needs to be implemented as (or wrapped by) a Valta-gated tool — not merely preceded by a polite check-in step. See Design notes below.

Install

npm install -g valta-mcp

Or run directly without installing:

npx valta-mcp

Configuration

Get an API key at valta.co/dashboard → Settings → API Keys → Create key.

Add to your MCP client's config (e.g. Claude Desktop's claude_desktop_config.json):

{
  "mcpServers": {
    "valta": {
      "command": "npx",
      "args": ["valta-mcp"],
      "env": {
        "VALTA_API_KEY": "sk_valta_your_key_here"
      }
    }
  }
}

Tools

Tool	What it does
valta_check_balance	Read an agent's wallet balance and limits
valta_request_spend	Request authorization for a spend — approved or denied by the real spend gate
valta_create_wallet	Create a new named wallet with optional spend limits
valta_freeze_agent	Kill switch — freeze an agent's wallet, blocking further spend
valta_unfreeze_agent	Resume a frozen agent's wallet
valta_get_audit_trail	Read the hash-chained audit trail of spend decisions

Every tool's description in src/index.ts states plainly what it enforces and what it doesn't — read those before wiring this into anything that touches real money.

Design notes

Why there's no "non-bypassable system prompt." An earlier draft of this project considered shipping a system-prompt instruction block claiming to make spend gating "non-bypassable." That claim doesn't hold up, for two reasons:

1. MCP servers don't control the host's system prompt. Claude Desktop, Claude Code, Cursor, and other hosts each own their own system prompt. An MCP server provides tools and, in some cases, a limited instructions field — it cannot inject a binding, universally-enforced rule into the conversation.
2. A system prompt is a request to the model, not a code-level constraint. Even where a host does surface server instructions, an agent can still be prompt-injected, jailbroken, or simply routed through a different tool entirely that doesn't call this server at all. Claiming otherwise would be exactly the kind of gap this project exists to close — trusting the model to self-police is the failure mode Valta's spend gate was built to avoid in the first place.

What actually provides a guarantee: spend that is authorized by calling valta_request_spend is genuinely checked against real limits, server-side, independent of the calling model's behavior. The gap is spend that happens through some other channel the agent has access to. Closing that gap requires the payment-triggering action itself to be gated — either by making it a Valta-provided tool, or by having whatever executes the real payment call Valta's API internally before proceeding. This server is the piece that makes gated spend possible; it is not a universal guarantee that ungated spend is impossible.

License

MIT.