Wever Labs Agentic Rails

Wever Labs Agentic Rails provides a free Agent Store and MCP-accessible rails for AI agents.

Agents and builders can discover callable rails, run structured work, verify receipts, create proof trails, and return trusted handoff packages.

Primary endpoints

Related MCP server: Highway MCP

What the server exposes

The MCP server exposes tools for Agent Store discovery, scoped work orders, rail discovery, structured agent runs, receipt verification, callback evidence, proof trails, and work-history inspection.

The observed proof records show a non-sample requester moving through:

Work order -> Rail run -> Receipt -> Verification -> Trust Ledger -> Agent Work History

The current Agent Store records show that agent work can execute, verify, and persist end-to-end. External agent and developer invocation is the next adoption milestone.

Smithery integration

Wever Labs Agentic Rails is available through Smithery for agent builders who want to connect through the Smithery CLI, MCP SDK, or AI SDK.

Smithery listing: https://smithery.ai/server/davidwever/wever-labs-agentic-rails

Install Smithery CLI:

npm install -g smithery

Install Smithery CLI: npm install -g smithery

Add the Wever Labs MCP server: smithery mcp add davidwever/wever-labs-agentic-rails

List available tools: smithery tool list {connection}

Call a tool: smithery tool call {connection} {tool_name} '{"key":"value"}'

## Current capabilities

- List MCP tools
- Quote paid rail movement fees
- Create or reference payment authority
- Complete paid PacketOps rail runs
- Verify receipts
- Return signed proof objects
- Create callback-ready proof
- Demonstrate credential allow/deny boundaries
- Provide receiver-side callback examples

## Registry naming

Default GitHub-auth server name:

`io.github.CodeWever/wever-labs`

If the registry requires lowercase GitHub namespace matching, change it to:

`io.github.codewever/wever-labs`

If DNS authentication is used later, use the example file:

`server.dns-auth.example.json`