# Warrant

Warrant records an agent’s decisions, supporting evidence and unresolved questions. Use it to hand work to another agent with the grounds and limits intact.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 5 repo stars
- Author: s0fractal
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/s0fractal/warrant/tree/HEAD/integrations/mcp-server
- Updated: 2026-10-24
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/warrant

## What it does

Warrant stores duties, work transitions, evidence and questions that remain open. Each record shows what the agent established, what conditions it closed and what still needs examination. A receiving agent gets a compact account of the result rather than a full activity log.

Another agent can accept a specific closure item. The assessment can also show when its grounds have become stale or challenged; computations are checked by replay on the recipient’s side. A replayed basis is marked UNVERIFIED unless its provider is installed and accepted in the recipient’s profile.

The server works with Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed and Gemini CLI. The Claude Code setup command runs `uvx warrant-verify`. The project README also describes installing from a local checkout, with Python ≥ 3.10 and the `cryptography` dependency.

## Requirements
- Needs: Python and uv (commands run via uvx)
- Environment variables: `WARRANT_STORE`, `WARRANT_KEY`, `WARRANT_ACTOR` — keys and tokens: ask the human for the values, never commit them or put them in shared files.

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add --env WARRANT_STORE=<WARRANT_STORE> --env WARRANT_KEY=<WARRANT_KEY> --env WARRANT_ACTOR=<WARRANT_ACTOR> --transport stdio mcp-server -- uvx warrant-verify
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from warrant-verify <command>

### Codex

**Platform CLI**
```bash
codex mcp add mcp-server --env WARRANT_STORE=<WARRANT_STORE> --env WARRANT_KEY=<WARRANT_KEY> --env WARRANT_ACTOR=<WARRANT_ACTOR> -- uvx warrant-verify
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from warrant-verify <command>

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "mcp-server": {
            "command": "uvx",
            "args": [
                "warrant-verify"
            ],
            "env": {
                "WARRANT_STORE": "<WARRANT_STORE>",
                "WARRANT_KEY": "<WARRANT_KEY>",
                "WARRANT_ACTOR": "<WARRANT_ACTOR>"
            }
        }
    }
}
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from warrant-verify <command>

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "mcp-server": {
            "type": "stdio",
            "command": "uvx",
            "args": [
                "warrant-verify"
            ],
            "env": {
                "WARRANT_STORE": "<WARRANT_STORE>",
                "WARRANT_KEY": "<WARRANT_KEY>",
                "WARRANT_ACTOR": "<WARRANT_ACTOR>"
            }
        }
    }
}
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from warrant-verify <command>

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add -e WARRANT_STORE=<WARRANT_STORE> -e WARRANT_KEY=<WARRANT_KEY> -e WARRANT_ACTOR=<WARRANT_ACTOR> mcp-server uvx warrant-verify
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from warrant-verify <command>

## When to use it
- record the basis and outcome of an agent’s work
- hand off conditions another agent still needs to accept
- separate established results from unverified claims
- surface open questions and outstanding duties
- replay a computation on the recipient’s side

## Example prompts
- Record the empty-input check, including what it did not test.
- Show me the duties I still need to accept.
- Summarize this task’s result with its grounds and open questions.

## Similar
- [Context7](https://affpapa.org/en/nevibecode/mcp/context7.md) — MCP server · 5563 installs
- [Search Google Flights](https://affpapa.org/en/nevibecode/mcp/search-google-flights.md) — MCP server · 2595 installs
- [Math-MCP](https://affpapa.org/en/nevibecode/mcp/math-mcp.md) — MCP server · 5958 installs
- [pipeworx gateway](https://affpapa.org/en/nevibecode/mcp/pipeworx-gateway.md) — MCP server · 419019 installs
- [SparkForge](https://affpapa.org/en/nevibecode/mcp/sparkforge.md) — MCP server · 219896 installs
- [OpenWebSearch](https://affpapa.org/en/nevibecode/mcp/openwebsearch.md) — MCP server · 494 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
