# Agenda Intelligence MD

Checks the evidence behind AI output, including references, quotes, numbers, and gaps. Useful for review, RAG checks, and CI.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 8
- Author: vassiliylakhonin
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/vassiliylakhonin/agenda-intelligence-md
- Updated: 2026-10-24
- HTML: https://affpapa.org/en/nevibecode/mcp/agenda-intelligence-md

## What it does

Agenda Intelligence MD inspects evidence packets attached to claims. It resolves source references, compares declared quotes with supplied text, and applies deterministic checks to wording, numbers, and negation. The result includes packet status, evidence gaps, and suggested reviewer actions.

The project exposes a CLI, a Python service, and a local MCP interface with 32 tools. You can connect it to Claude Code over HTTP or launch it through `uvx`; the Python package requires Python 3.9 or later.

The check only uses the material you provide. It does not retrieve missing sources, verify source authenticity, or establish factual truth, and lexical matching can miss paraphrases. A human still needs to review the result before consequential action.

## Requirements
- Needs: Python and uv (commands run via uvx)
- Remote server: `https://agendamcp.dpdns.org/mcp/agent`

## Install

### Claude Code (native platform)

1. **Platform CLI: remote server**
```bash
claude mcp add --transport http agenda-intelligence-md https://agendamcp.dpdns.org/mcp/agent
```

2. **Platform CLI**
```bash
claude mcp add agenda-intelligence-md -- uvx agenda-intelligence-md
```
uvx runs the command named after the package; if the README names another one, use uvx --from agenda-intelligence-md <command>

### Codex

1. **Platform CLI: remote server**
```bash
codex mcp add agenda-intelligence-md --url https://agendamcp.dpdns.org/mcp/agent
```

2. **Platform CLI**
```bash
codex mcp add agenda-intelligence-md -- uvx agenda-intelligence-md
```
uvx runs the command named after the package; if the README names another one, use uvx --from agenda-intelligence-md <command>

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "agenda-intelligence-md": {
            "url": "https://agendamcp.dpdns.org/mcp/agent"
        }
    }
}
```
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "agenda-intelligence-md": {
            "command": "uvx",
            "args": [
                "agenda-intelligence-md"
            ]
        }
    }
}
```
uvx runs the command named after the package; if the README names another one, use uvx --from agenda-intelligence-md <command>

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "agenda-intelligence-md": {
            "type": "http",
            "url": "https://agendamcp.dpdns.org/mcp/agent"
        }
    }
}
```
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "agenda-intelligence-md": {
            "type": "stdio",
            "command": "uvx",
            "args": [
                "agenda-intelligence-md"
            ]
        }
    }
}
```
uvx runs the command named after the package; if the README names another one, use uvx --from agenda-intelligence-md <command>

### Gemini CLI

1. **Platform CLI: remote server**
```bash
gemini mcp add -t http agenda-intelligence-md https://agendamcp.dpdns.org/mcp/agent
```

2. **Platform CLI**
```bash
gemini mcp add agenda-intelligence-md uvx agenda-intelligence-md
```
uvx runs the command named after the package; if the README names another one, use uvx --from agenda-intelligence-md <command>

## When to use it
- resolve claim references against supplied sources
- compare quoted text with its source
- spot numbers without matching support
- flag incomplete evidence packets
- route a RAG answer toward revision

## Example prompts
- Inspect this claims-and-sources packet and report broken references or unsupported numbers.
- Review the RAG answer with its inline citations and list the fixes a reviewer should make.
- Compare every declared quote with the supplied source text, then prepare the findings for human review.

## Similar
- [gws-calendar-agenda](https://affpapa.org/en/nevibecode/skills/gws-calendar-agenda.md) — skill · 49600 installs
- [foundation-meeting-agenda](https://affpapa.org/en/nevibecode/skills/foundation-meeting-agenda.md) — skill · 703 installs
- [competitive-intelligence](https://affpapa.org/en/nevibecode/skills/competitive-intelligence.md) — skill · 4500 installs
- [notion-meeting-intelligence](https://affpapa.org/en/nevibecode/skills/openai-notion-meeting-intelligence.md) — skill · 3100 installs
- [competitive-intelligence-analyst](https://affpapa.org/en/nevibecode/agents/competitive-intelligence-analyst.md) — subagent · 1561 installs
- [business-intelligence](https://affpapa.org/en/nevibecode/skills/business-intelligence.md) — skill · 1500 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