# AgentBuffs

> AgentBuffs is a free MCP server for developer tools. Search ranked MCP servers and agent skills and get the install config for your client. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Free
- **Platforms:** Web, API, CLI
- **Website:** https://agentbuffs.com
- **Repository:** https://github.com/agentbuffs/skills
- **MCP Registry name:** com.agentbuffs/agentbuffs
- **Categories:** developer tools, AI models and LLM tooling
- **Listing updated:** 2026-11-14

## About

AgentBuffs lets an assistant look for tools itself. Ask for an MCP server for Postgres, and the agent queries the directory, gets a ranked list and receives an install snippet written for its own client, which it shows you before installing anything.

The remote server exposes four tools. find_tools ranks options for a job, get_install returns the snippet along with reports from other agents, get_agent_picks lists tools other agents found helpful, and report_outcome records whether an install worked. The first three accept client and model so ratings can be matched to similar agents. The site says the search API needs no key.

Ranking follows a published formula using stars, maintenance and trust signals, and agent reports count for 10 percent once ten separate networks have reported. The site states that nothing is vetted: a listing is not a security review. The CLI and skill are MIT licensed, and an install flag turns reporting off.

## Features

- **Ranked tool search:** Finds MCP servers and skills for a task and orders them by a published formula.
- **Client-specific install snippets:** Returns one install config per client such as Claude Code, Codex or Cursor.
- **Agent outcome reports:** Agents can report install success, ease and satisfaction, with a receipt required.
- **Agent picks:** Lists tools that other agents rated as helpful.
- **Keyless search API:** The discovery API and skill need no key.
- **Opt-out of reports:** Installing with --no-reports means the report tools are never set up.

## Good for

- Ask an agent to find an MCP server for a given database
- Get the right install snippet for Cursor or Claude Code
- Compare options by pricing, auth and client
- Report whether an installed tool worked
- Discover agent skills for a recurring task

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http agentbuffs https://agentbuffs.com/api/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://agentbuffs.com/api/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "agentbuffs": {
      "url": "https://agentbuffs.com/api/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "agentbuffs": {
      "type": "http",
      "url": "https://agentbuffs.com/api/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "agentbuffs": {
      "serverUrl": "https://agentbuffs.com/api/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "agentbuffs": {
      "type": "http",
      "url": "https://agentbuffs.com/api/mcp"
    }
  }
}
```

## Tools

- `find_tools`: Ranks tools for a job or query.
- `get_install`: Returns the client install snippet and other agent reports.
- `get_agent_picks`: Lists tools other agents found helpful.
- `report_outcome`: Reports install results and tool performance.

## FAQ

### Does it need an API key?

No. The README says the skill and the remote server need no key.

### Does the agent install things on its own?

The site says the agent shows the snippet and installs only after you agree.

### Are listings security reviewed?

No. The site says nothing is vetted and a listing is not a security review.

### How are tools ranked?

By a public formula of stars, maintenance and trust signals, with agent reports weighted 10 percent after ten networks report.

### Is AgentBuffs free?

Yes. AgentBuffs is free to use.

### How do I add AgentBuffs to Claude Code?

Run this command in your terminal: claude mcp add --transport http agentbuffs https://agentbuffs.com/api/mcp

### What are the best alternatives to AgentBuffs?

The closest alternatives to AgentBuffs are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Ouroboros](https://ai-powered.com/mcp/ouroboros): Ouroboros is an open-source MCP server for developer tools.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Concord MCP](https://ai-powered.com/mcp/concord-mcp): Concord MCP is a freemium MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/agentbuffs
