mcpField
A
AI-SLOT Management

AI-SLOT Management

Local
github.com/iannilDeveloper Tools0Updated 2d ago
Repository

About AI-SLOT Management

An experimental Node.js package that exposes AI-SLOT management operations to an Agent through local stdio. It requires a separately started local management service, and publication still requires Owner approval.

Key features

  1. 01Seven management tools over stdio
  2. 02Discover and read slots
  3. 03Create and preview drafts
  4. 04Publish and roll back content
  5. 05Check status

Use cases

  1. 01Discover available slots
  2. 02Prepare and preview content drafts
  3. 03Publish Owner-approved artifacts
  4. 04Roll back a publication and check status

MCP support

Exposes management tools through MCP over local stdio and forwards requests to the fixed service at `http://127.0.0.1:4194`. It does not start the service or provide an approval tool.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "iannil-ai-slot-management": {
      "command": "npx",
      "args": [
        "-y",
        "@ai-slot/mcp"
      ],
      "env": {
        "AI_SLOT_AGENT_TOKEN": "<AI_SLOT_AGENT_TOKEN>"
      }
    }
  }
}
Run commandnpx -y @ai-slot/mcp
Environment variables to set
AI_SLOT_AGENT_TOKENSecret本地管理服务配置的 Agent 令牌；不得使用 Owner 令牌。

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@ai-slot/mcp0.1.3

Tools

7

Saved tool list · updated 1w ago

This tool takes no parameters.

Details

From the project README

@ai-slot/mcp

English | 简体中文

An experimental AI-SLOT management MCP package requiring Node.js 22 or later. It exposes tools over local stdio and forwards requests to a fixed management service at http://127.0.0.1:4194. It does not start that service, open the version store, or grant approval permissions. Compatibility across versions is not yet guaranteed for the 0.x interface.

Installation and usage

npm install @ai-slot/mcp

First, start the service by following the repository's management reference service instructions. Set the AI_SLOT_AGENT_TOKEN environment variable to the token configured for the Agent by the service. Do not give the Owner token to the Agent or commit real tokens to a configuration repository.

The command in your client configuration can point to node_modules/.bin/ai-slot-management-mcp in the project where the package is installed. From that project's terminal, you can run:

npx --no-install ai-slot-management-mcp

The process communicates using MCP over standard input and output; no manual input is required. The programmatic API exports createManagementClient and createManagementServer.

Capabilities and limitations

Seven tools support slot discovery, reading, draft creation, preview, publication, rollback, and status checks. There is no approval tool. The Agent can submit the corresponding artifact only after the Owner approves it on the reference management page. Retrieve the actual tool list through MCP tools/list.

The initial release supports only the fixed local HTTP service. Remote MCP, arbitrary target URLs, multiple tenants, and automatic approval by AI are not supported. The published package does not include the reference management page, CMS plugins, or service configuration; these are in the repository. Validation with the official SDK, real stdio, HTTP, and the complete browser workflow has passed. External AI clients and real customer pilots still await validation.

MIT license. Report issues at https://github.com/iannil/ai-slot-component/issues.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.