WorkBuddy Connector MCP server for ComfyUI (stdio, auth_mode: token)
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"workbuddy-mcp-comfyui","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# ComfyUI Connector(WorkBuddy) 为腾讯 WorkBuddy 生态开发的 ComfyUI 连接器:MCP stdio + 用户自填 Token 模式 (规范 06 / 06D 附录C),让 AI 通过自然语言运行 ComfyUI 工作流并取回生成的图片。 ## 目录结构 ``` comfyui/ ├── connector-meta.json # 连接器元信息(type: mcp, auth_mode: token, minWorkbuddyVersion: 4.24.0) ├── mcp.json # MCP Server 连接配置(stdio,凭证经 env 注入) ├── token-schema.json # 用户自填 Token 表单(COMFY_URL + 可选 COMFY_TOKEN) ├── comfy_mcp_server/ # MCP Server(Python,官方 mcp SDK FastMCP) │ ├── __init__.py │ ├── __main__.py # python -m comfy_mcp_server 入口 │ └── server.py # 5 个 tool:list_checkpoints / list_samplers / run_workflow / get_status / get_image ├── skills/ │ └── SKILL.md # Skill 文件(按 05 规范,教 AI 如何使用工具) ├── icon.svg # 连接器图标 ├── tests_smoke.py # 冒烟测试(MCP stdio 客户端调工具,验证错误路径) ├── pyproject.toml # hatchling 打包(PyPI 包名 workbuddy-mcp-comfyui) └── README.md ``` 依赖仅 `mcp`(v1,FastMCP)+ `httpx`(Python ≥ 3.10)。 ## 工具一览 | Tool | 用途 | |---|---| | `list_checkpoints()` | 列出 checkpoint 模型(GET /object_info/CheckpointLoaderSimple) | | `list_samplers()` | 列出采样器与调度器(GET /object_info/KSampler) | | `run_workflow(workflow)` | 提交 API 格式工作流(POST /prompt),返回 prompt_id,只提交不等待 | | `get_status(prompt_id)` | 查询进度与输出(GET /history/{id},未完成时对照 GET /queue) | | `get_image(filename, subfolder?, type?)` | 取回图片(GET /view),以 MCP ImageContent 返回 | 标准链路:`list_checkpoints` 确认模型名 → `run_workflow` 提交 → 轮询 `get_status` → 用 history outputs 里的 filename/subfolder/type 三元组调 `get_image`。 COMFY_TOKEN 仅在用户加了反向代理鉴权时才填;非空时 server 自动附加 `Authorization: Bearer` 头,原生 ComfyUI 留空。httpx 超时 30s。 ## 本地测试 ```bash cd <项目目录>/connectors/comfyui # 1. 语法与 JSON 校验 python3 -m py_compile comfy_mcp_server/*.py for f in connector-meta.json mcp.json token-schema.json; do python3 -m json.tool "$f" > /dev/null && echo "$f OK"; done # 2. 起本地构建的 server 并调用工具(ComfyUI 未运行时验证错误路径) # server 经 uvx --from 本地目录构建;客户端复用 home-assistant/.venv 里的 mcp 包 export UV_DEFAULT_INDEX=https://pypi.tuna.tsinghua.edu.cn/si…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.