REA
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
Agent-friendly ComfyUI workflow skills for OpenClaw, Hermes Agent, Codex, and Claude Code; complementary to Comfy official local MCP.
作者提供的原文
如果你是作者,可把这段代码贴到 README。徽章会链接回本页。
[](https://mcp.popnx.com/en/skills/huangyuchuh-comfyui-skills-openclaw)在同一工作中把 AI 连到服务的 MCP
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
Codebase intelligence for TypeScript and JavaScript, exposed as structured MCP analysis tools.
MCP server for terminal commands, file operations, and process management
同类工作,排序与 Skill 列表相同
| 本项 ComfyUI_Skills_OpenClaw | rea | andrej-karpathy-skills | graphify | MoneyPrinterTurbo | |
|---|---|---|---|---|---|
| 作者 | HuangYuChuh | morluto | multica-ai | Graphify-Labs | harry0703 |
| Skill 数 | 1 | 1 | 1 | 1 | 1 |
| GitHub 星标 | 423 | 470,165 | 234,528 | 150,695 | 144,727 |
| 最近代码更新 | 2026年8月26日 | 2026年11月20日 | 2026年4月20日 | 2026年11月20日 | 2026年11月19日 |
| 许可证 | MIT | MIT | – | Apache-2.0 | MIT |
HuangYuChuh 提供的 1 个 Skill(指令集),AI 读取后照做 · 不是 MCP server
MCP = 手
把 AI 连到外部服务或数据,例如读取 Notion 或开发票。需要有程序或服务器在运行。
Skill = 手册
教 AI 怎么做某项工作的文本文件,不连接任何服务。
有些会重叠:少数 Skill 带有调用工具的脚本,少数 MCP 只提供知识。请从你想完成的工作出发。
Claude CodeCodex
依据开发者 README 和 repo 中的文件夹
GitHub 账号 HuangYuChuh(个人),已创建 2 年
★ 423 星,目前用户较少 · 代码最近更新 2026年8月26日
有 34 个可以在电脑上运行的脚本:scripts/clean_generated_ui.sh, scripts/serve_pages.sh, scripts/update_frontend.sh, ui/run_ui.command 另外还有 30 个文件
会联网的脚本:scripts/shared/health.py, ui/open_ui.py
没有发现 hook 或写入电脑设置的操作
根据仓库 commit 723d3a7 中的文件自动检查(只读取文件,没有运行任何内容)。这不是安全保证,请自己阅读后再决定
第一步:你用哪个 AI?
选择你用的应用,本页只显示该应用的安装方法
无法上传到 claude.ai,因为描述超过 200 个字符(499 个字符)。请改用"让 AI 帮你安装"
复制这句话,粘贴到 Claude Code 的对话中
如果只用于一个项目,请粘贴到在该项目文件夹中打开的对话里 · AI 在安装前一定会先告诉你路径
AI 会先说明要做什么,每一步都会征求你的同意
重新打开 Claude Code,输入 /ComfyUI_Skills_OpenClaw,或直接提出相关任务,Claude 会自动选用
Checked against official docs 2026年11月12日 · Source: code.claude.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
ChatGPT 桌面应用内置 Codex,Codex 从 ~/.agents/skills 文件夹读取 Skill
在 ChatGPT 桌面应用中打开 Codex,把这句话粘贴到对话中
AI 在安装前一定会先告诉你路径
AI 会先说明要做什么,每一步都会征求你的同意
把 comfyui-skill-openclaw 文件夹放到 ~/.agents/skills/comfyui-skill-openclaw
~ 是你的主文件夹($HOME)
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
不能使用你自己安装的 Skill,只能使用目录中插件自带的 Skill
想在 ChatGPT 中使用这个 Skill,请按上面的方法使用 ChatGPT 桌面应用
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
复制这句话,粘贴到 Codex 的对话中 · 适用于 ChatGPT 桌面应用中的 Codex、Codex CLI 和 IDE 扩展
如果只用于一个项目,请粘贴到在该项目文件夹中打开的对话里 · AI 在安装前一定会先告诉你路径
AI 会先说明要做什么,每一步都会征求你的同意
Codex 会自动发现新安装的 Skill;如果没有出现,请重启 Codex · 使用方法:在 ChatGPT 中输入 @ 并选择 Skill · 在 Codex CLI 或 IDE 扩展中运行 /skills,或输入 $ 加 Skill 名称
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
复制这句话,粘贴到 Cursor 的对话中
如果只用于一个项目,请粘贴到在该项目文件夹中打开的对话里 · AI 在安装前一定会先告诉你路径
AI 会先说明要做什么,每一步都会征求你的同意
在 Cursor 的 Agent 聊天中输入 / 并选择 ComfyUI_Skills_OpenClaw
Checked against official docs 2026年11月12日 · Source: cursor.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
适用于其他能读取 Skill 的应用。这句话列出了 Claude Code、Codex 和 Cursor 的路径,AI 会按你的应用选择 · 复制这句话,粘贴到应用的对话中
如果只用于一个项目,请粘贴到在该项目文件夹中打开的对话里 · AI 在安装前一定会先告诉你路径
AI 会先说明要做什么,每一步都会征求你的同意
Checked against official docs 2026年11月12日 · Source: code.claude.com, learn.chatgpt.com, cursor.com
其他支持 Agent Skills 的应用(如 Gemini CLI、GitHub Copilot)使用不同文件夹,请查看该应用的指南,再把含 SKILL.md 的文件夹复制过去。
已于 2026年11月11日 对照 Claude Code、Codex 和 Cursor 官方指南核实 skill 文件夹
以下文字来自开发者的 repo,不是 PopNx MCP 的指令
显示前 8,000 个字符 · 在 GitHub 打开完整文件
---
name: comfyui-skill-openclaw
description: |
Run registered ComfyUI workflows through the fast comfyui-skill CLI, and use the official
local Comfy MCP for live template, node, model, validation, and orchestration capabilities.
**Use this Skill when:**
(1) The user requests to "generate an image", "draw a picture", or "execute a ComfyUI workflow".
(2) The user has specific stylistic, character, or scene requirements for image generation.
(3) The user asks you to import, register, sync, or configure saved ComfyUI workflows for later reuse.
license: Apache-2.0
metadata:
version: "1.1.0"
platforms: [macos, linux, windows]
prerequisites:
commands: ["comfyui-skill"]
env_vars: []
requires:
bins: ["comfyui-skill"]
cliHelp: "comfyui-skill --help"
hermes:
tags: [image-generation, comfyui, ai-art, workflow, stable-diffusion, flux]
related_skills: []
---
# ComfyUI Agent SKILL
> **Prerequisites**: Install the CLI: `pip install -U comfyui-skill-cli`. The official local Comfy MCP is optional and should be connected using the [official guide](https://docs.comfy.org/agent-tools/mcp#local-comfy-mcp-connection) when MCP capabilities are needed. All shell commands must run from this project's root directory (where this `SKILL.md` is located).
>
> [!IMPORTANT]
> **Directory Sensitivity**: The CLI reads `config.json` and `data/` from the current directory.
> You **MUST** `cd` into the project root before running any command.
> **Symptom**: `list` returns `[]` or `server status` reports not found → you are in the wrong directory.
## Quick Decision
- A matching registered workflow exists → use the **CLI Execution Flow (Step 1-4)**. This is the default for repeated, batch, and latency-sensitive work.
- The request has no matching workflow, or requires live template/model/node discovery → use the **official local Comfy MCP**.
- The user supplies a raw workflow that needs real environment validation → use MCP `validate_workflow`; use CLI `run --validate` for a registered workflow.
- The user asks to install nodes, download models, restart, update, or switch versions → use MCP only after explaining the effect and getting confirmation.
- User says "generate image / draw a picture" → first inspect registered workflows; if none fit, discover a current MCP template instead of guessing a model name.
- User says "import workflow / add workflow" → `comfyui-skill --json workflow import <path>`
- User says "img2img / use this image" → first `comfyui-skill --json upload <image>`, then execute
- User says "inpainting / mask this area" → `comfyui-skill --json upload <mask> --mask`, then execute
- User says "show previous results" → `comfyui-skill --json history list <id>`
- User says "what failed / check job status" → `comfyui-skill --json jobs list --status failed`
- User says "which server has more VRAM" → `comfyui-skill --json server stats --all`
- User says "what nodes are available" → use MCP `nodes` for live ecosystem discovery; use `comfyui-skill --json nodes list` when only the CLI inventory is needed.
- User says "dry run / test without executing" → `comfyui-skill --json run <id> --validate`
- User says "open management UI" → `python3 ./ui/open_ui.py`
## Core Concepts
- **Skill ID**: `<server_id>/<workflow_id>` (e.g., `local/txt2img`). If server is omitted, the default server is used.
- **Schema**: Each workflow has a `schema.json` that maps business parameter names (e.g., `prompt`, `seed`) to internal ComfyUI node fields. Never expose node IDs to the user.
- **Server**: One or more ComfyUI instances configured in `config.json`. Check health with `server status`.
- **Two interfaces, one Skill**: CLI is the efficient execution path for known workflows. Official MCP is the adaptive path for discovering and orchestrating the current ComfyUI ecosystem. They share the same business intent and must not submit the same job twice.
## Official Local Comfy MCP
When the Agent host exposes the official Comfy MCP tools, call those tools directly. Start every MCP flow with `server_info` and use the live tool schemas instead of assuming parameters or tool availability.
When the host does not expose MCP tools but can run shell commands, use the included stdio bridge:
```bash
python ./scripts/comfy_mcp.py probe
python ./scripts/comfy_mcp.py tools
python ./scripts/comfy_mcp.py call search_templates --arguments '{"query":"background removal"}'
```
The bridge reads `mcp_command`, `mcp_args`, `mcp_python`, `mcp_cwd`, and `mcp_env` from the selected server in `config.json`. `probe` and `tools` are read-only. A `call` has the side effects of the named official tool.
Read [references/comfy-mcp.md](./references/comfy-mcp.md) when the request needs MCP discovery, raw-workflow validation, template adaptation, job management, installation, lifecycle control, or partner models.
Routing invariants:
- Prefer CLI for a known `<server_id>/<workflow_id>`; MCP is not a mandatory extra hop.
- Prefer MCP when the answer depends on the currently installed templates, nodes, models, hardware, or official tool contract.
- For long MCP runs: `run_workflow(wait=false)` or `run_template(wait=false)` -> `job(action="wait"|"status")` -> `fetch_outputs`.
- Once either path returns a `prompt_id`, continue polling that path. Never fall back by submitting again.
- Never set `confirm_spend=true` without the user's explicit approval for that paid call.
## Command Reference
| Command | Purpose |
|---------|---------|
| `comfyui-skill --json server status` | Check if ComfyUI server is online |
| `comfyui-skill --json server stats` | Show VRAM, RAM, GPU, versions (`--all` for multi-server) |
| `comfyui-skill --json list` | List all available workflows and parameters |
| `comfyui-skill --json info <id>` | Show workflow details and parameter schema |
| `comfyui-skill --json submit <id> --args '{...}'` | Submit a workflow (non-blocking) |
| `comfyui-skill --json status <prompt_id>` | Check execution status |
| `comfyui-skill --json run <id> --args '{...}'` | Execute a workflow (blocking, real-time streaming) |
| `comfyui-skill --json run <id> --validate` | Validate workflow without executing |
| `comfyui-skill --json upload <path>` | Upload image to ComfyUI (for img2img workflows) |
| `comfyui-skill --json upload <path> --mask` | Upload mask image (for inpainting workflows) |
| `comfyui-skill --json nodes list` | List all available ComfyUI nodes |
| `comfyui-skill --json jobs list` | List server-side job history (`--status failed` to filter) |
| `comfyui-skill --json deps check <id>` | Check missing dependencies |
| `comfyui-skill --json deps install <id> --repos '[...]'` | Install missing custom nodes |
| `comfyui-skill --json workflow import <path>` | Import workflow (auto-detect, warns about deprecated nodes) |
| `comfyui-skill --json history list <id>` | List execution history for a workflow |
---
## Execution Flow
### Step 1: Query Available Workflows
```bash
comfyui-skill --json list
```
Returns a JSON array of all enabled workflows with their parameters.
- `required: true` parameters → **ask the user** if not provided.
- `required: false` parameters → infer from context (e.g., `seed` = random number), or omit.
- Never expose node IDs; only use business parameter names (e.g., prompt, style).
- If multiple workflows match, pick the most relevant one or list candidates.
### Step 2: Parameter Assembly
Assemble parameters into a JSON string. Example:
```
{"prompt": "A beautiful landscape, high quality, masterpiece", "seed": 40128491}
```
If critical parameters are missing, ask the user (e.g., "What visual style would you like?").
### Step 3: Pre-flight Dependency Check
**Always** run before first execution of a workflow:
```bash
comfyui-skill --json deps check <server_id>/<workflow_id>
```
- If `is_ready` is `true` → proceed to Step 4.
- If `is_ready` is `false`:
1. Present missing nodes and models to the user.
2. If user agrees to install, run:
```提问、分享技巧或报告无法使用。按“有帮助”数量排序。