← 返回 Skills 市场
219
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install photo-retouching
功能描述
Execute RunningHub ComfyUI workflows via API. Use when you need to run ComfyUI workflows on RunningHub cloud platform, submit tasks, query status, and retrie...
安全使用建议
What to consider before installing:
- The skill needs your RunningHub API key (32-character) though the registry metadata didn't declare it; the scripts look for RUNNINGHUB_API_KEY or a config.json saved in the skill workspace. Expect the key to be saved in plaintext to config.json unless you change the code or location.
- Using the skill will upload local images to RunningHub (external service). Only upload images you are comfortable sending to that service.
- The chrome_automation script attempts to control a local Chrome DevTools instance (127.0.0.1:9222) using a raw socket/WebSocket handshake. That requires launching Chrome with remote debugging enabled and grants the script the ability to interact with your browser — run that only in a controlled environment and inspect the code first.
- There is no install manifest or dependency list; you should ensure Python and required packages (e.g., requests) are installed in an environment you control. Consider running the skill in an isolated environment or container if you are unsure.
- If you plan to use this skill, review and (if desired) modify how the API key is stored (avoid long-lived plaintext storage), inspect the code yourself (upload/submit paths), and prefer providing the API key at runtime rather than saving it permanently.
- If you want higher assurance, ask the publisher for: (1) declared required env vars in the registry metadata, (2) a requirements.txt or install script, and (3) clarification on the Chrome automation use-case or remove that component if you don't want local browser control.
功能分析
Type: OpenClaw Skill
Name: photo-retouching
Version: 1.0.0
The skill bundle is a legitimate tool for executing ComfyUI workflows on the RunningHub cloud platform. It includes a standard API client (runninghub_client.py), batch processing capabilities (batch_runner.py), and a helper script for Chrome DevTools Protocol (CDP) automation (chrome_automation.py) to assist with web-based tasks. While the CDP script uses low-level socket communication to avoid dependencies, its logic is restricted to probing the RunningHub website for UI elements and lacks any malicious data exfiltration or unauthorized access capabilities. The instructions in SKILL.md are consistent with the stated purpose, and the inclusion of a referral link is a common non-malicious practice.
能力评估
Purpose & Capability
Name/description align with the included Python client and helper scripts: upload images, submit workflows, poll/query tasks, batch runs, and a hybrid/Chrome automation flow. All code relates to RunningHub workflow execution and image handling — no unrelated cloud providers or credentials are requested.
Instruction Scope
SKILL.md and scripts instruct the agent to upload local images and send them to RunningHub (expected for this skill). The Chrome automation script attempts to open/control a local Chrome DevTools endpoint (127.0.0.1:9222) and execute JS via a raw WebSocket handshake — this is outside typical simple API usage and can interact with the user's local browser (requires Chrome launched with remote debugging). The instructions also encourage saving the API key to the skill workspace config file, meaning secrets will be stored on disk in plaintext unless the user changes that behavior.
Install Mechanism
No install spec is provided (instruction-only install), but multiple Python scripts are included and they implicitly require Python and the 'requests' library. The lack of declared dependencies or an install step is an omission (user must ensure Python and dependencies are present). No external downloads or archive extraction occur.
Credentials
Metadata declares no required environment variables, yet the code and SKILL.md clearly expect an API key (RUNNINGHUB_API_KEY or saved config.json). The skill will store the API key in a local config.json file (plain text) by default. These are proportional to the stated purpose but the registry metadata omission is an inconsistency and a usability/security risk (plaintext credentials). The scripts do not request unrelated credentials.
Persistence & Privilege
The skill does not request always:true and does not modify other skills. It does persist the provided API key into a config.json inside the skill workspace by default (normal for CLI tools but means a secret is kept on disk). The Chrome automation requires access to a local Chrome DevTools instance (privileged local control of the browser if enabled).
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install photo-retouching - 安装完成后,直接呼叫该 Skill 的名称或使用
/photo-retouching触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Initial release of RunningHub ComfyUI workflow runner skill.
- Run ComfyUI workflows on the RunningHub cloud via API.
- Supports uploading images as workflow inputs.
- Includes CLI and Python API usage instructions.
- Enables submitting tasks, querying status, and downloading results.
- Provides troubleshooting tips and API endpoint references.
元数据
常见问题
Photo Retouching 是什么?
Execute RunningHub ComfyUI workflows via API. Use when you need to run ComfyUI workflows on RunningHub cloud platform, submit tasks, query status, and retrie... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 219 次。
如何安装 Photo Retouching?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install photo-retouching」即可一键安装,无需额外配置。
Photo Retouching 是免费的吗?
是的,Photo Retouching 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Photo Retouching 支持哪些平台?
Photo Retouching 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 Photo Retouching?
由 uiueux(@uiueux)开发并维护,当前版本 v1.0.0。
推荐 Skills