← 返回 Skills 市场
levdavid1

Don't download

作者 Levdavid1 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
233
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install test-skill-1
功能描述
Remove image background to transparent PNG. Powered by RMBG-2.0, commercially-safe model. Extract subjects for overlays, product photography, logos, and cuto...
安全使用建议
This skill appears to do what it says (remove image backgrounds), but take these precautions before installing or using it: - Confirm the auth server and API endpoints (metadata lists https://auth.bria.ai and engine.prod.bria-api.com). If the sources/homepage are unknown, verify them independently. - Be aware it will create and store a token at ~/.bria/credentials. If you don't want long-lived credentials on disk, do not authenticate or remove the file after use. - The skill tells you to upload local images to an unspecified hosting service. That step could leak private images if you choose an untrusted host — decide which host to use or refuse to upload sensitive files. - It requires network access to the Bria endpoints and will poll the auth server during device authorization; review network policies or run in a sandbox if you need isolation. - If you want more assurance, request the skill author/publisher, a homepage or repository link, and a declaration of required config paths/env vars. If those are missing or the author is unknown, treat the skill more cautiously.
功能分析
Type: OpenClaw Skill Name: test-skill-1 Version: 1.0.0 The skill bundle provides instructions for an AI agent to perform image background removal using the Bria AI service. It implements a standard OAuth2 device authorization flow, manages credentials locally in `~/.bria/credentials`, and interacts with legitimate Bria AI endpoints (auth.bria.ai and engine.prod.bria-api.com). The code logic in SKILL.md is transparent and aligns entirely with the stated purpose without evidence of malicious intent or data exfiltration.
能力评估
Purpose & Capability
The skill claims to remove image backgrounds and its runtime instructions call the Bria image API — that is coherent. However the skill accesses a local credentials file (~/.bria/credentials) and writes a persistent API token there even though 'required config paths' were not declared in the metadata. Declaring config paths or required env vars would be expected for a skill that stores credentials.
Instruction Scope
The SKILL.md instructs the agent to read and write ~/.bria/credentials and to perform a device-auth flow (polling the auth server). It also tells the operator to upload local images to 'a hosting service' but does not specify one — this vagueness could cause images to be uploaded to arbitrary third-party hosts, risking data exposure. The instructions also assume curl and python3 are available and include loops that poll external endpoints.
Install Mechanism
No install spec or code files are present (instruction-only), so nothing will be written to disk by an installer. This is lower risk than skills that download and install binaries.
Credentials
The skill does not declare required environment variables, yet it reads/uses BRIA_AUTH_SERVER (with a default) and expects to set/use BRIA_API_KEY (persisted to disk). It does not request unrelated credentials. The missing declaration of the credential path and env usage is a proportionality / transparency issue.
Persistence & Privilege
The skill persists an API token to ~/.bria/credentials, giving it ongoing access to the Bria API until that file is removed. The skill does not request 'always: true' or other elevated platform privileges and does not modify other skills, but persistent token storage is a lasting presence worth noting.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install test-skill-1
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /test-skill-1 触发
  4. 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 the remove-bg skill. - Enables background removal from images, returning transparent PNGs using the RMBG-2.0 model. - Includes secure, production-ready authentication via Bria device authorization (with credential file setup). - Supports both synchronous and asynchronous processing with status polling. - Easy integration for overlays, product photography, logos, and cutouts.
元数据
Slug test-skill-1
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Don't download 是什么?

Remove image background to transparent PNG. Powered by RMBG-2.0, commercially-safe model. Extract subjects for overlays, product photography, logos, and cuto... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 233 次。

如何安装 Don't download?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install test-skill-1」即可一键安装,无需额外配置。

Don't download 是免费的吗?

是的,Don't download 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Don't download 支持哪些平台?

Don't download 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 Don't download?

由 Levdavid1(@levdavid1)开发并维护,当前版本 v1.0.0。

💬 留言讨论