← 返回 Skills 市场
拍照识万物
作者
ZhangYangGZ11
· GitHub ↗
· v2.0.0
· MIT-0
122
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install vision-identify
功能描述
万物识别智能体 - 通过图片识别物品、动植物、人物、商品、文字等一切可见事物。 当用户上传图片并询问"这是什么"、"帮我识别"、"拍照识物"、"识图"、"图片识别"、"这张图里有什么"时触发。 核心特点：不仅识别物品，更分析场景、推断意图、给出精准选购方案，防止买错。
安全使用建议
This skill is internally consistent, but before installing check: (1) Privacy — uploaded images will be sent to the platform/model (references mention Tencent Mixed-Yuan/Turbo); avoid uploading sensitive photos (faces, IDs, private documents) unless you accept that processing by the model provider. (2) Source verification — the package lists a homepage (wtmvw.com) and an author handle; verify the site/author if you require provenance. (3) Misidentification risk — image-based identification can be wrong; treat medical/legal/safety-critical inferences cautiously. (4) If you need stricter data handling, confirm whether your deployment keeps images local or sends them to external APIs before enabling the skill.
功能分析
Type: OpenClaw Skill
Name: vision-identify
Version: 2.0.0
The 'vision-identify' skill bundle consists entirely of documentation and prompt instructions (SKILL.md, README.md, _meta.json) designed to guide an AI agent in performing image recognition and providing shopping advice. There is no executable code, no evidence of data exfiltration, and no malicious prompt injection intended to hijack the agent's behavior for harmful purposes.
能力评估
Purpose & Capability
Name/description (image recognition + scene analysis + shopping advice) match the SKILL.md content. The skill does not request unrelated credentials, binaries, or config paths; all declared functionality is achievable with image input and model outputs.
Instruction Scope
SKILL.md gives detailed, bounded instructions on how to analyze images, infer intent, and produce shopping recommendations in specific formats. It assumes the platform will forward uploaded images to a multimodal model (references mention WorkBuddy/Tencent Turbo). This is expected for a vision skill, but users should note images will be included in model context (privacy consideration).
Install Mechanism
No install spec and no code files — instruction-only skill. Nothing is downloaded or written to disk as part of the skill package itself, which reduces install-time risk.
Credentials
The skill requires no environment variables, no credentials, and no configuration paths. That is proportional to its stated purpose (uses platform model and image input only).
Persistence & Privilege
Default privileges (not always:true, user-invocable, model invocation allowed) are appropriate. The skill does not request persistent system-level presence or modify other skills or system configs.
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 vision-identify - 安装完成后，直接呼叫该 Skill 的名称或使用
/vision-identify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
v2.0.0
Initial release v2.0.0:
- 万物识别：动植物、商品、五金配件、电子产品、场景地点等
- 场景分析：观察物品状态，推断用户真实意图
- 精准选购：关联推荐防买错，线上线下双渠道
- 兼容提醒：容易买错的地方用 ⚠️ 标注
元数据
常见问题
拍照识万物 是什么？
万物识别智能体 - 通过图片识别物品、动植物、人物、商品、文字等一切可见事物。 当用户上传图片并询问"这是什么"、"帮我识别"、"拍照识物"、"识图"、"图片识别"、"这张图里有什么"时触发。 核心特点：不仅识别物品，更分析场景、推断意图、给出精准选购方案，防止买错。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 122 次。
如何安装 拍照识万物？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install vision-identify」即可一键安装，无需额外配置。
拍照识万物 是免费的吗？
是的，拍照识万物 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
拍照识万物 支持哪些平台？
拍照识万物 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 拍照识万物？
由 ZhangYangGZ11（@zhangyanggz11）开发并维护，当前版本 v2.0.0。
推荐 Skills