lov-describe-image SKILL.md skillA
lov-describe-image from lovstudio/skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add lovstudio/skills/lov-describe-image@v0.2.1Pins v0.2.1 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- lovstudio/skills
- Path
skills/describe-image/SKILL.md- Versions recorded
- 4
- Latest version
- v0.2.1, committed Sep 25, 2026
- First version recorded
- Aug 20, 2026
- Size
- 3,371 bytes
- Repository stars
- 72
- License
- MIT
给纯文本模型"看图"：调用智谱 GLM-4V-Flash（免费）把图片转成文字描述。当用户要求描述/识别/读取图片，或任务里出现图片文件（截图/图表/照片）而当前模型无法直接看图时，优先调用本 skill。需要图片磁盘路径，可附带具体问题。.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# 看图说话 · Image Narrator 给纯文本模型"看图"：把图片发送给视觉模型（智谱 GLM-4V-Flash，免费），返回文字描述，让本身不具备视觉能力的模型（如 deepseek-v4-flash）能间接"看到"图片内容。 ## Triggers ### Activate when - 用户要求"看看/描述/识别/读懂/读取/识别这张图"，或询问图片里写了什么。 - 任务上下文里出现图片文件（截图、图表、照片、海报、含文字的图），而当前模型无法直接看图。 - 需要从含文字的图片（界面截图、文档扫描、图表）中提取具体信息，例如检查异常、转录文字、核对内容。 ### Do not activate when - 用户只是要求生成、编辑、压缩或转换图片，而不是理解图片内容。 - 当前模型本身具备视觉能力、能直接看到图片，无需转述。 - 用户明确要求走 dsh 的原生多模态管线、项目内置读图工具，或指定了其他视觉后端。 ## Usage ```bash python3 ~/.claude/skills/lov-describe-image/scripts/describe_image.py "<图片路径>" "[具体问题]" ``` - `<图片路径>` 必填：绝对路径，或相对当前工作区的路径（建议绝对路径，避免歧义）。 - `[具体问题]` 可选：想从图中知道的具体信息；省略则让视觉模型完整描述（物体/场景/布局/颜色/风格/所有文字）。 ## Behavior - 读取文件 → 超 5MB 自动用 `sips` 降采样到 2048px → base64 编码 → 调 GLM-4V-Flash（`max_tokens` 上限 1024）。 - 返回的文本就是视觉模型的回答，直接向用户转述，不要添油加醋，也不要声称"看到了原图"。 - 失败时（未设 key / 网络 / 尺寸超限）会打印明确错误，按提示处理即可。 - 小模型对细节核对（逐字转写）置信度有限；重要结论建议用更强的视觉模型（如 Gemini）交叉验证。 ## Prerequisites - 环境变量 `ZHIPU_API_KEY`（智谱 key，格式 `id.secret`，注册地址：https://bigmodel.cn/apikey/platform）。 - macOS 自带 `sips`（大图降采样）与 `python3`。 ## 通用反馈闭环 用户在 Skill 驱动任务中提出修改意见时，继续当前产物前必须执行： 1. 先判断意见是 `task-specific`（仅本次）还是 `reusable`（可跨任务复用）。 2. `task-specific` 只修改当前任务，不改 Skill。 …
Read the whole file at its exact version.
How to install
mdr add lovstudio/skills/lov-describe-image@sha256:4f7c677899e7b9a8The line at the top of this page pins the label v0.2.1, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_giuhos7ccox6xvic)
2 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 4 versions of this file, between Aug 20, 2026 and Sep 25, 2026, about one every 12 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| v0.2.1 latest | 2026-09-25 | 4b67fd1 | 3,371 B | A | view · diff |
| v0.2.0 | 2026-09-07 | a09848e | 3,330 B | A | view · diff |
| v0.2.0 | 2026-08-23 | 00fafa8 | 3,318 B | A | view · diff |
| v0.1.0 | 2026-08-20 | 8b4d13a | 2,596 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (3371 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
lovstudio/skills · 72 stars · license MIT · pushed 2026-11-06 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_giuhos7ccox6xvic GET https://markdownregistry.com/api/v1/resolve?ref=lovstudio/skills/lov-describe-image GET https://markdownregistry.com/api/v1/blob/4f7c677899e7b9a8282aef7c88d970d1b0d2d5b8f90e06b971a3ce4db90dae9e
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.