# Analyze Images with GLM-4V

Image review often needs structured visual reasoning, not only a short caption. This skill guides GLM-4V prompts for descriptions, visual questions, OCR, and detail extraction.

## Install

```bash
npx skillstore add ck991357/glm4v-analyze-image
```

## Metadata

- Status: approved
- Slug: ck991357-glm4v-analyze-image
- Version: 1
- Author version: 1
- Skillstore revision: r1
- Version status: invalid
- Tree hash: 4d545e3197ec2649e57bed6d98a4890494a4ce4b9a4ce9047371a1601144748d
- Author: CK991357
- GitHub username: CK991357
- License: MIT
- Repository: https://github.com/CK991357/gemini-chat/tree/main/src/skills/glm4v\_analyze\_image
- Ref: 02f077c174c5335e2f5d02ca15e77b70d9543e58
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network
- Quality score: 79
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/ck991357-glm4v-analyze-image
- Manifest: https://skillstore.pages.dev/api/skills/ck991357-glm4v-analyze-image/manifest

## Capabilities

- Describes visible objects, scenes, actions, and spatial relationships in an image.
- Answers targeted questions about people, objects, counts, and visible activity.
- Extracts readable text from images when the text is clear enough for OCR.
- Analyzes colors, shapes, layout, atmosphere, and technical visual details.
- Shows the expected GLM-4V tool parameters for image URL and prompt input.

## Use Cases

- Product Image Review: Summarize product photos, identify visible attributes, and flag missing details for catalog preparation.
- Document Screenshot Reading: Extract visible text and layout details from screenshots for faster triage and note taking.
- Research Image Inspection: Ask focused visual questions about diagrams, scenes, or reference images during research workflows.

## Prompt Templates

### Basic Image Description

```
Describe this image in simple terms. Include the main subject, setting, visible objects, and overall mood.
```

### Focused Visual Question

```
Answer this question using only visible evidence from the image: [question]. State when the image is unclear.
```

### OCR and Layout Check

```
Read any visible text in the image. Then describe the layout, headings, tables, labels, and important visual hierarchy.
```

### Detailed Inspection Report

```
Analyze the image step by step. Cover objects, relationships, text, colors, technical details, uncertainty, and follow-up checks.
```

## Limitations

- Requires an accessible image URL for the model call.
- Does not support real-time video streams or 3D model analysis.
- Cannot reliably analyze unclear, low-resolution, or blocked images.
- Sensitive or inappropriate images may be rejected by model safety rules.

## Best Practices

- Use a direct image URL that the model can access without authentication.
- Ask specific questions when accuracy matters for counts, text, or small details.
- Request uncertainty notes for ambiguous areas or low-quality images.

## Anti Patterns

- Do not submit private or sensitive images through public URLs.
- Do not rely on the model for identity, medical, legal, or safety-critical conclusions.
- Do not use vague prompts when the task requires exact fields or focused evidence.

## Security Audit

- Audited at: 2026-07-05T06:39:02.17\+00:00
- Summary: Static command findings are false positives caused by Markdown code fences and inline backticks in SKILL.md. The URL findings are placeholder image examples, and the entropy finding is explained by readable multilingual Markdown. No prompt injection, hidden command execution, or malicious data-exfiltration intent was found.

## Stats

- Views: 234
- Downloads: 25
- Favorites: 0
- Popularity score: 0