multimodal-understanding
Skill0 starsChanged 7 months agoMITView on GitHub
Analyze images and visual content using the NoDesk AI Gateway multimodal API. ALWAYS use this skill when the user mentions an image file path, image URL, or asks to analyze/describe any image. Do NOT use the built-in image tool — always route through this skill for better accuracy.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- jinsi-ai/agent-skillsword for word, first seen Apr 3, 2026
Other versions1 project uses a version of this.
Word for word · 1
- jinsi-ai/agent-skills★ 0100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository33 other files it gives its agents.
- algorithmic-artSkill
- amap-mapsSkill
- bilibili-mcpSkill
- brand-guidelinesSkill
- canvas-designSkill
- agent-browserSkill
- docxSkill
- pptxSkill
- xlsxSkill
- doc-coauthoringSkill
- docxSkill
- excalidraw-diagram-generatorSkill
- feishuSkill
- frontend-designSkill
- image-generationSkill
- internal-commsSkill
- mcp-builderSkill
- multi-search-cnSkill
- pdfSkill
- pdfSkill
- planning-with-filesSkill
- pptxSkill
- read-wechat-articleSkill
- research-proposalSkill
- skill-creatorSkill
- slack-gif-creatorSkill
- smart-web-scraperSkill
- theme-factorySkill
- video-generationSkill
- webapp-testingSkill
- web-artifacts-builderSkill
- wechat-mp-writerSkill
- xlsxSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 17
No setup needed. Scripts read the API key from ~/.deskclaw/nanobot/config.json, checking providers.custom → anthropic → openai (first …
- Very specific: line 23
When the user provides a local file path (e.g.
- Very specific: line 33
When the user provides a public image URL (e.g.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Multimodal Understanding
Analyze images via NoDesk AI Gateway. The API is OpenAI Vision-compatible.
Configuration
No setup needed. Scripts read the API key from ~/.deskclaw/nanobot/config.json, checking providers.custom → anthropic → openai (first non-empty key wins).
Workflow
Local file (most common)
When the user provides a local file path (e.g. /Users/.../photo.jpg, ~/Desktop/img.png):
bash <skill_dir>/scripts/multimodal-call.sh --file '<image_path>' '<user_prompt>' [max_tokens] [model]The script auto-detects the image type and base64-encodes it. No manual conversion needed.
Image URL
When the user provides a public image URL (e.g. https://example.com/photo.jpg):
bash <skill_dir>/scripts/multimodal-call.sh --url '<image_url>' '<user_prompt>' [max_tokens] [model]The URL must be publicly accessible. No download or base64 conversion needed.
Parameters
max_tokens: default 1000. Use 300 for short descriptions, 2000+ for OCR / detailed analysis.model: defaultgpt-4o.
Present results
Extract choices[0].message.content from the JSON response and show it directly to the user.
Error Handling
| HTTP Status | Meaning | Action |
| 401 / 403 | Invalid API key | Ask user to verify their key |
| 429 | Quota exceeded or rate limited | Tell user to check quota on gateway dashboard |
| Timeout | Image too large or gateway busy | Retry once; if still failing, tell user |
Important
- Local file path → use
--filemode. Tell users: instead of pasting/dragging images, type the file path as text for best results. - Image URL → use
--urlmode. URL must be publicly accessible. - Each image costs ~200–800 prompt tokens. Multimodal and LLM chat share the same token quota.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.