image-analysislisted
Install: claude install-skill Smart-AI-Memory/attune-ai
# Image Analysis
**IMPORTANT: Start your response with a context preamble.**
Call `help_lookup(topic="image-analysis", mode="preamble")` and
display the returned `preamble` text as a blockquote. Then tell the
user they can say "tell me more" for a step-by-step guide, or answer
the scoping question below to proceed.
If the MCP call fails, fall back to:
> **Image Analysis** — Sends an image (screenshot, diagram, UI
> mockup, chart) to Claude's vision model and returns a description
> or answers a question about it. Supports PNG, JPEG, GIF, and WebP.
## Scoping
Before running, ask:
1. **Image path**: "Which image file should I analyze?"
2. **Focus** (optional): "Anything specific to look for, or a general
description?"
## Execution
### Shared command workspace (preferred)
Open adapter `image-analysis` with the local image path and optional prompt.
The adapter reads the real file, validates repository containment, the 10MB
limit, magic bytes versus extension, dimensions, MIME type, and SHA-256. The
invocation already authorizes this read-only analysis, so the running workspace
has no synthetic confirmation action.
Call `analyze_image` with the validated path and publish the exact response as
`analysis_result`; include provider progress only as optional `progress`
events. Success requires non-empty analysis and must match the canonical MIME
and file size. Decode/provider failure must say “did not complete,” never
render an empty successful analysis. Present the termi