Skill

klayoutclaw:visual

Captures current KLayout layout as PNG image and GDS file using gdstk, matplotlib, and numpy for visual inspection of geometry. Provides file paths for direct viewing.

Python

design

npx claudepluginhub caidish/klayoutclaw --plugin klayoutclaw

Tool Access

This skill uses the workspace's default tool permissions.

Preview

Save the current KLayout layout to a temporary GDS file, convert it to PNG, and return both file paths for visual inspection.

Supporting Assets

scripts/capture.py

SKILL.md

Similar Skills

klayoutclaw:display

Toggles KLayout layer visibility on/off or shows only specific layers (e.g., metal, mesa) for visualization during design review. Includes toggle_layer.py and show_only.py scripts.

2 files

klayoutclaw

pencil-mcp-get-screenshot

328

Captures a visual screenshot of a specific node in a Pencil .pen file via get_screenshot MCP tool. Use to verify design operations, check alignment or spacing issues, or review components after batch_design changes.

4 files

partme-ai-full-stack-skills

visual-mockup

Creates interactive drag-and-drop mockups for node diagrams or matplotlib plots for charts to confirm layouts before coding CeTZ/Fletcher diagrams with 4+ nodes.

1 file

workflows

Stats

Stars15

Forks3

Last CommitApr 17, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

KLayout Visual Capture

Save the current KLayout layout to a temporary GDS file, convert it to PNG, and return both file paths for visual inspection.

Prerequisites

KLayout running with KlayoutClaw plugin (v0.3+)

A layout with geometry must be open

Python packages: gdstk, matplotlib, numpy (available in the instrMCPdev conda env)

Script

capture.py — Capture layout as PNG

python scripts/capture.py [--output path.png] [--gds path.gds] [--dpi 200]

--output — PNG output path. If the user has a working directory or stack path, save there instead of /tmp.

--gds — GDS output path. Same rule: prefer the user's working directory over /tmp.

--dpi — Image resolution (default: 200)

Returns the paths to both the GDS and PNG files, printed to stdout.

Example:

python scripts/capture.py # Output: # GDS saved: /tmp/klayoutclaw_capture.gds # PNG saved: /tmp/klayoutclaw_capture.png

How It Works

Calls save_layout via MCP to write the current layout to a temp GDS file

Runs tools/gds_to_image.py from the KlayoutClaw repo to convert GDS to PNG

Prints both file paths

The GDS-to-PNG conversion uses gdstk to parse the GDS and matplotlib to render all layers with distinct colors and a legend.

After Capture

Use the Read tool on the PNG path to view the image directly in the conversation. This gives immediate visual feedback on layout geometry.

KLayout Visual Capture

Save the current KLayout layout to a temporary GDS file, convert it to PNG, and return both file paths for visual inspection.

Prerequisites

KLayout running with KlayoutClaw plugin (v0.3+)
A layout with geometry must be open
Python packages: gdstk, matplotlib, numpy (available in the instrMCPdev conda env)

Script

capture.py — Capture layout as PNG

python scripts/capture.py [--output path.png] [--gds path.gds] [--dpi 200]

--output — PNG output path. If the user has a working directory or stack path, save there instead of /tmp.
--gds — GDS output path. Same rule: prefer the user's working directory over /tmp.
--dpi — Image resolution (default: 200)

Returns the paths to both the GDS and PNG files, printed to stdout.

Example:

python scripts/capture.py
# Output:
# GDS saved: /tmp/klayoutclaw_capture.gds
# PNG saved: /tmp/klayoutclaw_capture.png

How It Works

Calls save_layout via MCP to write the current layout to a temp GDS file
Runs tools/gds_to_image.py from the KlayoutClaw repo to convert GDS to PNG
Prints both file paths

The GDS-to-PNG conversion uses gdstk to parse the GDS and matplotlib to render all layers with distinct colors and a legend.

After Capture

Use the Read tool on the PNG path to view the image directly in the conversation. This gives immediate visual feedback on layout geometry.