Kicad emsim

Skill AvatarSD/KiSkill/skills/kicad-emsim

Agent skills + headless Python engine giving an AI coding agent full control of KiCad — edit, verify, review & fabricate schematics and PCBs, no GUI.

Install

npx -y skills add AvatarSD/KiSkill --skill kicad-emsim

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Electromagnetic field simulation of a KiCad board via gerber2ems + openEMS (dockerized, headless): S-parameters, impedance, and E-field PNG visualizations fed back to the agent. Use when asked to simulate a board's EM behavior, check impedance of traces, or visualize fields.

SKILL.md

1.9 KB, as published. Nobody here has run it

EM simulation (gerber2ems + openEMS)

Engine: $(kx root)/kicad_lib/emsim.py. STATUS: chain VALIDATED — meander_loose example simulated end-to-end (rc 0; |S11|≈1 + 455 ps delay on the open meander = physically correct). Image gerber2ems (6.6 GB); rebuild if missing: docker build -t gerber2ems $(kx root)/.tools/gerber2ems/. NEVER mount over /home/docker — it shadows the image toolchain; emsim.run mounts at /home/docker/sim.

Flow

emsim.prepare(pcb, workdir) — gerbers+drill+pos via fab.py (names already match), stackup.json translated from the board's (setup (stackup ...)), simulation.json template.
Fill ports in simulation.json (width/length in µm, impedance, layer index, excite flag) and traces (start/stop port indices) — ports sit at position-file footprint locations (use simulation-port footprints on the board for clean port placement).
emsim.run(workdir, export_field=True) — dockerized gerber2ems -a --export-field: geometry → FDTD → postprocess.
Read the outputs: ems/*.csv S-params/impedance; field PNGs are the agent feedback — Read them, check field concentration against expectations (return paths, gaps, stubs). VTR dumps open in ParaView for the user.

Notes

Frequency range/grid in SIM_TEMPLATE: 0.2–6 GHz, sane FR-4 defaults; finer grids explode runtime (FDTD is O(cells × steps)).
Worked examples: .tools/gerber2ems/examples/* (differential pair, filters, stubs) — copy a simulation.json from the closest one.
Docker daemon does its own networking — image build/pull works even though the agent sandbox blocks direct DNS.