Skills that teach AI coding agents (Claude Code, Cursor, Codex, Copilot and others)
how to run, test, debug and profile embedded firmware on the Wokwi simulator:
set up CI, run ESP-IDF projects from idf.py, script buttons and sensors, soak-test
for memory leaks. One folder per skill, each a SKILL.md in the open
Agent Skills format, so the same files work in
every agent that reads skills.
Agents can already write firmware. What they cannot do is find out whether it works without a board on the desk. These skills close that loop.
| Skill | Use it to |
|---|---|
find-memory-leaks |
Catch slow leaks in ESP-IDF firmware: the simulator watches the heap while a scenario drives hours of use in seconds |
test-lvgl-ui |
Regression-test an LVGL (or any) display UI in CI with taps, swipes and pixel-exact screenshot compare |
debug-esp32-firmware |
Decode a panic backtrace, reproduce the bug with scripted inputs, trace a bus to VCD, attach GDB without a probe |
verify-esp-idf-build |
Run and check an ESP-IDF build with idf.py wokwi right after building, no devkit attached |
Claude Code, as a plugin (all skills, namespaced /wokwi:<skill>):
/plugin marketplace add wokwi/skills
/plugin install wokwi@wokwi-skills
Claude Code, one project, no plugin: copy skills/<name>/ into the repo's
.claude/skills/ (or ~/.claude/skills/ for every project).
Any other agent: copy the folder to wherever that agent loads skills from, or paste
the SKILL.md body into the instructions file it reads (AGENTS.md, .cursorrules,
copilot-instructions.md). The files are plain Markdown with a small YAML header.
MIT, see LICENSE.