Add m5stack-uiflow skill

[ishamehra]

Summary

Adds an agent skill for deploying and debugging MicroPython firmware on M5Stack devices (CoreS3 SE, Core2, Core2 for AWS) using UIFlow 2.0.

• Audio playback — PCM formats, Speaker.playRaw() buffer limits, PSRAM streaming, chunked playback, voice TTS prefetching
• Touch handling — ghost touch filtering (CoreS3 SE hardware defect at x=58, y=202), debouncing patterns
• Display — ILI9342C color codes, flicker-free timer updates, setTextColor() background bug
• Transport modes — WiFi polling, USB serial bridge, demo mode
• Device onboarding — M5Burner flashing, boot_option NVS config, Thonny file upload, CH9102 driver setup (macOS)
• Serial bridge — protocol, DTR reset prevention, command parity, Thonny port conflicts
• Cross-device portability — confirmed zero-change portability between CoreS3 SE and Core2 (Feb 2026)

Structure

skills/m5stack-uiflow/
├── SKILL.md (324 lines — core procedures)
├── LICENSE.txt
└── references/
    ├── audio.md              — audio playback deep dive
    ├── audio-streaming.md    — detailed audio debugging logs
    ├── hardware.md           — device comparison, onboarding, portability
    └── serial-bridge.md      — serial bridge, BLE, iOS integration

Uses progressive disclosure: SKILL.md is lean with essential procedures, reference files are loaded by Claude only when needed.

Context

Built from months of hands-on M5Stack development in the Microsteps project (routine coaching app with hardware companion tile). Every pattern and gotcha comes from real bugs encountered and solved on actual hardware.