Hi!
I’m looking for Claude skills that do more than just generic C++ and focus more on embedded devices. The goal is resource-efficient code (and debugging) that can handle the memory and hardware limits of ESP32 with ESP-IDF.
- Memory Allocation Strategy
Generating code that distinguishes between IRAM, DRAM, and PSRAM. This includes the correct use of MALLOC_CAP_SPIRAM vs MALLOC_CAP_DMA and MALLOC_CAP_INTERNAL based on the specific hardware constraints.
- Flash vs. RAM Execution
Proper use of IRAM_ATTR for ISRs and critical functions to prevent “Cache disabled” crashes during flash operations.
- Proactive Heap Management
Avoiding fragmentation in long-running nodes and implementing smart buffering instead of naive string/vector operations.
- Static Analysis & Binary Inspection
Skills that can analyze ‘size’ and ‘nm’ output to monitor .iram0.text, .dram0.data, and .flash.rodata growth, ensuring the model understands the impact of its code on the final memory footprint.
- Virtual Validation (QEMU)
Workflows where Claude can generate code, compile for a virtual Xtensa target, and use GDB/QEMU to verify execution flow and stack high-water marks before flashing.
- Crash & Backtrace Analysis
Tuning the model to accurately map register dumps and backtraces directly to ELF symbols for rapid iteration.
If anyone has built or knows about any skills that handle any of the points above, I’d love some pointers.
Thanks!