cmbacktrace-integration

# CmBacktrace Integration ## Overview Use this skill to integrate CmBacktrace as a Cortex-M crash diagnosis tool. The key is to preserve fault context, provide correct firmware metadata, route output safely, and verify that backtraces map to the running ELF. ## When To Use Use this skill when: - The target is Cortex-M and the user wants automatic fault diagnosis or stack backtraces. - The issue involves HardFault, fault handlers, register dumps, firmware name/version/hardware version, or crash logs. - The project uses or plans to use CmBacktrace. Do not use this skill for non-Cortex-M faults. Use `embedded-fault-debug` or the architecture-specific debug skill instead. ## First Questions Ask for: - MCU/core, compiler, optimization level, and build flags. - Fault handler code and whether the system resets immediately. - Output backend for crash text: UART, RTT, flash, EasyLogger, shell, or custom. - Firmware metadata values and where they are defined. - Whether the ELF/MAP file matching the running image is available. - Current fault output or absence of output. ## Integration Checklist 1. Integrate in the fault path. Call CmBacktrace from the HardFault/fault handling path before reset and before clearing fault state. 1. Provide metadata. Firmware name, hardware version, software version, and platform identifiers should be meaningful enough for field logs. 1. Keep output safe. Crash output must not block forever, allocate heavily, or depend on services tha...