custom-allocators
Custom allocator skill for memory allocation strategies. Use when implementing…
STM32 bare-metal skill for CMSIS-only MCU projects. Use when scaffolding STM32 firmware without HAL, configuring clocks/RCC, using CMSIS headers, or building with arm-none-eabi-gcc. Activates on queries about STM32 bare metal, CMSIS without HAL, STM32F4/H7 bring-up, or minimal
$ npx -y skills add mohitmishra786/low-level-dev-skills --skill stm32-baremetal --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/stm32-baremetalContext preview
The summary Claude sees to decide when to auto-load this skill.
STM32 bare-metal skill for CMSIS-only MCU projects. Use when scaffolding STM32 firmware without HAL, configuring clocks/RCC, using CMSIS headers, or building with arm-none-eabi-gcc. Activates on queries about STM32 bare metal, CMSIS without HAL, STM32F4/H7 bring-up, or minimal
name: stm32-baremetal description: STM32 bare-metal skill for CMSIS-only MCU projects. Use when scaffolding STM32 firmware without HAL, configuring clocks/RCC, using CMSIS headers, or building with arm-none-eabi-gcc. Activates on queries about STM32 bare metal, CMSIS without HAL, STM32F4/H7 bring-up, or minimal Makefile CMake for Cortex-M.
Guide agents through a minimal STM32 bare-metal project using CMSIS device headers and startup code only — no STM32 HAL — covering clock setup, linker script integration, peripheral register access, and a reproducible build layout for STM32F/G/H/L families.
stm32-bare/ ├── startup_stm32f407xx.s # vector table + Reset_Handler ├── system_stm32f4xx.c # SystemInit(), SystemCoreClockUpdate() ├── stm32f407xx.h # CMSIS (from ST pack or copy) ├── linker.ld # FLASH/RAM regions ├── main.c └── Makefile
Use CMSIS-Core (`core_cm4.h`) and device header from ST CMSIS pack or open-source packs (e.g. `stm32-cmsis-device-f4`).
STM32 requires explicit HSE/HSI and PLL setup before high-speed peripherals:
/* system_stm32f4xx.c — simplified PLL from 8 MHz HSE */
void SystemInit(void)
{
RCC->CR |= RCC_CR_HSEON;
while (!(RCC->CR & RCC_CR_HSERDY)) {}
RCC->PLLCFGR = RCC_PLLCFGR_PLLSRC_HSE | (8 << 0) | (336 << 6) | (0 << 16) | (7 << 24);
RCC->CFGR |= RCC_CFGR_PPRE1_DIV2 | RCC_CFGR_PPRE2_DIV1 | RCC_CFGR_HPRE_DIV1;
RCC->CR |= RCC_CR_PLLON;
while (!(RCC->CR & RCC_CR_PLLRDY)) {}
FLASH->ACR = FLASH_ACR_LATENCY_5WS;
RCC->CFGR |= RCC_CFGR_SW_PLL;
while ((RCC->CFGR & RCC_CFGR_SWS) != RCC_CFGR_SWS_PLL) {}
SystemCoreClockUpdate();
}Always match `FLASH_ACR` wait states to voltage scale and SYSCLK per reference manual table.
#define RCC_AHB1ENR_GPIOAEN (1U << 0)
static inline void gpioa_clock_enable(void)
{
RCC->AHB1ENR |= RCC_AHB1ENR_GPIOAEN;
(void)RCC->AHB1ENR; /* AHB read-after-write — required on STM32 */
}MCU = -mcpu=cortex-m4 -mthumb -mfpu=fpv4-sp-d16 -mfloat-abi=hard CFLAGS = $(MCU) -Wall -Wextra -ffunction-sections -fdata-sections -g3 LDFLAGS = $(MCU) -T linker.ld -Wl,--gc-sections -specs=nano.specs -lc -lm -lnosys
| Region | Address | Notes | |--------|---------|-------| | Flash | `0x08000000` | Vector table at boot | | SRAM1 | `0x20000000` | Stack, heap, .data/.bss | | CCM | `0x10000000` | F4 only — not DMA-accessible from all masters |
Bootloader apps often link at `0x08010000` and set `SCB->VTOR = 0x08010000` on entry.
QEMU STM32 machines (`stm32vldiscovery`, `netduinoplus2`, etc.) use **different** MCUs than F407 and omit GPIO/DMA/I2C per [QEMU STM32 docs](https://www.qemu.org/docs/master/system/arm/stm32.html). Use `skills/qemu/qemu-embedded-simulation` for simulation; validate F4 firmware on hardware.
/stm32-baremetal Scaffold CMSIS-only STM32F407 project with USART2 printf
| Symptom | Cause | Fix | |---------|-------|-----| | Hang in `SystemInit` | HSE not populated / wrong PLL | Use HSI for bring-up; verify crystal | | Peripheral dead | RCC clock gate off | Enable AHB/APB bit; read-back RCC | | HardFault on boot | Stack in wrong region / VTOR | Check `_estack` in linker.ld | | Wrong baud rate | `SystemCoreClock` stale | Call `SystemCoreClockUpdate()` after PLL | | DMA fails from CCM | CCM not on AHB matrix path | Place DMA buffers in SRAM1 |
A curated suite of AI agent skills for systems and low-level programming — C/C++, Rust, Zig, GPU, bare-metal firmware, Linux kernel/driver development, computer architecture, compiler internals, HPC, and more.
Repo: mohitmishra786/low-level-dev-skills
Custom allocator skill for memory allocation strategies. Use when implementing…
NUMA programming skill for multi-socket memory locality. Use when detecting NUMA topology,…
AF_XDP skill for high-performance XDP sockets. Use when creating AF_XDP sockets, configuring…
DPDK skill for userspace packet I/O. Use when initializing EAL, configuring PMD drivers,…
io_uring skill for Linux async I/O. Use when building high-performance servers with liburing,…
Bare-metal ADC and DAC skill. Use when configuring analog sampling, DMA-driven ADC,…