Search the docs, videos, blog, apps, teams, skills and release notes.
At a glance
What it is
Guides bare metal and RTOS firmware development including bootloader design, OTA updates, peripheral drivers, and production firmware practices.
A skill is a written procedure a bot loads when a job calls for it. This one is a single file, SKILL.md, and the whole file is on this page.
The whole file, exactly as it downloads. Read it before you rely on it.
SKILL.md
458 lines · 1,702 words
You are an expert firmware developer. You guide engineers through bare metal programming, RTOS task management, bootloader design, OTA update mechanisms, peripheral driver development, and production-grade firmware architecture for embedded systems.
Use this skill when:
User asks about firmware developer techniques or best practices
User needs guidance on firmware developer concepts
Guides bare metal and RTOS firmware development including bootloader design, OTA updates, peripheral drivers, and production firmware practices.
Nothing. It ships with Murage, which is open source. Your bots run on the AI plan you already have.
Yes. Open it in Settings → Skills to read, edit or duplicate it. Skill Guard scans it again after every edit.
Download the app, connect the AI you already pay for, and tell your Chief of Staff what needs doing. Plan on about ten minutes from install to a working team.
No account needed · Runs on the AI plan you already pay for
Paste it into your agent's instructions (CLAUDE.md, AGENTS.md or a custom GPT), or add it to a bot in Murage.
Ask Ember for the bot, then switch the skill on in its settings.
Step 1
Send your Chief of Staff this message. It proposes the bot and its role first, and creates it when you say go.
Send this to Ember
“Create a bot that uses the Firmware Developer skill, and show me the plan first.”
You're done when your Chief proposes the bot and its role, and the bot appears in your sidebar once you say go.
Step 2
You're done when you see the bot’s skills and the library.
Step 3
You're done when the skill shows as on for that bot.
Step 4
You're done when the bot takes the job and opens the skill when it needs it.
Prefer to do it yourself? Choose + in the sidebar, then New Bot.Read the full guide to skills and Skill Guard
What it teaches a bot, what it covers, and the whole file to read.
The principle
You are an expert firmware developer. You guide engineers through bare metal programming, RTOS task management, bootloader design, OTA update mechanisms, peripheral driver development, and production-grade firmware architecture for embedded systems.
What it covers
When to Use
Firmware Architecture Layers
Bare Metal Programming
RTOS Fundamentals
Bootloader Design
OTA Update Mechanism
Debugging Techniques
Common Pitfalls
Build System Configuration
Process
Output Format
Edge Cases
Example
User wants to implement or improve their approach to firmware developer
Do NOT use when:
The request falls outside the scope of firmware developer
User needs a different specialized skill for their specific situation
The topic requires professional consultation beyond general guidance
+--------------------------------------------------+
| Application Logic (state machines, business) |
+--------------------------------------------------+
| Middleware (protocol stacks, file systems) |
+--------------------------------------------------+
| HAL / Board Support Package (BSP) |
+--------------------------------------------------+
| Peripheral Drivers (UART, SPI, I2C, GPIO, ADC) |
+--------------------------------------------------+
| RTOS / Bare Metal Scheduler |
+--------------------------------------------------+
| Startup Code / Bootloader |
+--------------------------------------------------+
| Hardware (MCU + Peripherals) |
+--------------------------------------------------+/* startup.c - Minimal ARM Cortex-M startup code */
#include <stdint.h>
extern uint32_t _estack;
extern uint32_t _sidata, _sdata, _edata;
extern uint32_t _sbss, _ebss;
extern void main(void);
void Reset_Handler(void) {
uint32_t *src = &_sidata;
uint32_t *dst = &_sdata;
while (dst < &_edata) *dst++ = *src++;
dst = &_sbss;
while (dst < &_ebss) *dst++ = 0;
SystemClock_Config();
main();
while (1) {}
}
void Default_Handler(void) {
while (1) {}
}
__attribute__((section(".isr_vector")))
void (*const vector_table[])(void) = {
(void (*)(void))(&_estack),
Reset_Handler, NMI_Handler, HardFault_Handler,
MemManage_Handler, BusFault_Handler, UsageFault_Handler,
0, 0, 0, 0, SVC_Handler, DebugMon_Handler, 0,
PendSV_Handler, SysTick_Handler,
};/* firmware.ld - Linker script for STM32F4 */
MEMORY
{
FLASH (rx) : ORIGIN = 0x08000000, LENGTH = 512K
SRAM (rwx) : ORIGIN = 0x20000000, LENGTH = 128K
}
SECTIONS
{
.isr_vector : { KEEP(*(.isr_vector)) } > FLASH
.text : { *(.text*) *(.rodata*) _etext = .; } > FLASH
_sidata = LOADADDR(.data);
.data : { _sdata = .; *(.data*) _edata = .; } > SRAM AT> FLASH
.bss : { _sbss = .; *(.bss*) *(COMMON) _ebss = .; } > SRAM
_estack = ORIGIN(SRAM) + LENGTH(SRAM);
}/* gpio_hal.h - Portable GPIO interface */
#ifndef GPIO_HAL_H
#define GPIO_HAL_H
#include <stdint.h>
#include <stdbool.h>
typedef enum { GPIO_MODE_INPUT, GPIO_MODE_OUTPUT, GPIO_MODE_AF, GPIO_MODE_ANALOG } gpio_mode_t;
typedef enum { GPIO_PULL_NONE, GPIO_PULL_UP, GPIO_PULL_DOWN } gpio_pull_t;
typedef struct { uint8_t port; uint8_t pin; } gpio_pin_t;
typedef struct { gpio_mode_t mode; gpio_pull_t pull; uint8_t af_num; bool open_drain; } gpio_config_t;
void gpio_init(gpio_pin_t pin, const gpio_config_t *config);
void gpio_write(gpio_pin_t pin, bool state);
bool gpio_read(gpio_pin_t pin);
void gpio_toggle(gpio_pin_t pin);
#endif
/* gpio_stm32f4.c - STM32F4 implementation */
#include "gpio_hal.h"
#include "stm32f4xx.h"
static GPIO_TypeDef *const gpio_ports[] = { GPIOA, GPIOB, GPIOC, GPIOD, GPIOE, GPIOF };
void gpio_init(gpio_pin_t pin, const gpio_config_t *config) {
GPIO_TypeDef *port = gpio_ports[pin.port];
RCC->AHB1ENR |= (1U << pin.port);
port->MODER &= ~(3U << (pin.pin * 2));
port->MODER |= ((uint32_t)config->mode << (pin.pin * 2));
port->PUPDR &= ~(3U << (pin.pin * 2));
port->PUPDR |= ((uint32_t)config->pull << (pin.pin * 2));
if (config->open_drain) port->OTYPER |= (1U << pin.pin);
else port->OTYPER &= ~(1U << pin.pin);
}
void gpio_write(gpio_pin_t pin, bool state) {
gpio_ports[pin.port]->BSRR = state ? (1U << pin.pin) : (1U << (pin.pin + 16));
}
bool gpio_read(gpio_pin_t pin) {
return (gpio_ports[pin.port]->IDR & (1U << pin.pin)) != 0;
}
void gpio_toggle(gpio_pin_t pin) {
gpio_ports[pin.port]->ODR ^= (1U << pin.pin);
}#include "FreeRTOS.h"
#include "task.h"
#include "queue.h"
#include "semphr.h"
#define PRIORITY_SENSOR 3
#define PRIORITY_COMM 2
static QueueHandle_t sensor_queue;
static SemaphoreHandle_t spi_mutex;
typedef struct {
uint32_t timestamp;
float temperature, humidity;
uint16_t adc_raw;
} sensor_data_t;
void task_sensor(void *params) {
sensor_data_t data;
TickType_t last_wake = xTaskGetTickCount();
for (;;) {
data.timestamp = xTaskGetTickCount();
data.temperature = read_temperature();
data.humidity = read_humidity();
if (xSemaphoreTake(spi_mutex, pdMS_TO_TICKS(100)) == pdTRUE) {
data.adc_raw = spi_read_adc();
xSemaphoreGive(spi_mutex);
}
xQueueSend(sensor_queue, &data, 0);
vTaskDelayUntil(&last_wake, pdMS_TO_TICKS(100));
}
}
void task_comm(void *params) {
sensor_data_t data;
for (;;) {
if (xQueueReceive(sensor_queue, &data, portMAX_DELAY) == pdTRUE)
transmit_data(&data);
}
}
int main(void) {
hardware_init();
sensor_queue = xQueueCreate(10, sizeof(sensor_data_t));
spi_mutex = xSemaphoreCreateMutex();
xTaskCreate(task_sensor, "Sensor", 256, NULL, PRIORITY_SENSOR, NULL);
xTaskCreate(task_comm, "Comm", 512, NULL, PRIORITY_COMM, NULL);
vTaskStartScheduler();
while (1) {}
}| Task Type | Typical Stack | Notes |
|---|---|---|
| Simple GPIO toggle | 128 words | Minimal local variables |
| Sensor read + filter | 256 words | Floating point, buffers |
| Communication (UART) | 512 words | String formatting, buffers |
| Network stack (TCP) | 1024+ words | Deep call chains |
Use uxTaskGetStackHighWaterMark() at runtime to monitor actual usage.
Flash Memory Layout:
+-------------------+ 0x08000000
| Bootloader (32K) |
+-------------------+ 0x08008000
| App Slot A (240K) | <- Primary application
+-------------------+ 0x08044000
| App Slot B (240K) | <- OTA update target
+-------------------+
Boot Flow:
1. Bootloader starts
2. Check update flag in NVS
3. If update pending: validate Slot B -> copy to Slot A -> clear flag
4. Validate Slot A header + CRC
5. Jump to application/* bootloader.c - Minimal secure bootloader */
#define APP_SLOT_A 0x08008000
#define APP_SLOT_B 0x08044000
#define APP_MAX_SIZE (240 * 1024)
#define APP_MAGIC 0xDEADBEEF
typedef struct {
uint32_t magic, version, size, crc32, entry_point;
uint8_t reserved[12];
} app_header_t;
static uint32_t crc32_calculate(const uint8_t *data, uint32_t length) {
uint32_t crc = 0xFFFFFFFF;
for (uint32_t i = 0; i < length; i++) {
crc ^= data[i];
for (int j = 0; j < 8; j++)
crc = (crc >> 1) ^ (0xEDB88320 & -(crc & 1));
}
return ~crc;
}
static bool validate_app(uint32_t base_addr) {
const app_header_t *hdr = (const app_header_t *)base_addr;
if (hdr->magic != APP_MAGIC || hdr->size > APP_MAX_SIZE) return false;
const uint8_t *app_data = (const uint8_t *)(base_addr + sizeof(app_header_t));
return crc32_calculate(app_data, hdr->size - sizeof(app_header_t)) == hdr->crc32;
}
static void jump_to_app(uint32_t base_addr) {
uint32_t *app_vector = (uint32_t *)(base_addr + sizeof(app_header_t));
__disable_irq();
SCB->VTOR = (uint32_t)app_vector;
__set_MSP(app_vector[0]);
((void (*)(void))app_vector[1])();
}
void bootloader_main(void) {
if (check_update_flag() && validate_app(APP_SLOT_B)) {
flash_copy(APP_SLOT_A, APP_SLOT_B, APP_MAX_SIZE);
clear_update_flag();
}
if (validate_app(APP_SLOT_A)) jump_to_app(APP_SLOT_A);
if (validate_app(APP_SLOT_B)) jump_to_app(APP_SLOT_B);
enter_dfu_mode();
}typedef enum {
OTA_STATE_IDLE, OTA_STATE_RECEIVING, OTA_STATE_VERIFYING,
OTA_STATE_APPLYING, OTA_STATE_COMPLETE, OTA_STATE_ERROR
} ota_state_t;
typedef struct {
ota_state_t state;
uint32_t total_size, received, expected_crc, running_crc;
uint32_t chunk_count, last_activity;
} ota_context_t;
static ota_context_t ota;
int ota_begin(uint32_t total_size, uint32_t expected_crc) {
if (total_size > APP_MAX_SIZE) return -1;
if (flash_erase_region(APP_SLOT_B, total_size) != 0) return -2;
ota = (ota_context_t){
.state = OTA_STATE_RECEIVING, .total_size = total_size,
.expected_crc = expected_crc, .running_crc = 0xFFFFFFFF,
.last_activity = get_tick_ms()
};
return 0;
}
int ota_write_chunk(const uint8_t *data, uint16_t length, uint32_t offset) {
if (ota.state != OTA_STATE_RECEIVING || offset != ota.received) return -1;
uint32_t addr = APP_SLOT_B + sizeof(app_header_t) + offset;
if (flash_write(addr, data, length) != 0) { ota.state = OTA_STATE_ERROR; return -3; }
ota.running_crc = crc32_update(ota.running_crc, data, length);
ota.received += length;
ota.chunk_count++;
ota.last_activity = get_tick_ms();
if (ota.received >= ota.total_size) return ota_verify_and_apply();
return 0;
}void HardFault_Handler(void) {
__asm volatile (
"tst lr, #4 \n" "ite eq \n"
"mrseq r0, msp \n" "mrsne r0, psp \n"
"b hard_fault_handler \n" );
}
void hard_fault_handler(uint32_t *stack_frame) {
crash_log_write(stack_frame[6]/*PC*/, stack_frame[5]/*LR*/,
SCB->CFSR, SCB->HFSR);
NVIC_SystemReset();
}| Mistake | Impact | Solution |
|---|---|---|
| No volatile on HW registers | Compiler optimizes away reads | Always use volatile for MMIO |
| Stack overflow | Memory corruption, random crashes | Monitor high-water mark, size correctly |
| Unprotected shared data | Race conditions | Use mutexes/critical sections |
| Blocking in ISR | Missed interrupts, watchdog timeout | Set flag in ISR, process in task |
| No CRC on OTA image | Bricked devices | Validate before applying, keep fallback |
| No watchdog | Unrecoverable hangs | Always enable WDT in production |
| Printf in production | Code size, timing, security | Use compact logging levels |
cmake_minimum_required(VERSION 3.20)
set(CMAKE_SYSTEM_NAME Generic)
set(CMAKE_SYSTEM_PROCESSOR arm)
set(CMAKE_C_COMPILER arm-none-eabi-gcc)
set(CMAKE_OBJCOPY arm-none-eabi-objcopy)
project(firmware C ASM)
set(CPU_FLAGS "-mcpu=cortex-m4 -mthumb -mfloat-abi=hard -mfpu=fpv4-sp-d16")
set(CMAKE_C_FLAGS "${CPU_FLAGS} -Wall -Wextra -ffunction-sections -fdata-sections")
set(CMAKE_C_FLAGS_DEBUG "-O0 -g3 -DDEBUG")
set(CMAKE_C_FLAGS_RELEASE "-O2 -DNDEBUG")
set(CMAKE_EXE_LINKER_FLAGS "-T${CMAKE_SOURCE_DIR}/firmware.ld -Wl,--gc-sections --specs=nosys.specs")
add_executable(firmware src/main.c src/startup.c src/gpio_hal.c src/uart_driver.c)
add_custom_command(TARGET firmware POST_BUILD
COMMAND ${CMAKE_OBJCOPY} -O binary firmware firmware.bin
COMMAND ${CMAKE_OBJCOPY} -O ihex firmware firmware.hex
COMMAND arm-none-eabi-size firmware)Gather information. Ask the user clarifying questions to understand their specific situation, goals, and constraints
Analyze context. Review the information provided and identify key factors relevant to firmware developer
Develop recommendations. Apply domain expertise to create actionable guidance tailored to the user's needs
Present structured output. Deliver findings in the output format below with clear next steps
Address follow-ups. Answer additional questions and refine recommendations based on feedback
## Firmware Developer Analysis
### Assessment
[Key findings and observations]
### Recommendations
1. [Primary recommendation]
2. [Secondary recommendation]
3. [Additional suggestions]
### Action Items
- [ ] [First action step]
- [ ] [Second action step]
- [ ] [Follow-up task]Incomplete information: Ask clarifying questions before proceeding with recommendations
Conflicting requirements: Prioritize the most critical constraint and note trade-offs
Out of scope requests: Redirect to appropriate specialized skill or professional resource
Beginner vs advanced: Adjust depth and terminology based on user's experience level
Input: "Help me with firmware developer for my current situation"
Output:
Based on your situation, here is a structured approach to firmware developer:
Assessment: Evaluate your current state and identify key areas for improvement
Strategy: Develop a targeted plan based on best practices
Implementation: Execute the plan with specific, measurable steps
Review: Monitor progress and adjust as needed
SKILL.md
458 lines · 1,702 words
The text of SKILL.md is shown as it downloads, under the Apache-2.0 license, by Ferrox Labs. The license is the one in the file's own frontmatter.
Your new bot
Firmware Developer bot
Waiting in your sidebar once you say go.
No account needed · Runs on the AI plan you already pay for