What it solves
A hardware watchdog alone cannot identify which cooperative task stopped making progress.
How it works
1Task kicks + clock→2deadline state→3timeout/reset callbacks
Use it when
Track a sensor task deadline
Request reset after repeated misses
Report which task starved
Quick Start
#include "mwdt.h"
/* Initialize caller-owned tasks and clock callbacks; kick tasks and call the checker. */Engineering evidence
- State model and C/C++ consumption contracts are documented
Known limits
- Volatile RAM state only
- No built-in hardware watchdog, scheduler, logging, thread or ISR safety