Honggfuzz dies during fuzzing

I am running a fuzzer benchmark run and honggfuzz dies left and right with the same pattern:

```
[2026-08-04T21:58:11+0000][W][10] arch_checkWait():238 Persistent mode: pid=47 exited with status: SIGNALED, signal: 9 (Killed)
[2026-08-04T21:58:14+0000][W][10] arch_traceWaitForPidStop():1069 PID 50 not in a stopped state - status:14 (SIGNALED, signal: 14 (Alarm clock))
[2026-08-04T21:58:14+0000][W][10] arch_attachToNewPid():177 arch_traceAttach(pid=50) failed
[2026-08-04T21:58:14+0000][F][10] arch_prepareParentAfterFork():204 Couldn't attach to pid=50
```

This is on oss-fuzz targets. with only a few parallel instances this does not happen, but the more parallel instances are run, the more honggfuzz instances abort.

Reproduce the failure on an oss-fuzz target while increasing the number of parallel instances. Start by tracing the logged paths through arch_checkWait(), arch_traceWaitForPidStop(), arch_attachToNewPid(), and arch_prepareParentAfterFork(). Done means identifying why the processes are killed or cannot be attached and confirming that parallel fuzzing no longer causes honggfuzz instances to abort.

Written by the indexing model from the issue text.