Are other tests close to hitting the timeout in your runs as well? If the issue is just that your machine is much slower than the benchmark by which we've chosen the timeout values, a better fix would be to have test-skeleton.c grok an environment variable to apply a multiplier to the hard-coded timeouts.