Skip to content

Claude Skills by Anri-Lombard

github.com/Anri-Lombard

1 skillA× 10 installs0 views

BabysitAWatch a long-running ML training job overnight and step in when it breaks: NaN or exploding loss, loss spikes, CUDA OOM, crashes, hangs, SLURM jobs that die or time out, full disks. Shows progress bars, loss sparklines and loss charts. Waits without polling, diagnoses from the log, applies the fix the user allowed (usually resume from the last checkpoint with a safer setting), and leaves a morning report. Use when the user says "babysit", "watch my run", "keep an eye on training", "monitor th...ai-agentspythongoVotes: 0GitHub stars: 2

Claude Skills by Anri-Lombard — 1 skill, security-graded | Skills Directory