Check local screen sessions and remote endpoints for experiment status anomalies.
npx skills add https://github.com/skyllwt/AutoSci --skill exp-status-skyllwt
Please help me install this Agent Skill. Skill: /exp-status Source: https://github.com/skyllwt/AutoSci/tree/main/.claude/skills/exp-status Command: npx skills add https://github.com/skyllwt/AutoSci --skill exp-status-skyllwt
It helps you track which training experiments are still running, which have anomalies, and which are ready to be collected, so your research pipeline does not stall.
screen sessions or remote jobs to report alive, anomaly, and completion states./exp-run <slug> --collect to update wiki experiment pages./research --start-from stage4 once all experiments are completed.Run /exp-status --pipeline <pipeline-slug> to see progress for that pipeline and identify which experiments need collection or attention.
High-intent search queries and answers about installing and using this skill.
To monitor running training experiments, you can check local screen sessions or remote experiment endpoints to report alive, anomaly, and completion states. This prevents your research pipeline from stalling by identifying which runs need collection or attention.
Anomaly detection for ongoing training runs is handled by scanning active experiments and checking their local or remote status commands to summarize anomalies. It flags sessions that exhibit irregular states alongside those that are alive or ready for collection.
Auto-collecting finished screen sessions triggers a collection command like `/exp-run <slug> --collect` to update wiki experiment pages. This automates the gathering of completed training runs and conditionally updates wiki outputs for finished but uncollected sessions.
You do not strictly need remote job access; the status check supports both local screen sessions and remote experiment endpoints. You can monitor ongoing and completed training runs by reading wiki experiment pages and running status commands in either environment.
The best way to track completion readiness is unified experiment status monitoring, which scans running sessions and reports alive, anomaly, and completion states. This summarizes which experiments are ready to be collected to prevent pipeline bottlenecks.
duany049Monitor experiment screen sessions and classify status as running, anomaly, or completed.
MACJONSWOrchestrate literature survey, idea creation, novelty checking, and critical review into a ranked IDEA_REPORT.md.
duany049Execute pilot experiments from YAML specs and collect raw JSON results.
LambenthanMonitor experiment runtime states and flag anomalies from screen or GPU/SSH runs.