RL Status

SkillMonitoring & ops

One-to-one Codex counterpart for Claude `/rl:status`. Diagnose a Lego-RL run that is live or recently finished; identify the run, locate the real log, report progress, inspect metrics, and match known failure signatures. Use when the user asks how a run is doing, what step it is on, whether reward is improving, whether a run is broken, or asks `/rl:status`.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the RL Status skill

What this skill tells your AI

The instructions your AI receives, as published by legox/lego-rl in .agents/skills/rl-status/SKILL.md and read by ahel’s review.

Use this skill as the Codex entrypoint corresponding to Claude /rl:status.

Before taking task actions, read the canonical workflow in .claude/plugins/rl-plugin/skills/status/SKILL.md completely and follow it as the source of truth.

Invocation Mapping

  • Claude: /rl:status
  • Codex: $rl-status or natural language such as "how's the run doing", "what step is it on", "is reward going up", "is this run broken", or "diagnose the training run".

Codex Notes

Find the Lego-RL repo root first. It must contain scripts/lib/live_probe.sh and the .claude plugin path above.

Remain read-only. Do not kill, restart, clean, rotate, edit configs, delete logs, SSH to other nodes, or mutate Kubernetes. When evidence points off-box, report the symptom and the node or layer to confirm rather than asserting a cause you cannot observe locally.

Signals

GitHub stars
86
Forks
4
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
rl-status
Source
github.com/legox/lego-rl