relax-dev-debug

A development and debugging guide for the Relax reinforcement-learning project, including its distributed training workflows. Reinforcement learning trains systems through feedback, while Ray is a platform for running work across processes or machines.

In plain words
What is it for?
Changing code in the relax directory, investigating training problems, and validating requested training runs on a Ray cluster.
Why use it?
It helps keep code changes focused and provides project-specific checks when training or debugging is explicitly required.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/redai-infra/relax/dev
Any agent
npx skills add redai-infra/Relax --skill dev
Clone the repo
git clone --depth 1 https://github.com/redai-infra/Relax

Made for: Claude Code, Codex.

Per session 60 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,405 The whole file, excluding the scripts and references it only reads on demand.
Security scan E 3 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00060 $0.03405
Opus 5 $0.00030 $0.01702
Sonnet 5 $0.00012 $0.00681
Haiku 4.5 $0.00006 $0.00341

Measured 3d ago against content hash 7fe14e0df9ff, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade E, and why

relax-dev-debug scanned grade E with 3 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Downloads and executes remote codehighSupply chain

curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.

curl -s http://${CLUSTER_IP}:8265/api/serve/applications/ | python3 -m json.tool

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

**Strictly forbidden during this flow** (per `skills/ssh-ray-cluster/SKILL.md`): `ray stop`, `pkill -9 python`, `bash scripts/tools/kill_for_ray.sh`, `ray job stop` against unrelated jobs, `rm -rf /tmp/ray/`. `ray serve

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -s http://${CLUSTER_IP}:8265/api/serve/applications/ | python3 -m json.tool
skills/dev/SKILL.md · 263 lines

How it starts

The opening of the file, as written. The whole thing — 263 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Relax Development & Debugging

This skill covers two workflows: making code changes to the Relax project, and validating those changes by running a training job on a remote Ray cluster.

For project structure and coupling details, see AGENTS.md.


Part 1: Development

Minimum-change principle

Unless the user explicitly asks for a refactoring, apply the smallest diff that achieves the goal. This means:

  • Only touch files and lines directly required by the change. Don't "improve" nearby code, reformat untouched lines, or rename things that aren't part of the task.
  • Preserve the existing code style in each file — naming conventions, import ordering, indentation, comment style.
  • Don't introduce new dependencies unless the change itself demands it.
  • Don't alter function signatures, class hierarchies, or public APIs unless that is the stated purpose of the change.

If the user says "refactor" or "restructure", the minimum-change constraint is lifted — but confirm the scope before proceeding.


Part 2: Debugging (Remote Training Validation)

Do not enter this workflow on your own. Only proceed when the user explicitly asks to debug/validate on a remote cluster and provides a RAY_ADDRESS. Code changes alone do not trigger a debug run — the user decides when to test on real hardware.

The goal is to submit a training run to a remote Ray cluster via scripts/entrypoint/ray-job.sh, monitor it via ray job logs, and either confirm it works or fix errors and retry.


Standard launch flow (canonical)

Every training launch — the first one and every resubmit after a fix — follows the same three steps. Do not skip the pre-flight cleanup; stale Ray Serve apps from a failed previous job will silently break the new job at Router startup.

Read the full file on GitHub · 263 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 263 lines · 60 tokens per session scan E 7fe14e0df9ff

Subscribe to this mod's changes

relax-dev-debug is a skill published in the GitHub repository redai-infra/Relax (580 stars, last pushed 5d ago), licensed Apache-2.0. It adds 60 tokens to every session and 3,405 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it E with 3 findings (downloads and executes remote code, recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

verl-rl-training

Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.

Orchestra-Research/AI-Research-SKILLs · 51 tokens

openrlhf-training

High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.

Orchestra-Research/AI-Research-SKILLs · 72 tokens

openrlhf-training

High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.

liortesta/ClawdAgent · 72 tokens

verl-rl-training

Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.

ihatesea69/HieuNghi-AI-Skills · 51 tokens

openrlhf-training

High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.

ihatesea69/HieuNghi-AI-Skills · 72 tokens

grpo-rl-training

Expert guidance for GRPO/RL fine-tuning with TRL for reasoning and task-specific model training.

Orchestra-Research/AI-Research-SKILLs · 26 tokens