Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/redai-infra/Relaxnpx agentmods add skills/redai-infra/relax/redaccel-to-relaxWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/redai-infra/relax/redaccel-to-relax)<a href="https://agentmods.dev/skills/redai-infra/relax/redaccel-to-relax"><img src="https://agentmods.dev/badge/skills/redai-infra/relax/redaccel-to-relax/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/redai-infra/relax/redaccel-to-relax"><img src="https://agentmods.dev/badge/skills/redai-infra/relax/redaccel-to-relax.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00065 | $0.02600 |
| Opus 5 | $0.00032 | $0.01300 |
| Sonnet 5 | $0.00013 | $0.00520 |
| Haiku 4.5 | $0.00006 | $0.00260 |
Grade A, and why
redaccel-to-relax scanned grade A with 0 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 12d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 250 lines — stays where its author put it; the contents beside it link to each section on GitHub.
RedAccel → Relax Algorithm Migration
This skill guides migration of RL training algorithms (reward functions, agent environments, training scripts) from the RedAccel framework to Relax.
For mapping tables and code templates, see references/migration_mapping.md.
Migration overview
A RedAccel algorithm typically consists of:
| RedAccel Component | RedAccel Location | Relax Equivalent | Relax Location |
|---|---|---|---|
| Reward class (GRPO/Group) | aipet_rl/reward_*.py |
Async reward_func(args, sample) |
examples/<algo>/reward_<algo>.py |
| Agent / Tool env | aipet_rl/agent/*.py (ToolBase) |
BaseInteractionEnv subclass |
examples/<algo>/env_<algo>.py |
| LLM-as-Judge helpers | aipet_rl/llm_judge.py |
Inlined or standalone module | examples/<algo>/llm_judge.py |
| Training launch script | exps/*.sh (redaccel-cli) |
Shell script (ray job submit) |
examples/<algo>/run_<algo>.sh |
| Search / external tools | aipet_rl/agent/unified_search_async.py |
Same module, imported from example dir | examples/<algo>/search_tools.py |
The algorithm code lives under examples/<algo>/ in Relax — not inside the framework core.
Workflow
Step 0: Create the target directory
mkdir -p examples/<algo>
touch examples/<algo>/__init__.py
Step 1: Migrate reward function
This is the most critical step. RedAccel and Relax have fundamentally different reward interfaces.
RedAccel pattern (class-based, synchronous)
# RedAccel: registry-decorated class, __call__ returns List[dict]
from redaccel.verl.rewards.group.base import GroupRewards, group_rewards_registry
# or
from redaccel.verl.rewards.std.base import GRPORewards, rewards_registry
@group_rewards_registry.register() # or @rewards_registry.register()
class MyRewardClass(GroupRewards): # or GRPORewards
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.local_step = 0
def __call__(self, prompts, completions, solutions, **kwargs) -> List[dict]:
# Access: kwargs.get("extra_info")
# Returns: [(score, info_dict), ...]
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 250 lines · 65 tokens per session scan A fc99c49f171a
redaccel-to-relax is a skill published in the GitHub repository redai-infra/Relax (580 stars, last pushed 14d ago), licensed Apache-2.0. It adds 65 tokens to every session and 2,600 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
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.
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.
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.
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.
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.
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.