How do I install the Areno Add Algorithm skill?

Run npx skillmds@latest add inclusionai/areno-add-algorithm in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Areno Add Algorithm safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Areno Add Algorithm?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Areno Add Algorithm free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Areno Add Algorithm?

inclusionai (@inclusionai) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Areno Add Algorithm by inclusionai — Agent Skill

Name: Areno Add Algorithm
Author: inclusionai

Areno Add Algorithm

Add or modify an AReno algorithm, trainer, loss, advantage calculation, role model, or algorithm-specific configuration. Use for framework-level SFT, DPO, GSPO, GRPO, PPO, or new optimization method development. Do not use merely to run an existing algorithm.

inclusionai e279e01 4 files · 4.0 KB Updated Sep 17, 2026

File contents

Add an AReno Algorithm

Start from AlgorithmSpec and registration in areno/api/algorithms.py; do not add factory branches.

Develop on a dedicated local branch. Commit changes locally, then update the remote GPU checkout by fetching and pulling that branch; never patch source on the remote host. Use ModelScope for any model or dataset references used by the validation workload.

python .agents/skills/areno-add-algorithm/scripts/inspect_algorithms.py

Workflow

Define whether the algorithm is offline, rollout policy-only, or multi-role. Read references/ownership.md.
Specify input records, sequence construction, masks, role models, loss inputs, and metrics before coding.
Add the narrowest config type and preserve public defaults/compatibility.
Put batch/materialization logic in the trainer and tensor mathematics in areno/api/loss_fns/ or advantage helpers.
Register one AlgorithmSpec; load experimental implementations through areno/experimental/ when appropriate.
Add CPU tests for registration, config, masks, exact small-tensor math, and trainer dispatch.
Run the new algorithm end to end for at least two consecutive successful training steps using a real model and representative data. Verify finite losses, metrics, and gradients on both steps. For rollout algorithms, also verify bounded GPU rollout/train logprob consistency.

Completion requires registry discovery, deterministic mathematical tests, evidence from at least two successful end-to-end training steps, and role lifecycle checks where applicable. A one-step smoke train is useful for diagnosis but does not complete algorithm validation.

---name: areno-add-algorithmdescription: Add or modify an AReno algorithm, trainer, loss, advantage calculation, role model, or algorithm-specific configuration. Use for framework-level SFT, DPO, GSPO, GRPO, PPO, or new optimization method development. Do not use merely to run an existing algorithm.---# Add an AReno AlgorithmStart from `AlgorithmSpec` and registration in `areno/api/algorithms.py`; do not add factory branches.Develop on a dedicated local branch. Commit changes locally, then update theremote GPU checkout by fetching and pulling that branch; never patch source onthe remote host. Use ModelScope for any model or dataset references used by thevalidation workload.```bashpython .agents/skills/areno-add-algorithm/scripts/inspect_algorithms.py```## Workflow1. Define whether the algorithm is offline, rollout policy-only, or multi-role. Read [references/ownership.md](references/ownership.md).2. Specify input records, sequence construction, masks, role models, loss inputs, and metrics before coding.3. Add the narrowest config type and preserve public defaults/compatibility.4. Put batch/materialization logic in the trainer and tensor mathematics in `areno/api/loss_fns/` or advantage helpers.5. Register one `AlgorithmSpec`; load experimental implementations through `areno/experimental/` when appropriate.6. Add CPU tests for registration, config, masks, exact small-tensor math, and trainer dispatch.7. Run the new algorithm end to end for at least two consecutive successful training steps using a real model and representative data. Verify finite losses, metrics, and gradients on both steps. For rollout algorithms, also verify bounded GPU rollout/train logprob consistency.Completion requires registry discovery, deterministic mathematical tests, evidence from at least two successful end-to-end training steps, and role lifecycle checks where applicable. A one-step smoke train is useful for diagnosis but does not complete algorithm validation.

Frequently asked questions