Skip to content
BranchesTags

Repository files navigation

AREX-Skill

Turn Repos & Papers into Skills for Autonomous ML Research

An open library of 5,000+ verified, executable skills distilled from 1,000 ML repositories — and the agent that builds them.

AREX-Skill Library: 5000+ skills 1000 ML repositories DisCo CLI v0.2.0 License: Apache 2.0

English | 简体中文 

---
config:
  fontSize: 18
---
flowchart LR
    subgraph SRC["📚 Sources"]
        direction TB
        s0["🌐 Tens of thousands<br>of ML repos"] == "curate" ==> s1["⭐ 1,000 top repos<br>+ 📄 papers · ✍️ blogs"]
    end
    subgraph CRE["🤖 DisCo Creator Agent"]
        direction TB
        c1["🔍 Discover<br><i>capabilities</i>"] --> c2["🧠 Distill<br><i>into skills</i>"] --> c3["🧪 Verify<br><i>by execution</i>"]
        c3 -. refine .-> c2
    end
    subgraph LIB["🧠 AREX Skill Library"]
        direction TB
        l1["🧭 One router<br><i>routes any ML task</i>"] ~~~ l2["📖 5,000+ verified skills<br><i>20 areas · 178 families</i>"] ~~~ l3["🛠 Task-oriented skills<br><i>built per task</i>"]
    end
    subgraph FIN["🧑‍💻 Your Agent — unchanged"]
        direction TB
        u["Claude Code · Codex · DisCo<br><i>loads only what<br>the task needs</i>"] --> r["🔬 <b>Autonomous<br>ML research</b><br>🔓 new abilities<br>🏆 better results<br>⚡ fewer tokens"]
    end
    SRC ==> CRE ==> LIB ==> FIN
Loading

Same agent. Same budget. 2.3× the wins. MLE-bench sets agents loose on 75 Kaggle ML competitions. With AREX skills, vanilla Codex goes from winning medals in 31% of them to 73% — outscoring every public leaderboard entry. Skills, not agent engineering.

🧭 Table of Contents

📣 News

  • 2026-08-27: The library scales to 1,000 repositories and 5,000+ skills, with a rebuilt router covering every repo. Technical report coming soon, with full benchmark results on MLE-bench, PaperBench, Frontier-CS, and PassNet.
  • 2026-08-03: AREX-Skill launches with DisCo's Creator and Researcher workflows and the initial AREX-Skill Library release: 1,000+ operating skills for 170 widely used repositories.

💡 Why AREX-Skill

Research knowledge is everywhere. Agents still can't use it well.

📄 Papers 💻 Codebases 🤖 Agents
Explain why things work Contain working implementations Still have to rediscover everything

Papers, repositories, and blogs hold nearly all of the field's know-how — but they are written for human readers. They are loose, heterogeneous, and expose no interface an agent can call. So on every task, an agent burns its context window and execution budget searching, reading, and trial-and-erroring its way back to knowledge the field already wrote down.

We call the missing layer operating knowledge: the know-how that separates knowing a method from making it work. AREX-Skill pre-compiles it, once, into skills.

🧠 From Knowledge to Skills

We don't summarize repositories. We compile them into skills agents can execute.

---
config:
  fontSize: 18
---
flowchart TB
    subgraph D["📚 Descriptive Knowledge — written for humans"]
        direction LR
        d1["📄 <b>Papers</b><br><i>methods & why they work —<br>but no runnable path</i>"] ~~~ d2["💻 <b>Repos</b><br><i>working code —<br>but usage stays implicit</i>"] ~~~ d3["✍️ <b>Blogs</b><br><i>tricks & pitfalls —<br>scattered, unverified</i>"]
    end
    D == "⚗️ <b>Skill Distillation</b> — extract · operationalize · verify" ==> O
    subgraph O["🛠 Operational Knowledge — built for agents"]
        direction LR
        o1["🎯 <b>When to use</b><br><i>applicability conditions<br>the router can match</i>"] ~~~ o2["📋 <b>How to use</b><br><i>step-by-step workflows<br>with expected behavior</i>"] ~~~ o3["▶️ <b>What to run</b><br><i>commands, scripts,<br>ready-made tools</i>"]
        o4["✅ <b>How to validate</b><br><i>checks & expected<br>observations</i>"] ~~~ o5["🚑 <b>How to recover</b><br><i>known failures with<br>fixes attached</i>"] ~~~ o6["📎 <b>What to trust</b><br><i>evidence linked back<br>to the source</i>"]
    end
Loading

The output is not a summary and not a RAG index. It is Knowledge → Capability: a skill declares its use conditions, execution behavior, supporting evidence, validation steps, and failure handling — everything an agent needs to act, verify, and recover without re-deriving the source.

🧩 What Is an AREX Skill

An AREX Skill is a self-contained, agent-readable unit of operating knowledge. It follows the open Agent Skills format and is organized around SKILL.md, with optional supporting resources:

AREX Skill
│
├── 🧭 SKILL.md: Entry Point / Router
│   ├── applicability and scope
│   ├── task routing
│   ├── operating workflows
│   └── validation and troubleshooting
│
├── 📖 references/: Focused Instructions
│   ├── installation and configuration
│   ├── detailed workflows
│   └── troubleshooting and provenance
│
└── 🛠 scripts/: Executable Helpers
    ├── diagnostics and smoke tests
    ├── workflow utilities
    └── compatibility and reusable helpers

SKILL.md is the entry point for the skill's scope, operating instructions, and—when applicable—task routing. references/ and scripts/ are optional: references provide focused guidance and provenance, while scripts provide executable helpers for diagnostics, workflows, and repeatable checks.

When a repository exposes several workflows, its skills can be organized as a repository skill graph. A root skill routes the task to focused sub-skills:

Repository Skill Graph
│
├── root SKILL.md: entry point and router
└── sub-skills/
    ├── inference/SKILL.md
    ├── training/SKILL.md
    └── evaluation/SKILL.md

Each sub-skill is an individual AREX Skill and may have its own references and scripts. The links from the root to its sub-skills form a skill graph—an AREX-Skill extension that supports progressive disclosure, so the agent reads only the branch required by the task.

🧭 Router Design

The published repository collection currently uses a router that is itself an AREX Skill: repo-skills-router/SKILL.md. It describes available capabilities and links to repository skill graphs; it is not a separate black-box routing service or a required proprietary runtime.

The agent is expected to follow a progressive disclosure strategy: read the router first, identify the relevant repository or workflow, then load only the matching skill branch and its references/ or scripts/ when needed. This keeps the initial context focused while allowing deeper instructions and executable helpers to be loaded for the task at hand.

This SKILL.md router is the current portable baseline, not the only possible design. Depending on the agent runtime, latency, cost, privacy, and control requirements, users can build other routing layers, for example:

  1. Query subagent. Give the main agent a dedicated subagent that searches the skill library, selects candidate skills, and returns skill paths or a loading plan for the main agent to review and execute.
  2. Retrieval tool. Expose a tool backed by an embedding model, reranking model, or another retrieval component. The main agent submits the task, receives ranked skill candidates, and then loads the relevant graph branches.

Any alternative should preserve clear scope, progressive loading, provenance, and verification boundaries. AREX provides a portable SKILL.md router baseline, while users remain free to design the routing architecture that best fits their agent.

⚗️ How AREX-Skill Is Built: DisCo

The AREX-Skill Library is built through DisCo's discover, distill, and verify workflow.

---
config:
  fontSize: 18
---
flowchart LR
    S["📦 repo · paper · blog"] --> A["🔍 <b>Discover</b><br><i>map what the source<br>can actually do</i>"]
    A --> B["🧠 <b>Distill</b><br><i>write skills with evidence,<br>checks & recovery paths</i>"]
    B --> C["🧪 <b>Validate</b><br><i>execute examples & tests<br>in a real environment</i>"]
    C == "✅ passed" ==> L["📚 <b>Ship</b><br><i>into the library</i>"]
    C -- "❌ failed" --> E["🔁 <b>Evolve</b><br><i>repair & refine</i>"]
    E --> B
Loading

An ordinary repo-to-doc tool stops at Repo → Documentation. DisCo's Creator agent runs a full experimental loop — evidence-backed exploration, skill-graph generation, then verification with refinement: generated checks and native examples are executed, failures are repaired, and the loop repeats until the graph passes or the budget is spent. Skills ship only after they survive their own tests.

📊 Library Scale

AREX-Skill Library overview with 20 areas, 178 package families, 1,000 repositories, and 5,000-plus skills

This is not a demo. It is a growing AREX-Skill Library covering 20 research areas, 178 package families, 1,000 repositories, and 5,000+ verified skills — from training infrastructure, LLM alignment, and inference serving to robotics, genomics, and scientific computing. The catalog lists every graph with its upstream repository, source commit, and coverage.

🔄 Library Maintenance

Upstream repositories keep changing, so repo skills are maintained rather than treated as permanent snapshots. Our default maintenance target is to refresh the published repo skills about once a month, while major upstream changes, critical bugs, or security issues may trigger an earlier refresh. The monthly cadence is an operating target, not a fixed service-level commitment for every repository.

The community is welcome to refresh individual repo skills at any time. A refresh contribution should use current upstream evidence, record the new source commit and verification steps, and update the router or catalog when routing or coverage changes. Each skill's own SKILL.md license continues to apply to refreshes and contributions. See the Refreshing Repo Skills guide for the DisCo workflow, synchronization checklist, and pull request requirements.

🖼️ Skill Gallery

What a skill looks like in practice — source → skill → one prompt:

Source Skill Ask your agent
🔍 FAISS Vector search & index composition "Optimize this FAISS index for lower latency at recall ≥ 0.95."
⚡ vLLM High-throughput LLM serving "Benchmark vLLM vs SGLang on this model and report verified throughput."
🧠 Unsloth Efficient LLM fine-tuning "Fine-tune Llama on this dataset within 24 GB VRAM."
🔥 Diffusers Diffusion training & inference "Train a LoRA for this style and validate outputs."
🦾 LeRobot Robot learning workflows "Train and evaluate an ACT policy on this manipulation dataset."
🧬 AlphaFold2 Protein structure prediction "Fold these sequences and check confidence metrics."

Every graph follows the same contract: a routed entry skill, focused sub-skills for real workflows (data, training, evaluation, serving, troubleshooting), and validation steps the agent can actually run.

📈 Do Skills Make Agents Better Researchers

We hold everything fixed — Codex harness, GPT-5.5 (xhigh) backbone, same execution budget — and change exactly one thing: whether the agent has AREX-distilled skills.

MLE-bench (medal rate across 75 Kaggle competitions)
  without skills   ███████░░░░░░░░░░░░░░░░  31.1%
  with skills      █████████████████░░░░░░  72.9%   (+134% relative)

PaperBench (replication score, 20 papers)
  without skills   ███████░░░░░░░░░░░░░░░░  29.5
  with skills      █████████░░░░░░░░░░░░░░  39.6    (+34% relative)
Benchmark Metric Codex Codex + AREX-Skill Δ
MLE-bench (full, 75 tasks) Medal rate (Any Medal) % 31.11 72.89 +41.78
PaperBench (full, 20 papers) Replication score 29.45 39.59 +10.14
Frontier-CS (Agent Track, 188 tasks) Score 70.63 77.14 +6.51
PassNet (200 samples) AS Score 1.343 1.531 +14.0%

Highlights from the technical report (release coming soon):

  • Beyond agent engineering. Vanilla Codex + skills tops the strongest public MLE-bench entries (72.89 vs 64.44) — no custom harness, no modified control loop, only distilled operating knowledge.
  • The advantage grows with difficulty. On MLE-bench High-tier tasks the score rises from 13.3% to 62.2% (4.7×); skills matter most exactly where unguided trial-and-error is most expensive.
  • Recovery, not just polish. The largest gains land on tasks where the no-skill agent nearly fails (PaperBench rice: 7.9 → 48.5; Frontier-CS tasks scoring <50 gain +26.6 on average).
  • Efficiency, not brute force. On Frontier-CS, Codex + AREX-Skill Pareto-dominates the leaderboard's Claude Code entries on score, tokens, steps, and tool calls, using ~3× fewer tokens.

💡 Why it helps 

Without AREX                    With AREX
────────────                    ─────────
Search the web                  Route to skill
Inspect the repo                Execute known workflow
Guess an approach               Validate against checks
Debug from scratch              Optimize the target metric
Retry, repeat…

Benchmarks show that performance improves; the operating pattern shows why: the agent enters a productive region of the solution space early and spends its budget on the choices that move the metric. See a full Creator session building a FlagEmbedding graph and a Researcher session applying Gymnasium + Stable-Baselines3 skills to an auditable RL experiment.

🚀 Quick Start

Install the CLI, add the AREX-Skill Library, and open an interactive Researcher session:

# 1. Install the DisCo CLI (Node.js >= 22.19)
npm install -g @auto-ml-skills/disco

# 2. Install the AREX-Skill Library (1,000 repos + router)
disco repo-skills install

# 3. Open the interactive DisCo CLI (Researcher mode)
disco

At the DisCo prompt, enter research tasks directly:

Use the installed skills to benchmark vLLM and SGLang on this machine and report verified throughput for each.

Configure a model provider on first run with /login or environment variables (ANTHROPIC_API_KEY, OPENAI_API_KEY, GEMINI_API_KEY, …). DisCo is an interactive terminal CLI; -p / --print is only the optional non-interactive mode that processes one prompt and exits, which is useful for scripts and automation. Run disco --help to see all CLI commands and options.

Create your own skills (Creator mode)

DisCo's Creator mode distills new skill graphs from any repository or paper, then verifies them before import:

git clone https://github.com/FlagOpen/FlagEmbedding.git
disco --creator

Then enter the workflow request at the DisCo prompt:

/skill:distill-ml-knowledge Create and verify a repository skill graph for ./FlagEmbedding covering embedding inference and evaluation.

Researcher is the default mode; start explicitly with --creator or --researcher, or use /creator · /researcher in the UI. See DisCo Meta Skills for the complete catalog and portable installation guide, and DisCo Workflows for verification gates and maintenance workflows.

For installation and maintenance details, see the Installation Guide. It documents npm and source builds of the DisCo CLI, provider configuration, AREX-Skill Library install/status/update behavior (including conflict handling and recoverable backups), router controls, manual installation, and optional portable Creator meta-skill installation for other agents.

🤖 Works With Your Coding Agent

No new agent. No new workflow.

---
config:
  fontSize: 18
---
flowchart LR
    S["🧠 <b>AREX Skills</b><br><i>plain SKILL.md graphs<br>+ one library router</i>"] ==> G
    subgraph G["your existing agents — workflow unchanged"]
        direction LR
        A["<b>Claude Code</b><br><i>import from DisCo</i>"] ~~~ B["<b>Codex</b><br><i>import from DisCo</i>"] ~~~ C["<b>DisCo</b><br><i>native in<br>Researcher mode</i>"]
    end
Loading

Skills are plain SKILL.md graphs in the emerging agent-skills format. DisCo Researcher loads and routes the AREX-Skill Library natively; the Creator workflow can export selected skills and a scoped router to Codex or Claude Code. No proprietary skill runtime or workflow migration is required. The bundled DisCo CLI manages installation, routing, export, and updates.

To export selected repository skills from DisCo's managed collection into another compatible agent, run the Creator workflow after disco repo-skills install:

disco --creator

For Codex, which writes to ~/.agents/skills/repositories/, enter:

/skill:import-repo-skills-to-agent import vllm and sglang to Codex

For Claude Code, which writes to ~/.claude/skills/repositories/, enter:

/skill:import-repo-skills-to-agent import vllm and sglang to Claude Code

The workflow copies the selected skills into repositories/repo-skills/, builds or merges a scoped repositories/repo-skills-router/, and adds Codex-specific agents/openai.yaml policy files when the target is Codex. Restart the target agent after import so it reloads the new skills. For portable Creator meta skills and manual target setup, see DisCo Meta Skills.

🌐 The Bigger Vision

From repositories to skills. From skills to autonomous research.

Today's research knowledge is written for humans. AREX turns it into operating knowledge for AI researchers.

---
config:
  fontSize: 18
---
flowchart LR
    A["📦 <b>Repos + Papers</b><br><i>written for humans</i>"] --> B["🧠 <b>Research Skills</b><br><i>distilled once, verified</i>"]
    B --> C["🌐 <b>Skill Ecosystem</b><br><i>shared & inherited</i>"]
    C --> D["🤖 <b>Autonomous Agents</b><br><i>start where the field left off</i>"]
    D --> E["⚙️ <b>Automated<br>ML R&D</b>"]
Loading

Every skill distilled once is inherited by every agent afterward. As the library grows — more repositories, more papers, more task families — each research task starts a little further from zero. We believe ML research knowledge should exist not only as papers and repos, but as skills that AI researchers can directly call.

Naming note: AREX-Skill is the project and the published AREX-Skill Library. DisCo is the bundled skill-powered CLI/runtime that creates skills (Creator) and researches with them (Researcher).

🤝 Contributing

We welcome contributions in three areas:

  1. Add repo skills. Contribute a verified skill under skills/repositories/repo-skills/<skill-id>/, then update the sibling router and public catalog so agents can discover it.
  2. Refresh or extend existing skills. Update stale guidance or add deeper coverage using current upstream evidence. The default library cadence is roughly monthly, but focused community refreshes are welcome at any time; keep provenance, verification, license, and routing metadata aligned. See Refreshing Repo Skills for the detailed workflow and PR checklist.
  3. Improve the DisCo CLI. Contribute CLI/runtime or bundled repository and paper workflow changes under cli/.

For skill PRs, include the model and provider, source commit, and verification steps; update the router and catalog when routing changes. See the Contribution Guide for the complete checklist.

📚 Documentation

Page Description
Installation Guide DisCo npm/source installation, provider setup, Library install/status/update behavior, conflict backups, router controls, manual fallback, portable Creator meta-skill installation.
DisCo Workflows Modes, sessions, Researcher execution, Creator construction, deployment scopes.
Refreshing Repo Skills Refresh an existing repository skill with DisCo, synchronize runtime and publication metadata, verify the result, and prepare a contribution PR.
DisCo Meta Skills Complete Creator meta-skill catalog and portable installation into other agents.
AREX-Skill Library Library model, collection layout, installation.
Imported Repo Skills Catalog Every published graph with upstream baselines.
Repository Catalog Human-readable area -> family inventory of all published repository skills.
Architecture Repository layers, routing, authoring pipelines, deployment scopes.
Examples Sanitized end-to-end Creator and Researcher sessions.
Bundled Skills Reference Creator meta-skill contracts and artifact layouts.
DisCo CLI README CLI usage, runtime skill routing, packages.

🙏 Acknowledgement

DisCo's CLI and agent runtime are built on the foundation of earendil-works/pi, an open-source AI agent toolkit with a unified LLM API, agent loop, terminal UI, and coding-agent CLI.

AREX-Skill is also made possible by the open-source community on GitHub. The repository skills in this library build on high-quality ML, agent, data, bio/chemistry, vision, and infrastructure projects released by researchers and engineers around the world. We are grateful to everyone who makes that work available for the community to use and build on.

📄 License

Unless a file or component states otherwise, repository-level AREX-Skill materials are released under the Apache License 2.0.

⚠️ Every skill in the AREX-Skill Library has its own license. Before using, copying, modifying, or redistributing a skill, check the license metadata field in that skill's SKILL.md. The per-skill license is authoritative for that skill, not this repository's Apache-2.0 license. It may contain different terms, additional conditions, or restrictions. You are responsible for reviewing and complying with each skill's license.

The standalone DisCo npm package under cli/ is distributed under its own MIT License, with upstream attribution in cli/THIRD_PARTY_NOTICES.md.

📝 Citation

TBA

About

A Skill Library for Automated Machine Learning

Resources

Readme
Apache-2.0 license

Contributing

Contributing
Activity
Custom properties

Stars

12 stars

Watchers

0 watching

Forks

2 forks
Report repository

Releases

Packages

Contributors

Languages