SkillFoxx

distil labs Agent Skill

A skill guiding an agent to build a small task-specific model on the distil labs platform, from traces to training.

  • 189 ★ on the repository
  • Last commit
  • Apache-2.0
Skill

Extended access

Why this level

  • The skill sends data and traces to the distil labs platform, starts paid training operations and deploys endpoints, but confirms cost before each stage.
All reasons and checks

distil-labs/distil-cli-skillAuthor:distil-labs

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add skills/distil-labs-agent-skill

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • The skill sends data and traces to the distil labs platform, starts paid training operations and deploys endpoints, but confirms cost before each stage.

This is third-party code. Review the repository files before installing.

What it does

distil labs Agent Skill helps an agent train a small model for a narrow task: intent detection in support messages, invoice and contract parsing, API call routing. Every stage runs through the `distil` CLI or the REST API on the distil labs platform. The agent starts from the model you already run in production: it puts an endpoint in front to collect traces, or takes ready trace, train and test files. The order is building a test set, picking a teacher by evaluation, relabelling traces, generating synthetic data, training the student model and deploying behind a new endpoint. Before each stage the agent confirms settings and expected cost. An autoresearch mode lets the agent iterate on the model itself.

Who it is for. Teams that want to replace an expensive large model in production with a compact one trained for a single task and hand data preparation to an agent.

Good fit when

  • An intent classifier in support runs a large model where a small one would do.
  • Train and test sets must be built from production traces.
  • A student model for invoice and contract parsing is trained from a teacher's answers.
  • The agent runs model improvement loops itself in autoresearch mode.
  • A finished model is deployed locally or behind a platform endpoint.

Not a fit when

  • You need to train on your own infrastructure; every stage runs on the distil labs platform.
  • You have no distil labs account, or you are on Windows without WSL; the CLI runs inside WSL on Windows.
  • You need a general pretrained chatbot rather than a single-task model.

Example request

Help me build a model that labels the intent of customer messages, starting from the traces of our current model.

Limitations

Every stage runs on the distil labs platform, so an account is needed: sign in with `distil auth` or register with `distil signup`. Before each run the agent states the expected cost, and paid operations spend your account funds. API mode needs `pip install requests pyyaml` and tokens from `distil access-token`. The CLI installer is downloaded and run through `curl | sh`, so trust the source. `distil skill install` needs git and puts the skill in `~/.agents/skills`. On Windows the CLI works only inside WSL.

How to disable. Run `distil skill uninstall`. If you installed with `npx skills`, delete the skill folder from the agent's skills directory. Models and data on the platform are deleted separately.

FAQ

How do I install distil labs Agent Skill?

Run `curl -fsSL https://cli-assets.distillabs.ai/install.sh | sh`, then `distil skill install`. Without that CLI use `npx skills add distil-labs/distil-cli-skill`.

Which agents does it work with?

Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and any agent with the Agent Skills format. `distil skill install` puts it in the shared `~/.agents/skills` folder.

Which task types are supported?

Question answering (extraction and classification), chat completion and chat with a tool-calling loop.

Is an account required?

Yes. Sign in with `distil auth` or register with `distil signup`; the skill itself offers this at the start of a project.

What is autoresearch?

A second workflow in which the agent itself iterates on the model instead of building it in a single pass.

Discussion

No questions or reviews yet.

Skills

A tool that builds tested skills from a real workflow and governs their lifecycle through your own marketplace

PluginExtended access2.4K
Needs a VPN

A CLI for training agents: your agent works on a task, and a mentor model reviews the result and sends back remarks

CLIBroad access1.6K

One agent launches others in the background: Claude Code, Codex, Copilot CLI, Grok Build and Pi, with a model chosen per stage.

SkillBroad access119

A skill registry with a CLI and API: the agent describes a task and gets a recommendation with a trust score and install command

CLIExtended access371
Editors’ pick

A skill that distills chats, documents and public sources into a reusable person profile for AI agents and compatible bots

SkillExtended access26.5K

A CLI that detects your project's stack and installs matching agent skills from a vetted registry in one command

CLIExtended access7K

README in short

The docs split work into stages, workflows and references: separate files for each stage, two workflows, data preparation and CLI commands. Swapping the CLI for the REST API changes only the commands, while the stages stay the same.

README badge

Original SKILL.md

---
name: distil-cli
metadata:
  version: "14.1.0"
description: >
  Use when building or training a model on the distil labs platform end to end: preparing a
  dataset (config.yaml, job_description.json, traces, train and test rows), running or
  analyzing any pipeline stage (building a test set or a train set from traces by relabelling
  and synthetic data generation, teacher evaluation, model training, deployment), iterating
  when a teacher score or student score is too low, or deciding what to run next after a
  stage completes, or iterating on a model on its own until it reaches a target
  (autoresearch). Activate for phrasings like "build a model for X", "train a student model",
  "build a test set from my traces", "run a teacher eval", "generate synthetic data", "the
  score is low, what now", "add more training data", or "keep iterating until the model is as
  good as you can get it". Stages run through the distil CLI or
  the distil labs API. This skill owns the model-building logic on top of it. Also activate for
  distil labs lookups that are not a full build: distil CLI commands (distil auth,
  distil traces, distil dataset, distil teacher-evaluation, distil slm, distil deployment,
  distil inference-endpoint, distil api-keys, distil access-token), the distil labs REST API,
  config.yaml parameters, supported student and teacher models, what an evaluation metric
  means, or moving a seed dataset or training dataset into a Dataset.
---

# Building Models

Building task-specific small language models on the distil labs platform: from production
traffic or files on disk, through a test set, teacher evaluation and a train set, to a
finetuned, evaluated student served behind an inference endpoint.

Before the first stage, install the `distil` CLI and sign in per
`references/execution/README.md` § Set up the CLI, then settle the backend per § Choose the
backend.

## How the Skill Is Organised

The skill separates what to do from how to run it. Each fact has exactly one owning page, and
every other page links to it.

- **Stages** (`stages/`): one unit of pipeline work each, directly invocable. A stage file is a
  step-by-step procedure: what it needs, what to confirm with the user, how to run it, how to

Report a problem