Every premium and members-only resource for $9.99/month, as a no-commitment subscription. See the plans

PalateKit

Sign in
Menu

Request a resource

Describe the resource you are missing in one sentence. We will email you when it is published.

0 / 280

A page, a video or an image online.

Unlock PalateKit

Unlock everything

Plan

Paying means accepting our terms of sale.

Secure payment by Stripe. Receipts and invoices sent by email.

Welcome back

Sign in to install, download or save this resource. A free account is enough for free resources.

Building the evaluation harness before any fine-tuning v1.0.0

Use this resource to build an evaluation harness before starting a fine-tune

New

  • 572 downloads
  • 202 likes
  • 166 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install fine-tuning-eval-harness

With no option, the skill goes to ~/.agents/skills/; add --agent claude-code (or codex, cursor…) to target an agent.

First time: connect the CLI with

npx --yes palatekit@latest login

The agent builds a golden set from traces or a task spec, writes graders per failure mode and calibrates the judge against human labels. It also measures the base model to serve as the comparison point.

Mostly picked by

Updated November 2, 2026

Card for the “Building the evaluation harness before any fine-tuning” skill

Details

Works in
Install guides
Generates with
Your agent's model
Licence
MIT

Similar resources

View all

What you get

  • SKILL.md: entry gate, eval directory layout and phase 0 checklist
  • references/grader-templates.md: grader templates
  • references/judge-calibration.md: judge calibration against human labels

Good fit for

  • You start a fine-tune with no reliable measure
  • You want to turn traces into an evaluation set

Not for

    Tags

    • evaluation
    • graders
    • golden set
    • fine-tuning

    About

    You say

    “Build the evaluation harness for our support assistant from these 200 traces.”

    Card for the “Building the evaluation harness before any fine-tuning” skill

    Skill published by wshobson under the MIT license, delivered in its original version.

    Show more

    How to use it

    1. Install the skill in your agent's skills folder
    2. Ask your agent to build an evaluation harness before starting a fine-tune
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1.1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/grader-templates.md10.2 KB
    • references/judge-calibration.md6.6 KB
    • SKILL.md8 KB

    Origin and licence

    Listed unmodified

    Author
    wshobson
    Source
    github.com/wshobson/agents/tree/0874a27b17a22e2b7c2e2bcd725768e0803cad06/plugins/llm-finetuning/skills/eval-harness-first
    Licence
    MIT

    Notice

    « eval-harness-first » (wshobson, MIT) livré dans sa version d'origine : fiche en français et en anglais par PalateKit.

    User reviews

    –

    No ratings yet

    Loading reviews…

    Delete your review?

    Your rating and comment will be removed from the page.