Skip to content
Back to skills

Build Judgment List

ASecurity

Build graded relevance judgments (explicit or click-derived) and an offline harness before tuning. Reach for this when there's no eval.

  • 7 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 23, 2026
ai-agents

Works with

  • cli

Security analysis

A100/100

Scanned September 23, 2026

npx -y skills add mcorbett51090/RavenClaude --skill build-judgment-list --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Build Judgment List?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Build Judgment List
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/mcorbett51090-build-judgment-list/badge)](https://www.skillsdirectory.com/skills/mcorbett51090-build-judgment-list)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: build-judgment-list
description: "Build graded relevance judgments (explicit or click-derived) and an offline harness before tuning. Reach for this when there's no eval."
---

# Skill: Build judgment list

Tuning without a judgment list is guessing dressed as engineering (§3 #3).

## Step 1 — Sample the query mix
Representative queries weighted by real traffic (§3 #7).

## Step 2 — Grade relevance
Explicit graded labels or click-derived judgments, with position-bias caution (§3 #3 #6).

## Step 3 — Build the offline harness
Reusable NDCG/MRR/precision@k harness over the judgment list (§3 #3).

## Step 4 — Set the baseline
The current ranking's metrics — the bar every change must beat (§3 #1).

## Output
A graded judgment list and an offline harness with a recorded baseline.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…