ai-evaluation-dataset

ai-evaluation-dataset is a skill for Claude Code from alivirgo/Major-AI-Skills. It costs 31 tokens per session (330 once invoked), scanned A, original, MIT.

A workflow for building a versioned JSONL test dataset for an AI task. JSONL is a text format with one JSON record per line; the dataset can include expected behavior, forbidden behavior, scoring rules, sources, and separate test splits.

In plain words
What is it for?
Use it to create ordinary, edge, ambiguous, and abstention cases for comparing AI behavior against explicit acceptance criteria.
Why use it?
It helps prevent evaluation results from being misleading by separating development examples from held-out cases, removing duplicates and leaks, and recording disagreements and provenance.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: mentions Codex.

Part of the mas-ai-workflows plugin — 10 skills shipped together

Good fit Use it to create ordinary, edge, ambiguous, and abstention cases for comparing AI behavior against explicit acceptance criteria.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/alivirgo/major-ai-skills/ai-evaluation-dataset
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add alivirgo/Major-AI-Skills --skill ai-evaluation-dataset
Clone the repo
git clone --depth 1 https://github.com/alivirgo/Major-AI-Skills

Made for: Claude Code.

Or install mas-ai-workflows, the plugin that ships this one along with the rest of its 10 skills.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for ai-evaluation-dataset

README.md
[![agentmods](https://agentmods.dev/badge/skills/alivirgo/major-ai-skills/ai-evaluation-dataset/github.svg)](https://agentmods.dev/skills/alivirgo/major-ai-skills/ai-evaluation-dataset)
Your own site
<a href="https://agentmods.dev/skills/alivirgo/major-ai-skills/ai-evaluation-dataset"><img src="https://agentmods.dev/badge/skills/alivirgo/major-ai-skills/ai-evaluation-dataset/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for ai-evaluation-dataset

Your own site · 80×15
<a href="https://agentmods.dev/skills/alivirgo/major-ai-skills/ai-evaluation-dataset"><img src="https://agentmods.dev/badge/skills/alivirgo/major-ai-skills/ai-evaluation-dataset.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 31 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 330 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00031 $0.00330
Opus 5 $0.00015 $0.00165
Sonnet 5 $0.00006 $0.00066
Haiku 4.5 $0.00003 $0.00033

Measured today against content hash e5d4ea7c7ac8, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

ai-evaluation-dataset scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/mas-ai-workflows/skills/ai-evaluation-dataset/SKILL.md · 35 lines

What it actually says

Evaluation Dataset

Scope

Ask for the target task, deployment population, known failures, and which errors are unacceptable. Reuse the project's evaluator and data format when present; otherwise propose JSONL records with id, input, expected_behavior, forbidden_behavior, rubric, source, and split.

Procedure

Separate training examples, prompt-development examples, and held-out evaluation cases. Split by originating document, user, or conversation rather than individual rows when rows share information. Keep near-duplicates in the same split. Remove secrets and obtain permission before including private user content.

Checks

Include ordinary cases, boundary cases, ambiguous inputs, and explicit abstention cases. Label expected behavior before viewing candidate model outputs. For subjective outputs, use observable rubric criteria instead of a single preferred phrasing.

Failure Handling

Report case counts by split and slice, provenance, duplicate findings, and unresolved label disagreements. Freeze a dataset version and content hash before comparing models. Never report evaluation accuracy without actually running the evaluator.

Deliverable

Given several paraphrases of one support ticket, keep them in a single split; a random row split would leak the answer.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. today First seen · 35 lines · 31 tokens per session scan A e5d4ea7c7ac8

Subscribe to this mod's changes

ai-evaluation-dataset is a skill published in the GitHub repository alivirgo/Major-AI-Skills (1 stars, last pushed today), licensed MIT. It adds 31 tokens to every session and 330 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-12.