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.
npx skills add alivirgo/Major-AI-Skills --skill ai-evaluation-datasetgit clone --depth 1 https://github.com/alivirgo/Major-AI-SkillsWrote 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.
[](https://agentmods.dev/skills/alivirgo/major-ai-skills/ai-evaluation-dataset)<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.
<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>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00054 | $0.03732 |
| Opus 5.5 | $0.00022 | $0.01493 |
| Sonnet 5.5 | $0.00011 | $0.00746 |
| Haiku 4.5 | $0.00005 | $0.00373 |
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 21d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 335 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Evaluation Dataset AI Skill Guide (Claude)
Overview & Engine Architecture
An evaluation dataset is a governed test artifact, not a bag of questions. Scores are only comparable when every run pins an immutable dataset revision, a system fingerprint (prompt/model/tools/retriever), and a judge/scorer fingerprint. Most “the new prompt is better” claims are dataset drift, label edits, or train/eval leakage in disguise.
Claude operates as a Principal AI Evaluation Engineer, specializing in three-layer dataset design (golden / regression / production-holdout), group-aware splits, rubric-first labeling, contamination checks, and bridge reports when labels change.
┌─────────────────────────────────────────────────────────────┐
│ Evaluation Dataset Stack │
│ │
│ Layers (version independently) │
│ ├── golden/ 30–100 hand-verified canonical cases │
│ ├── regression/ 200–500 known failures + edge coverage │
│ └── holdout/ rolling production sample (drift) │
│ │
│ Case record │
│ ├── id, input, expected_behavior, forbidden_behavior │
│ ├── rubric / assertions, gold evidence (source+offsets) │
│ └── tags, split, provenance, ACL, created_at │
│ │
│ Run binding │
│ ├── dataset_version + content_hash │
│ ├── system_hash (prompt/model/retriever/tools) │
│ └── judge_hash (rubric/LLM-judge/deterministic scorers) │
└─────────────────────────────────────────────────────────────┘
When to use / when not to
Use when
- You need to compare prompts, models, retrievers, or agent tools honestly.
- Production failures should become regression cases.
- An LLM judge needs a human-calibration subset.
- RAG needs gold evidence labels (passages/offsets), not only answers.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 21d ago Changed · +300 lines · +23 tokens per session 8b2eccfc8d9a
- 1mo ago First seen · 35 lines · 31 tokens per session scan A e5d4ea7c7ac8
ai-evaluation-dataset is a skill published in the GitHub repository alivirgo/Major-AI-Skills (2 stars, last pushed today), licensed MIT. It adds 54 tokens to every session and 3,732 once invoked, about $0.0002 per session on Opus 5.5. 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.
Other skills, from other repositories
advanced-evaluation
Design LLM-as-judge rubrics, direct scores and pairwise comparisons. Use for judge models, evals, position bias and calibration against human labels.
agent-harness-fault-injection
Use when an agent workflow needs deterministic recovery evidence for sandbox, MCP/tool, worker, checkpoint, memory, or orchestration failures.
ai-prompt-regression-testing
Use when comparing prompt changes against a fixed evaluation set: record actual model settings, required checks, missing results, judge limitations and case-level regression evidence.
api-contract-testing
Design and review API contract tests across providers and consumers. Use for schema changes, compatibility regressions and client integration failures.
agent-evals
Build automated evaluation suites for AI agents using golden datasets, rubrics, and regression gates. Use when shipping agent features, validating prompt changes, or gating deployments on quality.
eval-harness
Professional Eval Harness Expert skill. Integrate LLM API workflows, safe system prompt guidelines, and agentic workflows.