DSH Plugin Evaluation Datasets

English | 中文 | 日本語

A growing collection of evaluation datasets for DSH plugins.

Each dataset is a profile (which metrics to use) and a cases file (test prompts and expected answers). Pick one that fits your plugin, run its cases, and use the results to understand how your plugin behaves.

Start here

1. Browse the datasets.
2. Choose one that matches your plugin and the scenarios you want to cover.
3. Open its profile and cases files.
4. Run the cases against your plugin and review the results.

Need a dataset that is not here yet? Use the AI-assisted authoring guide to draft one, then contribute it.

Datasets

Dataset	Plugin type	Covers	Cases	Metrics
Knowledge Query Basics	knowledge-query	Refunds, shipping, invoices	3	answer-matches-expected, duration

Knowledge Query Basics

A small starting set for plugins that look up clear facts from an installed knowledge source.

• ID: default-v1
• Version: 1.0.0
• Plugin type: knowledge-query
• Covers: refund request windows, standard shipping time, and electronic invoice delivery
• Cases: 3
• Profile: profiles/default-v1.json
• Cases: cases/default-v1.json

Included cases

Case	What it checks	Expected answer
Refund request window	The default deadline for a refund request	30 days
Standard shipping SLA	The promised time for standard shipping	3 business days
Electronic invoice channel	Where an electronic invoice is sent	Email address bound to the order

Metrics

• answer-matches-expected checks whether the final answer matches the expected answer. It decides whether a case passes.
• duration records how long the case takes. It does not change the pass/fail result.

Dataset files

Each dataset has two files:

profiles/<id>.json  Which metrics to use and where to find the cases
cases/<id>.json     Plugin types and test cases

A test case looks like this:

{
  "id": "case-id",
  "title": "A short name for the case",
  "prompt": "The input sent to the plugin",
  "expected": "The answer you expect"
}

Supported metrics

Metric type	Available now	Changes pass/fail
llm_judge	Yes	Yes
observation	Yes	No
tool_trace	Not yet	No
threshold	Not yet	No

Add a dataset

You can contribute a small dataset directly to this repository, or keep a larger dataset in its own repository and add it to the catalog.

• Read CONTRIBUTING.md before opening a pull request.
• Read DATASET_LISTING.md when adding an external dataset.
• Run these checks before submitting:

npm run validate
npm test

Useful links

• AI-assisted dataset authoring
• Contribution guide
• Dataset listing guide
• Governance
• Security policy
• CC0-1.0 license