npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@eval-kit/seed-suite

v0.1.1

Published

Reference eval tasks for research, coding, and support agents — multi-step YAML suites with real distractors.

Vulnerabilities

Powered byPowered by Snyk
  • 0High
  • 0Medium
  • 0Low

Links

Git

Maintainers

akadevvakadevv

Keywords

evalllmagentevaluationresearchcodingsupport

Readme

@eval-kit/seed-suite

Reference eval tasks for research, coding, and support agents — multi-step YAML suites with real distractors. Used by eval-kit — the scoring cockpit for research agents.

npm version License: MIT

Install

npm install @eval-kit/seed-suite
# or
pnpm add @eval-kit/seed-suite

What's in the box

Three suites, each multi-step with real-world distractors:

  • suites/research-agent-v1.yaml — document-grounded research workflows. Tasks include cosmological-claim writing, grammar-audit revisions, future-dated paper searches (the distraction), cover-letter drafting, direct-quote finding, and current-events lookups. Tests whether the agent grounds claims, refuses to fabricate, and catches unverifiable requests.
  • suites/coding-agent-v1.yaml — code-assistance workflows. Architecture preservation, hallucinated APIs, test writing, diff explanation, bug diagnosis, blanket-refactor pushback. Tests whether the agent reads existing code before suggesting changes and resists making sweeping rewrites without justification.
  • suites/support-agent-v1.yaml — customer-support workflows. Ambiguous refund triage, security escalation, de-escalation, policy-gap calibration, multi-issue triage. Tests whether the agent knows when to escalate, when to apologize, and when to refuse.

Every task has a golden_truth per step (what success looks like) and scoring_hints.dimensions (which of the five rubric dimensions apply). Every suite includes at least one distraction task — a request that should be refused or flagged, not satisfied.

Use the suites

After installing, the YAML files are at node_modules/@eval-kit/seed-suite/suites/. Run them via @eval-kit/core:

# from a project that has both packages installed:
npx eval-kit run node_modules/@eval-kit/seed-suite/suites/research-agent-v1.yaml \
  --adapter anthropic --model claude-sonnet-4-5

Or copy a suite into your own suites/ directory and customize:

cp node_modules/@eval-kit/seed-suite/suites/research-agent-v1.yaml suites/my-research.yaml
# edit suites/my-research.yaml to add your own tasks
npx eval-kit run suites/my-research.yaml --adapter anthropic

Programmatic access

import { suitePath } from "@eval-kit/seed-suite";

console.log(suitePath("research-agent-v1"));
// → /abs/path/to/node_modules/@eval-kit/seed-suite/suites/research-agent-v1.yaml

Why ported, not synthetic

Every task in the seed suite is from observed real usage — actual research workflows, actual coding sessions, actual support tickets. Fabricated "plausible-looking" tasks don't make the cut. The point of eval-kit is to measure agent behavior on the work people actually do, not on benchmark fixtures the agent might have memorized.

When adding tasks: prefer porting real workflows over inventing them. See the project brief §10 for the contribution guidelines.

Status

v0.1.0 — three reference suites covering research, coding, and support agents. Stable; new task domains land as new files (e.g. suites/legal-agent-v1.yaml) without versioning the existing suites.

Links

License

MIT