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

@jonathanleonel/smh

v0.1.1

Published

Spec-minded development runtime for Claude Code

Readme

spec-minded-harness

A spec-driven development runtime for Claude Code.

The problem

Coding agents are powerful but operationally chaotic. They drift, over-engineer, and lose context between sessions. The usual fix: better prompts. Only goes so far.

The real fix is structure.

The philosophy

  • Humans define intent. Agents execute bounded work.
  • A spec is a contract, not a suggestion.
  • The orchestrator thinks. The executor codes. Never mix them.
  • Human checkpoints are a feature, not a limitation.

How it works

Two Claude Code windows. Two distinct roles.

Orchestrator window ({project}-orch/)
Product thinking, backlog and spec refinement. Talks to Trello. Produces specs.

Executor window ({project}-code/)
Reads the spec. Opens a branch. Writes code. Runs tests. Pushes.

Everything generative runs inside Claude. The only shell work is git.

State machine

BACKLOG → REFINED → SPEC READY → CODING → LOCAL REVIEW → PR → DONE
                                                               ↑
                                               FAILURE ← (from anywhere)

| State | Owner | | ------------ | ---------------------- | | BACKLOG | Human | | REFINED | Human + Orchestrator | | SPEC READY | Human | | CODING | Executor agent | | LOCAL REVIEW | Human (local testing) | | PR | Human / async reviewer | | DONE | Human | | FAILURE | Human (any stage) |

Actions

| Command | What happens | | ------------------------------ | ----------------------------------------------------- | | setup-trello <board-url> | Configure Trello board, create missing columns | | setup-orch-repo (optional) | Publish the orch directory to a private GitHub repo | | setup-code-repo [github-url] | Wire up the code repo — clone existing or create new | | add | Create a card with just a title in BACKLOG | | refine | Discuss and populate the first card (or a given ID) | | spec | Write the formal spec to {code-repo}/specs/ | | execute | Prepare executor handoff — branch, spec path, context | | push | Commit, push, open PR | | success | Move card to DONE | | failure | Move card to FAILED, keep branch, log a learning card |

Setup

  1. Use this repo as a template on GitHub
  2. Run the init script from your repos folder:
    spec-minded-harness/scripts/init.sh my-project
  3. Open my-project/my-project-orch/ in Claude Code
  4. Run setup-trello https://trello.com/b/xxxxx/my-board — configures the board ID and creates any missing columns
  5. (Optional) Run setup-orch-repo to publish the orch directory to a private GitHub repo
  6. Run setup-code-repo to create the code repo locally and on GitHub, or setup-code-repo <url> to wire up an existing repo
  7. Done — start with add

Structure

spec-minded-harness/       ← this repo (public template)
{project}-orch/            ← your orchestrator instance (private)
{project}-code/            ← your actual project (existing or new)

Ideas worth building

Provider-agnostic state management
Abstract the Trello integration behind an interface so state management can be swapped for Jira, SQLite, JSON files, or any other backend without changing the workflow.

Eval suite
Automated end-to-end tests that simulate the full workflow — happy path and failure path — using the Anthropic SDK. Validates that agent behavior matches the spec on every change to the templates.

Mutation testing
Include mutation testing as part of the execute quality gates, so the harness validates test quality, not just coverage.

Git worktree isolation
Run each task in an isolated git worktree, preventing side effects between concurrent executions and making rollbacks trivial.

Run persistence
Log each run: prompts, outputs, diffs, decisions. For debugging, retrospective analysis, and future eval training.

Spec templates
Predefined spec templates for common task types: bug fix, new feature, refactor, API endpoint. Reduces the time from refine to spec.

TDD-first execution mode
When enabled, the executor writes failing tests first and pauses for human approval before coding. The implementation loop runs autonomously until all tests pass — no intermediate commits in red. The human only sees the green result.

npx create-spec-minded
Publish as an npm package so bootstrapping a new project is a single command without needing to clone the repo manually.


Inspired by real-world spec-driven workflows.