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

maverick-ai-dev-stack

v1.6.0

Published

A vendor-neutral engineering workflow for building with AI using explicit context, scope, verification and human review.

Readme

Maverick AI Dev Stack

Maverick is an open-source, vendor-neutral toolkit for building software with AI in real repositories—without accepting changes you cannot explain, verify, or review.

It is not a Claude, Codex, Cursor, or Copilot framework. It is the workflow layer that gives any coding agent explicit intent, scoped context, verification evidence, security checks, and a human decision before merge.

Spec-driven when you need it. Review-driven always.

Why Maverick?

AI can produce a diff quickly. The hard part is knowing whether that diff belongs in your codebase.

INTENT → CONTEXT → PLAN → BUILD → VERIFY → REVIEW → PR

Maverick helps teams answer the questions that actually matter:

  • What problem are we solving, and what is explicitly out of scope?
  • Which files, architecture rules, and constraints must the agent understand?
  • What checks were run, and what evidence do we have?
  • Did the change affect security, dependencies, observability, or rollback?
  • Who made the final human merge decision?

Features

| Area | What Maverick provides | |---|---| | Task workflow | Lightweight, Standard, Spec Driven, and Strict Review presets | | Context | Repository map, task packets, impact lookup, and focused artifacts | | Verification | Detected test/lint/typecheck/build commands with saved evidence | | Review | Diff, scope, dependency, security, and readiness reports | | Collaboration | Handoffs, proposed learnings, and human review gates | | Governance | Versioned repository policy, change budgets, and sensitive-path rules | | Integrations | Provider-neutral adapters for Codex, Claude Code, Cursor, and generic agents | | Automation | JSON output, GitHub Actions task gates, npm release pipeline, and GitHub Pages docs |

Quick start

Requires Node.js 20 or later.

npx maverick-ai-dev-stack init
npx maverick-ai-dev-stack map
npx maverick-ai-dev-stack task add-pagination --preset standard
npx maverick-ai-dev-stack packet add-pagination --agent codex
npx maverick-ai-dev-stack dashboard --open

Installation

Maverick runs in the repository you want to understand and change. It does not require an account, a cloud service, or a project-wide daemon.

Run once with npx

This is the quickest way to try Maverick and is the recommended option for occasional use:

cd path/to/your-repository
npx maverick-ai-dev-stack init

Install globally

Use a global installation when you want the maverick command available from every repository:

npm install --global maverick-ai-dev-stack
maverick init

Use this checkout while contributing

Clone the repository, install its dependencies, and run the CLI through the development script:

git clone https://github.com/jhonesgoncalves/maverick-ai-dev-stack.git
cd maverick-ai-dev-stack
npm install
npm run dev -- init

Confirm your setup at any time with maverick doctor. maverick init writes .maverick/config.json; task artifacts are written to .maverick/tasks/ by default. Both stay local to the repository and can be reviewed in Git.

Fill the generated task artifacts, use your preferred agent, then collect evidence and complete the review:

npx maverick-ai-dev-stack verify add-pagination
npx maverick-ai-dev-stack review add-pagination --diff
npx maverick-ai-dev-stack security add-pagination
npx maverick-ai-dev-stack readiness add-pagination

Choose the right preset

| Preset | Use it for | Generated workflow | |---|---|---| | lightweight | Small, low-risk, obvious changes | Brief → Context → Build → Verify → Review | | standard | Everyday features and fixes | Brief → Context → Plan → Build → Verify → Review → PR | | spec-driven | Ambiguous, complex, or domain-heavy work | Discover → Spec → Context → Plan → Build → Verify → Review → PR | | strict-review | Security, payments, data, migrations, or shared infrastructure | Brief → Context → Plan → Build → Verify → Security → Review → PR |

standard is the default. Every preset includes human review. Read the preset guide and Spec-Driven Development guide for details.

A complete task flow

# Understand the repository and create a task
maverick map
maverick task payment-change --preset strict-review --risk-doc

# Give a focused packet to any AI coding agent
maverick adapter install codex
maverick packet payment-change --agent codex

# Capture work and verify it
maverick checkpoint payment-change
maverick verify payment-change
maverick guard payment-change
maverick security payment-change
maverick deps payment-change
maverick review payment-change --diff --against main

# Collaborate and prepare the delivery
maverick handoff payment-change --to reviewer
maverick readiness payment-change
maverick pr payment-change

Generated reports stay with the task: EVIDENCE.md, CHECKPOINT.md, SECURITY-REPORT.md, DEPENDENCY-REVIEW.md, REVIEW-FINDINGS.md, SCOPE-GUARD.md, HANDOFF.md, and MERGE-READINESS.md.

CLI highlights

maverick init
maverick task <slug> --preset <preset>
maverick presets [show <preset>]
maverick map
maverick impact <path>
maverick packet|run <task> --agent <name>
maverick verify <task> [--json]
maverick checkpoint <task>
maverick guard <task>
maverick security <task>
maverick deps <task>
maverick review <task> --diff [--against main]
maverick policy init|show|check <task>
maverick handoff <task> --to <owner>
maverick readiness <task>
maverick pr <task>
maverick adapter list|install <name>
maverick ci github <task>
maverick status [--json]
maverick dashboard|dash|lens [--open]

Run maverick help for the current command reference. Most operational commands also support --json, making them useful in CI and other developer tools.

Repository policy and team learning

Create a repository-local policy with a change budget and sensitive paths:

maverick policy init
maverick policy check payment-change

Policies are plain JSON under .maverick/policy.json, so they can be reviewed with the code. Teams can capture proposed reusable practices without silently turning them into agent instructions:

maverick learn payment-change --rule "Run focused tests before requesting review."

AI agent integrations

Maverick ships safe, non-overwriting adapters for:

maverick adapter list
maverick adapter install codex
maverick adapter install claude-code
maverick adapter install cursor
maverick adapter install generic

See the Codex, Claude Code, Cursor, and generic agent integration guides.

Automation, CI, and releases

The repository pipeline runs documentation link checks, lint, type checks, CLI tests, demo tests, package build, package dry-run, documentation build, and GitHub Pages deployment on successful pushes to main.

Generate a task-level GitHub Actions gate when a task needs explicit workflow checks:

maverick ci github payment-change

To publish a release to npm, configure the repository secret NPM_TOKEN, update package.json, and push a matching tag:

npm version 1.0.1
git push origin main --follow-tags

Every non-release commit pushed to main creates a version, a vX.Y.Z tag, and an npm publication automatically. The release workflow follows Conventional Commits: feat: creates a minor release, ! or BREAKING CHANGE: creates a major release, and other commits create a patch release. It publishes with npm provenance using the NPM_TOKEN secret. The release checklist has the complete process.

Documentation

Run the documentation locally with npm run docs:dev.

To preview a screen-safe, fictional Lens project for screenshots or demos, run npm run dash:example. It starts a temporary local dashboard on http://127.0.0.1:4175 and opens it when possible.

Security and privacy

Maverick collects no telemetry by default. Never place secrets, production data, or unauthorized proprietary code into prompts. Treat repository instructions and generated output as untrusted until reviewed, use least-privilege tool permissions, and keep final merge accountability with a human.

Read the full security policy.

Contributing

Contributions are welcome—especially improvements that make AI-assisted engineering more explainable, testable, secure, and accessible across tools.

  1. Read the contribution guide and code of conduct.
  2. Create a focused task and keep your diff narrow.
  3. Run npm run check and npm run docs:build.
  4. Explain compatibility and verification evidence in your pull request.

See the roadmap for directions and the Maverick Lab for workflow experiments.

License

CLI code is released under the MIT License. Existing content and template licensing remain described in LICENSE-CONTENT.