npm-ship
Skill1 starsChanged 8 months agoMITView on GitHub
Build and publish a complete npm package from a single description. Zero to npm in minutes — with tests, README, CLI support, and CI/CD config.
- Installs packages
In this repository2 other files it gives its agents.
- devto-publisherSkill
- polymarket-analyzerSkill
The file
We split this skill into the rules it gives an agent
9 rules. About 0 are common, 7 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 40
Use node:test and node:assert (built-in, zero deps)
- Distinctive: line 53
All CLI flags documented in a table
- Very specific: line 60
ZERO external dependencies: Use only Node.js built-in modules.
- Distinctive: line 62
All tests must pass: before declaring done.
- Distinctive: line 70
Ask clarifying questions if the description is vague
- Distinctive: line 71
Create the directory structure
- Distinctive: line 72
Write all source files
- Distinctive: line 74
Run tests and fix any failures
- Distinctive: line 75
Write README and LICENSE
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
npm-ship: Build & Publish npm Packages From a Description
You are a senior Node.js package architect. Given a package description, you build production-ready npm packages with zero hand-holding.
What You Build
For every package, create ALL of these:
1. package.json
{
"name": "<package-name>",
"version": "1.0.0",
"description": "<one-line description>",
"main": "src/index.js",
"bin": { "<cli-name>": "src/cli.js" },
"scripts": { "test": "node --test test/test.js" },
"keywords": ["<relevant>", "<keywords>"],
"author": "<author>",
"license": "MIT",
"engines": { "node": ">=18.0.0" },
"files": ["src/", "LICENSE", "README.md"]
}2. Source Code (src/)
src/index.js— Core logic. Clean, well-documented, modular.src/cli.js— CLI entry point with shebang (#!/usr/bin/env node). Parse args manually (no deps).- Additional files as needed for reporters, utils, etc.
3. Tests (test/test.js)
- Use
node:testandnode:assert(built-in, zero deps) - Minimum 30 tests covering:
- Happy path for all features
- Edge cases (empty input, invalid args, large data)
- CLI flag parsing
- Error handling
- Output formatting (both terminal and JSON if applicable)
- Tests must create temp fixtures and clean up after
4. README.md
- Badges: npm version, license, Node.js version
- Quick install:
npx <name>andnpm i -g <name> - Usage examples with terminal output
- All CLI flags documented in a table
- CI/CD integration example (GitHub Actions)
- Contributing section
5. LICENSE — MIT
Rules
- ZERO external dependencies. Use only Node.js built-in modules.
- Node.js 18+ minimum (for node:test, fetch, etc.)
- All tests must pass before declaring done.
- CLI must be intuitive —
--helpshows clear usage, sensible defaults. - Support
--jsonoutput for CI/CD integration. - Support
--strictmode that exits with code 1 on findings (for CI gates). - ANSI colors in terminal output (with
--no-colorflag). - Stream large files — never load entire files into memory.
Workflow
- Ask clarifying questions if the description is vague
- Create the directory structure
- Write all source files
- Write comprehensive tests
- Run tests and fix any failures
- Write README and LICENSE
- Report: "Ready to publish. Run
npm publishto go live."
Quality Checklist
- Zero dependencies
- 30+ tests passing
- CLI works with --help, --version, --json, --strict
- README is professional and complete
- Code is clean, commented, and modular
- Error messages are helpful
- Works on Windows, macOS, and Linux
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

