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

@aneuhold/robot-instructions

v0.4.1

Published

Organized markdown instructions for AI coding agents

Readme

robot-instructions

NPM

A collection of organized markdown instructions, skills, and status line scripts for AI coding agents, published to npm as @aneuhold/robot-instructions.

📥 Consuming

Instructions are meant to be referenced directly from a markdown instruction file, skills are meant to be symlinked into a skills directory, and status line scripts are meant to be run by path. None of it is read through code.

See docs for different tools:

Updating

Update every global location the package is installed in, and link its skills:

for dir in ~/.claude; do
  package="$dir/node_modules/@aneuhold/robot-instructions"
  [ -d "$package" ] || continue
  (cd "$dir" && pnpm update --latest @aneuhold/robot-instructions)
  mkdir -p "$dir/skills"
  for skill in "$package"/src/skills/*/; do
    ln -sfn "${skill%/}" "$dir/skills/$(basename "$skill")"
  done
done

Locations without the package installed are skipped. Add further paths to the for list as other global locations come into use. Re-running is safe, since ln -sfn replaces a skill's existing link instead of nesting a new one inside it.

Project scope

Install the package as a dev dependency:

pnpm add -D @aneuhold/robot-instructions

Reference it from the project's robot file by its path under node_modules using it's particular syntax.

# Some overall title

Your normal instructions or caveats to the project instructions

<!-- Always put these at the bottom. Order matters, but the placement does not, so it's better to put
it where it will end up being output in the context window anyway -->

@node_modules/@aneuhold/robot-instructions/src/instructions/lang/typescript.md
@node_modules/@aneuhold/robot-instructions/src/instructions/runtime/node.md

Each referenced file is appended as its own labeled block after the referencing file. It is not spliced in at the @ line, which stays literal where it sits. Files linked into .claude/rules/ are delivered the same way.


User scope

Install the package into a global location, such as ~/.claude and import it from the global instruction file.

cd ~/.claude
pnpm init
pnpm add -D @aneuhold/robot-instructions

Run pnpm init first if ~/.claude/package.json does not exist.

The import then reads:

# Some overall title

Your general instructions

<!-- Leave the references at the bottom -->

@~/.claude/node_modules/@aneuhold/robot-instructions/src/instructions/framework/svelte.md
@~/.claude/node_modules/@aneuhold/robot-instructions/src/instructions/framework/sveltekit.md

~/.claude is not an ancestor of a project directory, so its node_modules stays invisible to project builds.


Status lines

src/statuslines/ holds shell scripts that run the "status line" concept in various robot systems.

  • context-usage.sh shows context-window usage and prompt cache warmth: ctx 27% (54k/1M) · cache warm (42m).
    • The percentage is green below 50, yellow from 50, and red from 80.
    • The cache reads green warm with the whole minutes left before it expires (<1m in the last minute), or red stale once it has expired.

Add the script to ~/.claude/settings.json:

{
  "statusLine": {
    "type": "command",
    "command": "bash ~/.claude/node_modules/@aneuhold/robot-instructions/src/statuslines/context-usage.sh",
    "refreshInterval": 300
  }
}

refreshInterval re-runs the script every 300 seconds on top of the event-driven updates, so the minutes left keep counting down while the session is idle. Claude Code also re-runs the script when the cache expires, so the switch to stale is on time regardless of the interval. See how status lines update.

The cache segment reads the prompt_cache field, which requires Claude Code v2.1.251 or later. It shows cache -- until the session's first API response.


🚀 Publishing

Just bump the version and push it up to main. To do both at once run:

pnpm pushpub

📁 Structure

All published content lives under src/, grouped by kind:

src/
├── instructions/   # Instruction documents
├── skills/         # Skill documents
└── statuslines/    # Status line scripts

Only src/**/* is included in the published package.