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

@mastersof-ai/harness

v0.1.0

Published

Multi-agent runtime built on the Claude Agent SDK

Downloads

34

Readme

@mastersof-ai/harness

Define agents in markdown. Control the entire system prompt. No hidden framework instructions coloring your agent's behavior.

Write an IDENTITY.md, run mastersof-ai, and your agent starts with exactly the context you gave it — nothing more. Persistent memory, built-in tools, sub-agents, sandboxing, and a terminal UI. Powered by your Claude Code subscription via the Claude Agent SDK.

Install

npm install -g @mastersof-ai/harness

Quick Start

mastersof-ai                          # first-run setup → starts default agent
mastersof-ai --agent analyst          # start a specific agent
mastersof-ai --message "hello"        # headless one-shot mode
mastersof-ai --resume                 # resume last session
mastersof-ai create my-agent          # scaffold a new agent
mastersof-ai --list-agents            # list available agents

On first run, ~/.mastersof-ai/ is created with three default agents:

  • cofounder — co-founder template with self-improvement tools (default)
  • assistant — general purpose
  • analyst — research and analysis

Creating Agents

mastersof-ai create my-agent

This creates ~/.mastersof-ai/agents/my-agent/ with a template IDENTITY.md. Edit the identity file to customize your agent's personality, instructions, and behavior.

Three example agents ship in defaults/agents/ (cofounder, assistant, analyst). They're copied to ~/.mastersof-ai/agents/ on first run — use them as templates for your own.

How It Works

  • Identity is markdown. Each agent is defined by an IDENTITY.md file — no code required.
  • Persistent memory. Agents read and write to ~/.mastersof-ai/agents/{name}/memory/. Context survives across sessions.
  • Built-in tools. Memory, workspace (file ops), web search/fetch, shell, task tracking, introspection, and model queries.
  • Sub-agents. Researcher (Sonnet), deep-thinker (Opus), and writer (Opus) handle delegated work in separate contexts.
  • Session management. Named sessions with resume, rename, and history.
  • Config-driven. Optional ~/.mastersof-ai/config.yaml for model selection and tool toggles.
  • Sandbox. Optional --sandbox flag runs the agent inside a bubblewrap container for filesystem isolation.

Configuration

Edit ~/.mastersof-ai/config.yaml:

model: claude-opus-4-6        # default model for all agents
defaultAgent: cofounder        # agent started with no --agent flag

tools:
  memory:
    enabled: true
  workspace:
    enabled: true
  web:
    enabled: true
  shell:
    enabled: true
  tasks:
    enabled: true
  introspection:
    enabled: true
  models:
    enabled: true

TUI Commands

Inside the TUI:

  • /sessions — list recent sessions
  • /resume [name|#N] — resume a session
  • /name <text> — rename current session
  • /new — start a fresh session
  • /quit — exit

Keyboard shortcuts:

  • Enter — send message
  • Ctrl+J — insert newline
  • Ctrl+G — open external editor
  • Escape — interrupt streaming / clear input
  • Ctrl+C (double) — exit

Auth

Uses your Claude Code subscription. No API key needed.

Sandbox

Run any agent in a bubblewrap sandbox for filesystem isolation:

mastersof-ai --agent cofounder --sandbox

The sandbox mounts system directories read-only, gives the agent read-write access to its memory, session state, and a configured project directory, and isolates PID/IPC namespaces. On first use, a default sandbox.json is created in the agent's directory. Edit it to customize mounts, environment variables, and network access.

Requires bwrap to be installed (apt install bubblewrap or equivalent).

Optional Dependencies

  • fd — used by find_files tool (fast file search)
  • rg (ripgrep) — used by grep_files tool (fast content search)

Both are optional. Tools return clear errors if the binaries are missing.

Web Search

Set BRAVE_API_KEY environment variable to enable the web_search tool. web_fetch works without it.

Troubleshooting

  • bubblewrap not found — Install bwrap (apt install bubblewrap or equivalent), or run without --sandbox.
  • API key not set — Web search requires BRAVE_API_KEY to be set in your environment.
  • No agents on first run — Check that ~/.mastersof-ai/agents/ was created and contains agent directories. Re-run mastersof-ai to trigger first-run setup.

License

MIT