wevm

wevm / incur

Public

CLI framework for agents and humans

134
5
100% credibility
Found Feb 28, 2026 at 118 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

incur is a lightweight TypeScript library for creating command-line tools optimized for seamless integration with AI agents through efficient discovery, token-saving output, and automatic call-to-action suggestions.

How It Works

1
👀 Discover incur

You hear about incur, a simple way to build helper programs that smart assistants love to use.

2
📦 Add to your project

Download the building blocks into your work folder so you can start creating.

3
Build your first helper

Write a quick command like 'greet someone' using easy examples that feel like magic.

4
Try it out

Run your new helper and see friendly messages pop up just as you imagined.

5
Share with assistants
Quick share

One command makes lightweight files for fast discovery.

🔌
Full connect

Set up a direct link for powerful, always-ready access.

🚀 Assistants team up

Watch your helpers chain commands together, saving time and getting perfect results every time.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 118 to 134 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is incur?

Incur is a TypeScript CLI framework for building command-line tools that humans and AI agents use seamlessly. Define commands with Zod schemas for args, options, env vars, and output, getting auto-parsing, validation, typed callbacks, and formats like token-efficient TOON or JSON. Run `skills add` or `mcp add` for agent discovery, `--llms` for manifests, and get global flags like `--help` or `--verbose` for free.

Why is it gaining traction?

It crushes token costs—up to 3x session savings over MCP or monolithic skills—via on-demand loading and CTAs that suggest next commands. Light API (create/command/serve) beats verbose alternatives in Go, Python, Rust, JS/Node, with streaming support and zero-boilerplate types. Agents parse outputs easily; humans get crisp help and chaining.

Who should use this?

Node.js/TS devs crafting agent tools, like GitHub Actions runners, repo managers, or token handlers for Copilot/Linux/Ubuntu/Windows. Perfect for CLI authors needing structured I/O without boilerplate, especially GitHub packages integrators tired of manual schemas.

Verdict

Promising for agent CLIs with excellent docs and tests, but 53 stars and 1.0% credibility signal early maturity—prototype here, not prod yet. Grab if you're in JS and want MCP/skills without the bloat.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.