onmax

onmax / unagent

Public

Unified primitives for AI coding agents.

18
2
100% credibility
Found Feb 05, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

unagent is a library for detecting AI coding agents and managing markdown-based skill files by discovering, installing, and tracking them from local paths or Git sources.

How It Works

1
🔍 Discover unagent

You hear about unagent, a friendly helper that lets you add special skills to your AI coding buddy to make it smarter.

2
📦 Bring it home

You easily add unagent to your setup so it can watch over your AI tools.

3
🧠 Spot your AI pal

unagent automatically recognizes which AI coding assistant you're using and gets ready to help.

4
📁 Hunt for skills

You look in folders or online places for simple skill files that teach your AI new tricks.

5
🛠️ Add the skills

You pick your favorite skills and tell unagent to install them safely with a record of what you added.

🚀 AI supercharged!

Your coding assistant now knows the new skills and helps you create amazing things faster.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 18 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 unagent?

Unagent provides unified TypeScript primitives for AI coding agents, letting you detect 40+ agents via env vars, discover markdown skills, and install them from GitHub repos or local paths with lockfile hashing. It delivers consistent APIs for queues (Cloudflare, Vercel, QStash), sandboxes (Deno, Vercel), vectors (Pinecone, Upstash, pgvector), workflows, and tasks—abstracting provider differences so agents run portably. Developers get agentic building blocks without rewriting for each platform.

Why is it gaining traction?

It auto-detects runtimes and agents, plus a Nitro playground for testing unified queue sends, sandbox execs, vector searches, and workflow starts across providers. Stands out for seamless provider swaps in agentic flows, like github unified remote ops or unified gaussian primitives for embeddings, sidestepping fragmented SDKs. Early hook: zero-config skills for unagented submissions to agentic apps.

Who should use this?

AI agent devs building prototypes on Cloudflare Workers or Vercel, needing portable sandboxes for safe code runs or unified vector stores for RAG queries. Backend teams handling github unified logs, appsync unified github pipelines, or bitwarden unified github auth in multi-provider setups.

Verdict

Promising for agentic stacks with strong TypeScript docs and playground, but 15 stars and 1.0% credibility signal early-stage—rigorously test before prod. Maturity lags in adoption; pair with established providers for reliability.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.