Ready-to-use AI agents for AgentsKit. Copy the source into your project — you own the code. No new framework dependency, no lock-in.
→ Browse agents at registry.agentskit.io
Each agent is a small, self-contained folder that wires published @agentskit/*
packages (a skill + tools + the runtime) into a one-call factory. The CLI copies
that source into your project — shadcn-style — so you can read it, edit it, and keep it.
npx agentskit add research # copies the research agent into ./agents/research/
npx agentskit add code-reviewimport { openai } from '@agentskit/adapters'
import { createResearchAgent } from './agents/research/agent'
const agent = createResearchAgent({ adapter: openai({ apiKey: process.env.OPENAI_API_KEY!, model: 'gpt-4o' }) })
const { content } = await agent.run('What changed in the EU AI Act in 2025?')| Agent | Category | What it does |
|---|---|---|
research |
research | Citation-first web research |
code-review |
coding | Deep, low-noise review: 7 lenses + adversarial verify → typed findings & patches; git diff / PR / files; Markdown / SARIF / PR comments; blocking CI gate |
knowledge-promoter |
ops | Turns curated private notes into public docs PRs: classify → sanitize → adversarial leak-gate → draft PR (never merges) |
More landing weekly. Contribute one →
- Framework — the OSS agent toolkit these agents are built on
- Playbook — engineering standards
- Registry — you are here
- AKOS — the enterprise OS for agents (orchestration, egress, RBAC)
MIT