IVDirectory listingPolaris · RA 02h 31m · Dec +89° 15′

Verified Launch

AgentKit

Build your AI engineering system on a battle-tested foundation of agents, skills, and workflows.

Visit Website(opens in a new tab) Back to the directory

Editor's Pick

AgentKit — Build Your AI Engineering System

Battle-tested agents, skills, and workflows so you don't build your AI development system from zero. Keep what works, change what doesn't, make it yours.

Explore AgentKit(opens in a new tab)
Ship faster with AI dev teams — one desktop control center for your coding agents.
Ship faster with AI dev teams — one desktop control center for your coding agents.

Features

  1. Specialized Subagents

    Role-based delegation: dedicated subagents for planning, coding, review, security audits, and deployment.

  2. Chained Skill Graph

    Skills connect into a guarded delivery loop — Brainstorm → Plan → Cook → Review → Ship — with deterministic gates.

  3. Context & Safety Hooks

    Just-in-time guardrails: pre-tool hooks prompt for approval before sensitive operations and inject project rules.

  4. Harness-Native Adapters

    Deep runtime integration with Claude Code, Codex, and Cursor, with native tool definitions and model routing.

Features

  1. 1 · Brainstorm

    Exhaust trade-offs, surface edge cases, and define non-goals before touching code.

  2. 2 · Advise

    Autonomous Fable-tier counsel reframes requirements, surfaces anti-patterns, and scores options.

  3. 3 · Plan

    Generates verifiable phase contracts, dependency DAGs, and interactive artifacts.

  4. 4 · Cook

    Tests-first execution, worktree isolation, and zero-breakage modularization.

  5. 5 · Code Review

    Multi-perspective quality audit, regression tests, and security scans.

  6. 6 · Ship

    Conventional commit splitting, PR verification, and CI convergence watch.

Technology stack

Production Harness
  • Claude Code
  • Codex
  • Cursor
Core Runtime
  • TypeScript
Delivery Engine
  • Skill Graph
Advisory (Fable 5)
  • Kongming Supervisor

Pricing

  • AgentKit Engineer

    $99/one-time

    • 60+ engineering skills, 30+ workflows
    • TDD planning & execution
    • Database & API development
    • Evidence-based code review
    • Full source ownership
    Explore Engineer(opens in a new tab)

  • AgentKit Bundle

    Popular

    $149/one-time

    • Engineer + Marketing (save $49)
    • All 108+ skills & 45 subagents
    • 3-device Desktop App license
    • Priority workflow updates
    • Discord community access
    Get the Bundle(opens in a new tab)

  • AgentKit Marketing

    $99/one-time

    • 12+ MCP integrations
    • Automated research & briefs
    • Programmatic SEO & content pipeline
    • Competitor & market intelligence
    • Multi-channel copywriting agents
    Explore Marketing(opens in a new tab)
“I really like its structured process with specialized agents. The /plan and /cook features are very impressive.”
AgentKit builder
From the AgentKit community

Questions

Is AgentKit an online course?

No. AgentKit is a battle-tested foundation of agents, skills, and workflows for coding agents — scaffolding you own and customize, not a course.

Which coding agents are supported?

Production-native support for Claude Code, Codex, and Cursor. Additional harnesses are in experimental preview or on the roadmap.

Do I keep ownership?

Yes — full source ownership. Keep what works, change what doesn't, replace or delete anything. No vendor lock-in.

Don't build your AI system from zero. Make it yours.

Start from a battle-tested foundation and adapt every agent, skill, and workflow to your team.

Get Started with AgentKit(opens in a new tab)