Independent guide · Not the seller
Claudekit.cc is now Agentkit.best
Agentkit.best (formerly Claudekit.cc) is a paid kit of skills, subagents and slash-command workflows for Claude Code and other coding agents. It suits developers who want a ready-made plan, build, test and review workflow instead of writing one. The old claudekit.cc domain now redirects to agentkit.best.
Affiliate disclosure: if you buy through our link, Claudekit.ai may earn a commission at no extra cost to you. We are an independent community site, not the seller, and not affiliated with Anthropic or Agentkit.best's makers. How we handle affiliates.
Facts checked on agentkit.best on · By Claudekit Team
- Price
- $99 per kit, $149 for both
- Billing
- One-time for the kits
- Refund
- 14-day money-back guarantee
- Works with
- Claude Code, Codex and more
What you get
These are the seller's own figures as listed on agentkit.best on Oct 26, 2026. They change between releases, so check the site before you buy.
- 108+ skills across engineering and marketing.
- 95+ slash commands that chain those skills into a delivery lifecycle, such as
/ak:brainstorm,/ak:plan,/ak:cook,/ak:code-reviewand/ak:ship. - 45 specialized subagents for roles such as planning, coding, testing and review.
- Instant licence-key access and, in the seller's words, "full source code ownership".
Pricing
| Plan | Price | Billing | What it covers |
|---|---|---|---|
| Engineer kit | $99 | One-time | Engineering skills, commands and subagents |
| Marketing kit | $99 | One-time | Marketing skills, commands and subagents |
| Bundle | $149 | One-time | Engineer + Marketing, priority updates, 1-year 3-device desktop app licence |
| Desktop app | $49/yr (1 device) or $99/yr (3 devices) | Annual | Optional companion app |
Prices checked on agentkit.best on . The site advertises a 14-day money-back guarantee. It does not say how long kit updates last, so read the current terms.
Supported tools
| Support level (seller's label) | Agents |
|---|---|
| Production | Claude Code, Codex, Antigravity, Pi, Oh My Pi |
| Beta | Cursor, DeepSeek Harness |
| Preview | OpenCode, Grok, GitHub Copilot |
How we use it
Claudekit.ai is built and run by AI agents with a human editor; you can watch them on Agents at work. The agents that change this site's code run Claude Code with Agentkit.best installed, and three workflows do most of the work:
-
/ak:planturns a request into a written plan with phases, files and acceptance criteria before any code changes. The plans live in our repository next to the code. -
/ak:cookcarries a plan through implementation, testing and code review, handing off between the kit's planner, developer, tester and reviewer subagents. -
/ak:fixhandles bugs: a debugger subagent looks for the cause first, and the fix comes with a test.
Every change ends as a pull request that has to pass our CI checks before it is merged. This page was drafted and built the same way.
Where we do not use it: the content factory that writes our guides and news runs claude -p
in an empty folder with tools switched off, and it does not load the kit. Loading dozens of agents and skills
for every single article would spend tokens on context the writer never uses. That is the main trade-off we
see: a large kit pays off for multi-step engineering work, not for one-shot tasks.
Pros and cons
Pros
- Named commands chain skills and subagents for you, which is the hardest part of a kit to design yourself.
- One-time price for the kits, not a subscription.
- Works across several agents, useful if your team mixes Claude Code and Codex.
- Saves the time it takes to write, test and tune dozens of skills and subagent prompts.
Cons
- It costs money while the building blocks (skills, subagents, hooks, MCP) are free features of Claude Code.
- A large library adds always-on context: Claude Code loads the name and description of every auto-invocable skill each session. It is why we keep the kit out of our writing pipeline.
- Strong opinions about process. If yours differs, expect to override or disable parts of it.
- The desktop app is a separate annual licence unless you buy the Bundle.
Who should not buy it
- You are new to Claude Code. Learn the building blocks first, or a large kit will hide them. Start with what a Claude kit is.
- You need one or two jobs done, such as PR review, release notes or blocking destructive commands. Our free, tested Claude Code kits each do one job, and Superpowers gives you a free plan-first, test-first workflow.
- You want the smallest possible context footprint or a process tied closely to your stack. A DIY kit fits better; our Agentkit.best vs DIY comparison walks through the trade-off.
- You hoped it would replace a Claude plan. It does not: you still pay for the agent and the model you run it with.
Agentkit.best vs DIY and free kits
| Agentkit.best | DIY kit | Superpowers | Claudekit.ai free kits | Official marketplace | |
|---|---|---|---|---|---|
| Price | $99 per kit or $149 bundle, one-time | Free | Free (MIT) | Free (MIT) | Free to install |
| What's inside | 108+ skills, 95+ commands, 45 subagents (seller figures) | Whatever you write | Workflow skills: brainstorming, plans, TDD, debugging, review | 5 single-job kits: Safe Mode, Token Diet, PR Review, Usage Meter, Release Notes | Individual plugins |
| Workflow | Strong, end to end | None until you add one | Strong | One job per kit | Per plugin |
| Runtimes | Claude Code, Codex and others | Claude Code | Claude Code plus many other agents | Claude Code | Claude Code |
| Best for | Full lifecycle on day one | Control and a small footprint | Free engineering discipline | One tested job, no lock-in | Mixing single pieces |
More detail: Agentkit.best vs DIY vs other kits and Claude kit pricing, which also covers theclaudekit.com's subscription kits.
About Claudekit.ai
Claudekit.ai helps people get real work done with Claude. We publish free Claude Code kits with the test level and date, guides and comparisons, and we show what our AI agents are doing on Agents at work. If you would rather have a Claude setup built for you, see Hire us. For the bigger picture, read our Claude kit guide.
Frequently asked questions
Is Agentkit.best the same as Claudekit.cc?
Yes. Claudekit.cc changed its name to Agentkit.best, and the claudekit.cc domain now redirects to agentkit.best. It is the same commercial kit of skills, subagents and workflow commands.
Is Claudekit.ai the same as Claudekit.cc or Agentkit.best?
No. Claudekit.ai is an independent community hub with free kits and guides. We do not make or sell Agentkit.best; we link to it as an affiliate and may earn a commission if you buy.
How much does Agentkit.best cost?
As listed on agentkit.best on Oct 26, 2026: the Engineer or Marketing kit costs $99 one-time, both cost $149 as a bundle, and the optional desktop app is $49 a year for 1 device or $99 a year for 3 devices. The site advertises a 14-day money-back guarantee.
Do I still need a Claude plan if I buy Agentkit.best?
Yes. The kit is a set of skills, subagents and commands that run inside a coding agent such as Claude Code or Codex. You still need a plan or account for the agent and model you run it with.
Is there a free alternative to Agentkit.best?
Yes. You can build your own kit from free Claude Code features, use free workflow libraries such as Superpowers, install single plugins from the official marketplace, or use the free, tested single-job kits on Claudekit.ai.