ClaudeKit Review 2026: Are the 5 Vertical Kits Worth It?
Honest ClaudeKit review: 5 kits, 101 commands, 82,197 measured tokens. Pricing, install steps, what each kit does, and who should skip it.
ClaudeKit ships 5 vertical kits — EngineerKit, MarketingKit, VideoKit, SEOKit, and EcomKit — covering 101 commands, 19 skills, and 13 read-only specialist agents across 82,197 measured context tokens. Pricing starts at $14.99/month per kit or $99 one-time per kit (as-shipped, no future updates). The honest answer: if your work spans more than one discipline, the Pro bundle at $29.99/month for any 3 kits is the sharpest deal. If you only write code, a cheaper engineering-only option exists and you should know that before you decide.
We write this as the team that built ClaudeKit, so the obvious bias is in favor. We have tried to keep every claim to verifiable numbers, name the gaps honestly, and point you to alternatives where they are genuinely better.
What Is ClaudeKit and Who Is It For?
ClaudeKit is a Claude Code extension system: you install a kit, and it adds slash commands, auto-loading knowledge skills, and read-only specialist agents to your Claude Code environment. The kits are vertical — each one is built around a job function, not just a technology.
The five kits as of v2 (June 2026):
| Kit | Commands | Skills | Agents | Measured Tokens |
|---|---|---|---|---|
EngineerKit (/eng) | 25 | 4 | 4 | 20,413 |
MarketingKit (/mkt) | 20 | 3 | 2 | 16,714 |
SEOKit (/seo) | 19 | 4 | 2 | 16,004 |
EcomKit (/ecom) | 20 | 3 | 2 | 16,464 |
VideoKit (/video) | 17 | 5 | 3 | 12,602 |
| Total | 101 | 19 | 13 | 82,197 |
Every install prints a token ledger so you see exactly what you are adding to your context. You can recount any time with ck tokens <kit> and run ck doctor if something looks off. That transparency was a deliberate design choice: context is the scarcest resource in a long agent run, and we measured the spread between a lean and a bloated skill at roughly 7x across 500+ community skills. Knowing your number matters.
How Does ClaudeKit Install?
Installation is two commands:
ck auth <your-key>— authenticates against your licenseck install <kit>— pulls the kit to~/.claudeglobally (add--localto scope it to a project)
Alternatively: /plugin marketplace add Madni-Aghadi/claudekit-<kit> from inside Claude Code.
After install, a token ledger prints automatically. ck list shows your active entitlements. ck doctor runs a diagnostic if commands are not resolving. The CLI is claudekits v0.1.3 on npm.
One rule worth knowing: the one-time $99/kit license ships the kit as-is with no future updates baked in. Subscriptions get updates as the kit evolves with the Claude Code platform. That is a real trade-off, not a footnote.
What Does Each Kit Actually Do?
EngineerKit (/eng)
25 commands covering the full software development loop. The flagship is /eng debug: root-cause-first diagnosis that works through the stack before proposing a fix. The daily-8 commands most engineers reach for every day are: catchup, plan, tdd, debug, verify, review, commit, and handoff. /eng ship and /eng fix-issue round out the set for release and issue triage workflows.
Every command ends with evidence — a diff, a verified file, a test result — not a reviewer gate that blocks the loop. This matters because v1's reviewer-gate architecture created friction without catching more bugs; the v2 approach trusts the operator and shows the output.
Full details on the EngineerKit page.
MarketingKit (/mkt)
20 commands for content and growth work. Two flagships: /mkt voice builds a voice file from your real posts so every output sounds like you, and /mkt humanize strips 14 detectable AI tells from a draft. The rest of the set covers the full content lifecycle: mkt-hooks for opening lines, mkt-repurpose (one piece to 5 formats), mkt-thread, mkt-newsletter, mkt-calendar, and mkt-launch for coordinating a product launch sequence.
If you have ever pasted AI copy into a tool that flagged it as AI-generated, /mkt humanize is the command that fixes that before it ships.
Full details on the MarketingKit page.
SEOKit (/seo)
19 commands for search and AI-answer optimization. Two flagships: /seo quick-wins surfaces pages sitting in positions 8-20 with low CTR — the fastest ROI in technical SEO — and /seo citations runs N measurement passes of AI citation rates with confidence intervals. The latter matters because AI Overviews now appear on 48% of Google queries (March 2026, up from 34.5% in December 2025), and knowing whether your pages are cited requires actual measurement, not guessing.
Also in the kit: seo-audit, seo-write, seo-check, seo-pseo for programmatic SEO at scale, and seo-extractable for structuring content so AI engines can lift it cleanly.
Full details on the SEOKit page.
EcomKit (/ecom)
20 commands for store operations. Flagship: /ecom no-sales runs a store triage against AOV-band benchmarks to diagnose conversion problems before you guess at fixes. The kit also covers: ecom-flows (Klaviyo flow templates), ecom-cart-recovery, ecom-amazon for marketplace listings, ecom-margin for contribution margin modeling, ecom-ads for paid acquisition copy, ecom-reviews for structured review mining, and ecom-bfcm for Black Friday / Cyber Monday planning.
AI-driven retail orders grew 15x from January 2025 to January 2026, and AI referral traffic is up 393% year-over-year in Q1 2026. EcomKit is built around that shift — the commands assume you are operating in an environment where AI is already touching the customer journey.
Full details on the EcomKit page.
VideoKit (/video)
17 commands for programmatic video production in Remotion. Flagship: /video clone recreates a reference video's visual style in Remotion and verifies the match. The kit also includes video-make (full video from a brief), video-demo (product demo scripts and assets), video-caption (accurate caption generation), video-data (data visualization in motion), and video-social (social cut-downs from a master).
5 skills and 3 read-only agents make VideoKit the highest agent-density kit in the suite. The agents are researcher/auditor roles — they inspect and report, they do not block or gate.
Full details on the VideoKit page.
How Does ClaudeKit Pricing Stack Up?
The pricing model has four tiers:
- Single kit: $14.99/month. Good if you have one clear job function.
- Pro (any 3 kits, swap 1/cycle): $29.99/month. The sharpest value for most operators who cross disciplines.
- All-Access (all 5 kits): $49.99/month (or $399/year).
- Lifetime per kit: $99 one-time, as-shipped. No future updates. 3 devices per license.
Annual options: $119/year single, $239/year Pro, $399/year All-Access.
Refund window: 14 days (not 30). 3 devices per license.
The honest comparison against the market:
| Option | Price | Token data | Verticals | Updates |
|---|---|---|---|---|
| ClaudeKit single kit | $14.99/mo or $99 lifetime | Yes, measured | 5 kits | Subscription: yes. Lifetime: no |
| ClaudeKit Pro (3 kits) | $29.99/mo | Yes, measured | Mix-and-match | Yes |
| wshobson/agents (free) | $0 | No | Engineering-adjacent | Community |
| claudeskills360 | $39 one-time | No | ~2,350 skills, no curation | No |
| ClaudeKit.cc | $99 one-time | No | Engineering only | No |
The key differentiator is measured token data at install time. No other option in this table publishes per-skill context cost before you run it. For anyone managing long agent sessions, that is not a minor detail.
See the full breakdown on the pricing page.
What Is the v2 Agent Architecture?
The v2 architecture has three component types:
- Commands — slash workflows (
/eng debug,/seo quick-wins, etc.). These are the primary interface. They run a defined workflow and end with evidence: a diff, a file, a report, a verified output. There are no reviewer gates that block completion. - Skills — auto-loading knowledge files. They inject domain context when relevant without requiring explicit invocation. Each one has a measured token cost published at install.
- Read-only specialist agents — reviewer, auditor, and researcher roles. They inspect and report. They do not orchestrate other agents, do not block commands from completing, and do not run Python tools. 13 total across the 5 kits.
What v2 does not have: orchestrator agents, blocking reviewer/quality-gate agents, runnable Python tools, --demo flags, or examples/ folders. The old /marketing, /founder, and /sales namespaces from v1 are gone. FounderKit and SalesKit were shelved; the 5 current kits are the product.
This architecture came out of a real problem: v1's reviewer-gate pattern created friction on every command run without a measurable quality improvement. Commands that end with evidence are faster, more auditable, and easier to build on. For the full reasoning, see the post on why we killed the reviewer gate.
What Are the Real Gaps?
Honest gaps, not spin:
No FounderKit or SalesKit. If your primary job is fundraising, investor relations, pitch decks, or outbound sales sequences, ClaudeKit does not have a kit for that right now. FounderKit and SalesKit were shelved in the v2 rebuild. There is no timeline for when or whether they return.
The $99 lifetime license does not include updates. "As-shipped" means exactly that. If Claude Code's platform changes significantly and commands break, the subscription is what gets patched. For a developer who wants to buy and forget billing, the lifetime option is still a real value — just understand what you are buying.
Engineering-only alternatives are cheaper. If you only write code, a dedicated engineering-only option like ClaudeKit.cc at $99 one-time is a fair competitor. You lose token transparency and vertical coverage, but if you genuinely only need /eng-style commands, the price math favors one-time. We will not argue otherwise.
14-day refund window is shorter than some competitors. It is 14 days, not 30. Factor that into your evaluation timeline.
Who Should Buy ClaudeKit?
A numbered decision framework:
- Buy Pro ($29.99/mo) if your work crosses 2-3 disciplines — say, engineering + SEO, or marketing + ecom. The kit-swap-once-per-cycle feature lets you rotate in VideoKit for a launch quarter without paying for all 5 permanently.
- Buy a single kit ($14.99/mo) if you have one clear job function and want to stay lean. Start with the kit that covers your biggest time sink.
- Buy lifetime ($99/kit) if you have a defined project, a fixed budget, and no interest in subscription billing. Understand you are buying a snapshot.
- Buy All-Access ($49.99/mo) if you are building an agency or running a cross-functional operation where all 5 verticals are active work.
- Skip ClaudeKit entirely if your work is exclusively software engineering and you are comfortable assembling free options. The ranked list of free skill repos is a good starting point, and we wrote it to be honest rather than self-serving.
How Does ClaudeKit Compare to Free Alternatives?
Free is a real option, and we should say that clearly. wshobson/agents ships roughly 192 agents and 84 plugins at no cost. davila7/claude-code-templates is a large free hub. The tradeoffs are: no token transparency, no vertical curation, no install tooling, and no support when something breaks.
The Agent Skills open standard (adopted by 32+ tools as of December 2025) and the skills.sh ecosystem (roughly 90k skills) mean the free pool is large and growing. The value of a curated kit over the free pool is curation density, measured token cost, and command coherence — not access to any capability that is otherwise locked.
For a structured head-to-head, see ClaudeKit kits vs free prompt packs.
FAQ
Is ClaudeKit worth it in 2026?
For operators whose work spans more than one discipline, yes. The Pro bundle at $29.99/month gives you any 3 kits with one swap per billing cycle, which covers most cross-functional workflows at a cost that pays back quickly if you are doing work that would otherwise take hours per task. If your work is exclusively software engineering, a cheaper engineering-only option is worth comparing first.
How many commands does ClaudeKit have total?
ClaudeKit v2 ships 101 commands across 5 kits: 25 in EngineerKit, 20 in MarketingKit, 19 in SEOKit, 20 in EcomKit, and 17 in VideoKit. There are also 19 skills and 13 read-only specialist agents. The full measured token footprint is 82,197 tokens across all 5 kits.
What happened to FounderKit and SalesKit?
Both were shelved in the v2 rebuild. The current product is 5 kits: EngineerKit, MarketingKit, VideoKit, SEOKit, and EcomKit. There is no public timeline for FounderKit or SalesKit returning. If your primary use case is fundraising or outbound sales, ClaudeKit does not currently have a kit for that work.
What is the refund policy?
14 days from purchase. Not 30 days. Each license covers 3 devices. Lifetime licenses are as-shipped with no future updates included.
Does ClaudeKit work with Claude Code's plugin marketplace?
Yes. You can install any kit with /plugin marketplace add Madni-Aghadi/claudekit-<kit> directly from inside Claude Code, in addition to the ck install <kit> CLI path. Both routes print the token ledger on completion.
What does "read-only specialist agents" mean?
In v2, agents are reviewer, auditor, and researcher roles that inspect and report. They do not orchestrate other agents, do not block commands from completing, and do not have write access to your codebase. There are 13 total across the 5 kits. The old blocking reviewer/quality-gate agent pattern from v1 was removed because it added friction without measurable quality improvement.
If your biggest time sink right now is engineering work, the EngineerKit page has the full command list and a breakdown of the daily-8 workflow. If it is content or growth, MarketingKit is the place to start. The pricing page has the full tier comparison and the kit-swap rules for Pro subscribers.
Give Claude Code a real team
Five kits, 101 commands, every token measured. Pick the team that matches your work and install it in five minutes.
See the kits
