What's new

New: /eng-remake — clone any website pixel-perfect, then rebrand it as your own product. 27 commands · 6 skills · 4 agents · 66,865 measured tokens.

• ✓/eng-remake <url>, <brand brief> — two strict stages: a faithful clone (browser-driven extraction with exact computed CSS, interaction sweeps, parallel section builders), then a remake layer that swaps palette, copy, logo, and imagery while the proven layout and interactions stay identical
• ✓Brand synthesis on paper before code — BRAND.md with a palette token map (contrast-checked), per-slot copy length envelopes, an AI-slop vocabulary ban list, and an original SVG logo
• ✓A QA loop that refuses to finish — build gate, visual audit at 1440/768/390, replay of every recorded interaction, and a zero-source-brand-strings sweep; exits only after two consecutive clean passes
• ✓/eng-remake:multi — whole-site mode: discovers pages from the nav and footer (capped at 8, or list routes explicitly), clones each as its own route with shared nav/footer components, and runs one brand and one QA gate across every page

/eng-site goes fully cinematic — opt-in AI image + looping hero video. 26 commands · 5 skills · 4 agents · 54,379 measured tokens.

• ✓/eng-site — the premium motion upgrade: a real looping cinematic hero video, generated from the hero still with Seedance 2.0 via Higgsfield, sealed as a seamless constant-velocity loop and finalized watermark-free for instant autoplay
• ✓Split-vendor AI assets — Gemini 2.5 Flash Image (Nano Banana) for three distinct stills, Higgsfield for the one hero clip; the free hand-built CSS cinematic hero stays the default, so a working site never depends on a paid vendor
• ✓Token ledger re-measured across the whole kit for an honest footprint — every file accounted for

New: /eng-site — a bespoke cinematic site builder. 26 commands · 5 skills · 4 agents · 36,060 measured tokens.

• ✓/eng-site — design + build a one-of-a-kind cinematic site (portfolio, open-source project, or product/launch) from one topic string; no fixed template, its own art-directed layout, palette, type, and GSAP/Lenis motion
• ✓Self-verifying — it opens the result in a real browser, fixes console errors, 404s, overflow, and mobile breakage, loops until two clean passes, then serves it on localhost
• ✓Free by default (hand-crafted CSS cinematic hero); AI stills (Gemini) and video (Seedance) are opt-in upgrades, never required

New kit: AI market research for Polymarket, Kalshi, and crypto. 18 commands · 4 skills · 3 agents · 23,217 measured tokens.

• ✓/tk-scan and /tk-decide — rank live markets by an edge score, then size the bet with correct prediction-market Kelly, or pass when there's no edge
• ✓/tk-whales and /tk-dashboard — track real taker flow and render a live trading terminal from public data, no API keys
• ✓/tk-backtest — score your forecast log with Brier calibration. Read-only and research-first by default; live execution is opt-in

New: /mkt infographic + /mkt doc. 22 commands · 5 skills · 2 agents · 21,537 measured tokens.

• ✓/mkt doc — paste any URL (a YouTube tutorial, an article, any page); it transcribes the source and renders the viral text-doc carousel: 1–4 centered PNG pages of bold numbered takeaways
• ✓/mkt infographic — turn any topic into a viral cheatsheet infographic: a high-res PNG, rendered locally in cream, purple, dark or indigo themes
• ✓Both render on your machine from one JSON spec — no Canva, no Figma — and every claim has to trace back to the source

25 commands · 4 skills · 4 agents · 20,413 measured tokens.

• ✓/eng debug — root-cause-first: evidence, ranked hypotheses, minimal fix, regression test
• ✓The daily-8 loop: catchup, plan, tdd, debug, verify, review, commit, handoff
• ✓Read-only specialist agents — every command ends with evidence, not a reviewer gate

20 commands · 3 skills · 2 agents · 16,714 measured tokens.

• ✓/mkt voice — builds a voice file from your real posts so drafts sound like you
• ✓/mkt humanize — strips the 14 AI tells before anything ships
• ✓/mkt repurpose — one piece becomes five formats without re-briefing

17 commands · 5 skills · 3 agents · 12,602 measured tokens.

• ✓/video clone — recreate a reference video's style in Remotion and verify the match
• ✓Script → storyboard → composition code → captions → publish, as commands
• ✓Data-driven renders: one composition, hundreds of variants

19 commands · 4 skills · 2 agents · 16,004 measured tokens.

• ✓/seo quick-wins — positions 8-20 and low-CTR snippets, ranked by click prize
• ✓/seo citations — N-run AI-citation measurement with confidence intervals, not a one-run coin flip
• ✓Extractable-content checks for the answer-engine era

20 commands · 3 skills · 2 agents · 16,464 measured tokens.

• ✓/ecom no-sales — store triage against AOV-band benchmarks, not vibes
• ✓Cart recovery, email flows, and margin math with every number cited
• ✓BFCM prep and ad diagnostics built on real conversion benchmarks

The ck CLI: auth once, install anywhere, see every token.

• ✓ck auth <key> then ck install <kit> — global to ~/.claude by default, --local for projects
• ✓Token ledger prints on every install; ck tokens recounts any kit
• ✓ck doctor diagnoses setups; ck list shows your entitlements

The site, built around the 5-kit lineup.

• ✓Dedicated pages for every kit with real, measured numbers
• ✓All-Access plan: every kit (and every future kit) for $49.99/mo
• ✓Checkout via Dodo Payments — license key by email, delivery through the ck CLI