A Claude plugin for AI-native product teams who ship things that actually matter.
48 skills. 17 slash commands. 6 domains. From AI feature design to shipping with conviction.
⚠️ Disclaimer: This repository is for informational and educational purposes only. Nothing here constitutes professional product, legal, or business advice. You are solely responsible for your product decisions. See DISCLAIMER.md for full terms.
Designing an AI feature? → /ai-feature
Making a hard product decision? → /instinct
Have a blurry idea that needs shape? → /sharpen
Writing your launch story? → /story
Ready to ship but not sure? → /ship
Auditing your design? → /design-audit
If this helps you build better, ⭐ the repo.
Most product tooling teaches you how to run process. This teaches you how to build better.
There's a specific kind of product work that existing tools don't address — the judgment calls, the taste questions, the "is this real or theater" moments that determine whether what you ship actually matters. That's what build-sharp is for.
Six domains. Each one covers a distinct part of building AI-native products well.
claude plugin add --marketplace Dragoon0x/build-sharp- Open Customize (bottom-left)
- Go to Browse plugins → Personal → +
- Select Add marketplace from GitHub
- Enter:
Dragoon0x/build-sharp
The skill files in each */skills/ folder follow a universal format and work with any AI tool that reads markdown skill files.
# Copy all skills for Cursor
for plugin in bs-*/; do
cp -r "$plugin/skills/"* .cursor/skills/ 2>/dev/null
doneDesign AI features that earn trust, not just ship fast.
Skills (6): ai-theater-check · ai-feature-brief · ai-ux-patterns · ai-failure-modes · ai-scope-cut · ai-trust-design
Commands (3):
/ai-feature— Full AI feature design: theater check → brief → failure modes → scope cut → trust arc/theater-check— Fast verdict: is this real value or AI theater?/ux-patterns— The right AI UX patterns for your specific feature
Examples:
/ai-feature AI that automatically categorizes support tickets by urgency/theater-check Smart sort that reorders tasks based on predicted priorityWhat are the failure modes for an AI writing assistant in a legal context?
Sharpen your judgment. Make better calls under real pressure.
Skills (6): decision-pressure-test · signal-vs-noise · conviction-check · assumption-killer · contrarian-take · fast-judgment
Commands (3):
/instinct— Full instinct session: pressure test + conviction check + assumption map/pressure-test— Fast pressure test on a single decision/signal— Separate signal from noise in your current inputs
Examples:
/instinct We're rebuilding onboarding before investing in retention/signal Three enterprise customers asked for SSO, two churned users cited onboardingWhat are the hidden assumptions in our decision to go upmarket?
From blurry to buildable in one session.
Skills (5): concept-sharpen · blur-to-brief · rapid-spec · idea-to-pitch · quick-prototype-plan
Commands (3):
/sharpen— Turn a vague idea into a sharp concept + 60-second pitch/brief— Convert a vague direction into a brief engineering can act on/prototype— Plan a prototype that answers your most important open question
Examples:
/sharpen Something that helps product teams know what to work on next/brief Improve the way users find past conversations in the app/prototype Test whether users will trust AI with their calendar scheduling
Tell the story that makes your product stick.
Skills (5): ai-product-story · positioning-cut · launch-message · differentiation-honest · investor-narrative
Commands (3):
/story— Full narrative: product story + positioning + honest differentiation/launch-message— The one message that carries your launch/pitch— Investor narrative: why now, your insight, why you
Examples:
/story [describe your product and who it's for]/launch-message Launching AI feature that halves time spent on PRDsHow is our positioning different from what our competitors claim?
Know what good looks like. Ship things that earn it.
Skills (5): good-vs-polished · ui-taste-check · interaction-intent · ux-theater-check · visual-honesty
Commands (2):
/design-audit— Full audit: taste, polish vs good, UX theater, visual honesty/taste-check— Fast opinionated check on a specific screen or interaction
Examples:
/design-audit Our onboarding flow — 4 screens from signup to first value/taste-check Our empty state for users who haven't connected data yetWhat's the difference between a product that's good vs just polished?
Ship with conviction, not just speed.
Skills (5): launch-decisions · pre-ship-audit · cut-or-keep · ship-go-nogo · post-ship-signal
Commands (3):
/ship— Full pre-ship workflow: decisions → audit → go/no-go → 72hr signal plan/go-nogo— Structured go/no-go call on a planned ship/cut-or-keep— Principled decision on what stays in or out of a launch
Examples:
/ship Launching AI weekly summary feature — target ship is Friday/go-nogo New onboarding ships tomorrow. Core flow works. Mobile not tested yet./cut-or-keep CSV export is 80% done but might take 2 more days
Judgment over process. This isn't a template library. Every skill encodes a real decision you have to make, not a document you have to fill in.
Opinionated over neutral. Good product work requires taking positions. Build-sharp takes positions. You can disagree — but you can't ignore.
Speed over ceremony. Every command is designed to produce something useful in one session. Not a research project. An output.
AI-native, not AI-flavored. This was built for teams actually shipping AI products. Not for teams adding AI features to existing products as an afterthought.
Built by Dragoon0x.
MIT License — see LICENSE.
Full disclaimer — see DISCLAIMER.md.
Usage terms — see TERMS.md.
If this helped you ship something better, tell someone.