bama
BAMA: Build Any Mac App. Agent skills to research, clone, design, build, optimize, and test native macOS apps with Swift and SwiftUI.
翻訳待ち
リポジトリ情報はまだ収集していません。
No updates detected yet. GitHub releases will appear here.
Claude Code plugin that tries to kill your startup idea with sourced research: score /100, GO / PIVOT / NO-GO, market metrics (TAM, CAC, LTV), country analysis, roast mode. Stop building things nobody wants.
翻訳待ち
Loading the full README.
Read the full README ↗/plugin marketplace add RemiPelloux/kill-my-idea
/plugin install kill-my-idea@kill-my-idea
▶ Watch a real run (50 s): 19 searches, 34 sources, the verdict, the report and the dashboard.
Live reports · Gallery · Examples · Security · Changelog · Roadmap · Sponsor
Ask an AI whether your idea is good and it will usually find a way to say yes. kill-my-idea does the opposite. It looks for a free alternative, competitors, people who already failed, and the numbers that don't add up. Then it gives you a score out of 100, a GO / PIVOT / NO-GO verdict, the ways the idea could still be saved, and a report where every number has a source.
The most common use: you see a viral thread promising easy money. Paste it in:
kill my idea: <paste the thread>
It checks every claim, names what the author is selling, redoes the math and gives a verdict. Real example, a viral thread about AI-written Amazon KDP books:
| The thread said | The check found |
|---|---|
| "$3,000 a month from 5 sales a day" | About $1,380 in royalties once printing costs are counted (KDP royalty rules) |
| "Anyone can do it" | About 300k new books a month compete for the same shelf (AI Policy Desk) |
| "$3M from books" | No evidence ever shown. The author sells KDP coaching |
| Verdict | PIVOT, 52/100. Run a 1-week niche test before writing a word (full card) |
We asked plain Claude and kill-my-idea about the same idea: an AI app that writes wedding speeches, at $19 each.
| Claude, no plugin | kill-my-idea | |
|---|---|---|
| Verdict | "It's a decent small business." | NO-GO, 44/100, confidence high |
| Sources | None | 40, linked |
| Competition | "Several AI speech tools exist." | 11 named: AI tools at $0.99–65 a speech, human writers at $100–655, and free ChatGPT does the job. A look-alike app made $9.98 in the last 30 days ($311 all-time, Stripe-verified) |
| Unit economics | "Margins are almost entirely marketing." | LTV $19.7 vs CAC ~$35 on paid search: LTV:CAC 0.56 (estimates, reasoning shown). Every paid sale loses money |
| What to do next | Spend about $200 on ads and see | Archive the $19 app. If you want one test: a 7-day Fiverr gig for a $59 human-edited speech. Fewer than 3 inquiries: stop |
Plain Claude's full answer is here, unedited. The kill-my-idea version is a live report.
All seven are real runs, mostly ideas we were about to spend time or money on. The coloring book is why this exists: we built it first, three versions and a day of work, and a five-minute search then found the same book for free. The plugin also checked itself: 63, because fifteen open-source idea validators already exist.
The gallery ranks 14 more real runs on popular side hustles (dropshipping, Etsy planners, AI headshots, a faceless YouTube channel, vending machines…), with every report one click away.
Every run writes report.html, a report you open in a browser and can print as an A4 PDF.
It opens with the answer, then shows what would change it:
It goes on to cover:
Every number shows whether it was sourced, estimated (hover for the formula) or not found. After a deep dive, the report shows how the score moved since the quick verdict. One click on Share card downloads a PNG verdict card for social media, and reports follow your language (English or French).
Every run updates ideas/board.html, a dashboard of every idea you've checked, and
ideas/BRAIN.md, one note that remembers you and your ideas.
This is the beginning of the README. Read the rest at the original. Read the full README ↗