A
auracura/idea-autopsy
A Claude skill that tells you if an AI-generated business idea is BUSINESS, BULLETIN or BULLSHIT — with…
View on GitHub ↗Link checked 2026-11-12 · HTTP 200 · opens the vendor's own site
Install & config
Client config (built from the install command this repo gives)
{
"mcpServers": {
"auracura-idea-autopsy": {
"command": "npx",
"args": ["skills"]
}
}
}Parsed from the README we fetched on 2026-10-29; we have not executed it.
command/args come from the install command the repository states. When it is ambiguous, the project README wins.
VS Code (.vscode/mcp.json - the key is servers)
{
"servers": {
"auracura-idea-autopsy": {
"command": "npx",
"args": ["skills"]
}
}
}Same server, but VS Code keeps it under <code>servers</code>.
Claude Code (command line)
claude mcp add --transport stdio auracura-idea-autopsy -- npx skills
Claude Code does not read a config file; add the server with this command.
Key names differ by client: Clients & CLI
- Star count
- 5 ★ · auracura/idea-autopsy · read from the GitHub search API · (checked 2026-11-13 20:00:50)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-10-30 · last seen 2026-11-14 · seen 16× · listed by 1 sources
- Score
- score 103 · 5 ★ · 1 sources · repo
- Install
npx skills
FAQ (answers come from our records)
How do I install it?
The repository provides this install command:
npx skills. It comes from the repository itself; we cannot guarantee it runs as-is in your client.Who is it for?
We filed it under these tags: 开源项目. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
auracura/idea-autopsy；first catalogued 2026-10-30；last seen 2026-11-14。There is no official certification or quality rating here — we record sources and dates only.# Idea Autopsy
以下摘录自该仓库的 README（片段） · 在 GitHub 读完整 README →
A Claude skill that tells you whether a business idea — especially one an AI suggested — is a BUSINESS, a BUSINESS, WRONG OWNER, a BULLETIN (really just content), or BULLSHIT. It is deliberately hostile. Its job is to kill the idea; if the idea survives, it hands you the smallest test that would prove it wrong, with a kill line.
## Install
With skills.sh:
bash
npx skills add auracura/idea-autopsy
Or manually: copy SKILL.md to .claude/skills/idea-autopsy/SKILL.md.
Then say "autopsy this:" followed by the idea and your constraints, e.g. "autopsy this: a booking app for mobile barbers. 6 hours/week, $2k/month in a year, solo barbers, Trinidad."
## Where it came from
I spent four hours letting an AI pitch me side businesses: invoice-chasing tools, Caribbean land registries, verified-handmade badges, graduate-to-trades content sites, an "Etsy without AI slop". Eleven of twelve died — to a free government website, to a chatbot that already answers the question, to a funded team already running the race, and to my own calendar.
The machine generated. I decided. The deciding was a repeatable set of tests, so I wrote them down. This is them.
## The tests, in order
1. The Wrapper Test — what does a general-purpose LLM not have that this idea needs? Private data, action, trust/licence, presence, a captive audience, or live specificity. None of the six = wrapper.
2. The Crowd and the Graveyard — name five incumbents with prices; name who tried this and died. In 2026 any pain that makes the news has ten apps within a quarter.
3. The Time Test — casework doesn't fit evenings. Count the hidden weekly chores.
4. The World-Condition Test — what does the idea ride, and what happens when it reverses?
5. The Money Maths — with the owner's own target, not a nicer one.
6. Fresh Eyes — isolated reviewers, one lens each, for anything still alive.
7. Verdict + smallest test + kill line.
## What it did in testing (examples/)
01-ai-meal-planner.md — textbook LLM slop. BULLSHIT. Found the graveyard on its own: Yummly (Whirlpool) shut 2024, Mealime (Albertsons) shutting Oct 2026.
02-tax-prep-service.md — a real service I'm building, disguised. First verdict: BULLSHIT on the stated hours. It found that the "paperless" filing route requires a signed form physically dropped at a government office during business hours, and that the government is closing the wedge. I didn't know that.
03-one-page-site-builder.md — Carrd, described as if proposed today. BULLSHIT, because Carrd exists. Fair — and a limit: the skill judges an idea today; it cannot tell you what would have worked in 2016.
04-handmade-marketplace.md — Goimagine, disguised. BULLSHIT. It identified the real company, its US$400k raise, its US$40k/month burn, and its closure on 23 March 2026.
05-tax-prep-constraints-fixed.md — the same tax service with a paid runner and more hours. Verdict moved to BUSINESS, WRONG OWNER: wrapper test passed, hours now fit, but the owned audience is 7× too small for the target. That is the behaviour you want — the verdict moves with the constraints, and it names which one still fails.
## Install
With skills.sh:
bash
npx skills add auracura/idea-autopsy
Or manually: copy SKILL.md to .claude/skills/idea-autopsy/SKILL.md.
Then say "autopsy this:" followed by the idea and your constraints, e.g. "autopsy this: a booking app for mobile barbers. 6 hours/week, $2k/month in a year, solo barbers, Trinidad."
## Where it came from
I spent four hours letting an AI pitch me side businesses: invoice-chasing tools, Caribbean land registries, verified-handmade badges, graduate-to-trades content sites, an "Etsy without AI slop". Eleven of twelve died — to a free government website, to a chatbot that already answers the question, to a funded team already running the race, and to my own calendar.
The machine generated. I decided. The deciding was a repeatable set of tests, so I wrote them down. This is them.
## The tests, in order
1. The Wrapper Test — what does a general-purpose LLM not have that this idea needs? Private data, action, trust/licence, presence, a captive audience, or live specificity. None of the six = wrapper.
2. The Crowd and the Graveyard — name five incumbents with prices; name who tried this and died. In 2026 any pain that makes the news has ten apps within a quarter.
3. The Time Test — casework doesn't fit evenings. Count the hidden weekly chores.
4. The World-Condition Test — what does the idea ride, and what happens when it reverses?
5. The Money Maths — with the owner's own target, not a nicer one.
6. Fresh Eyes — isolated reviewers, one lens each, for anything still alive.
7. Verdict + smallest test + kill line.
## What it did in testing (examples/)
01-ai-meal-planner.md — textbook LLM slop. BULLSHIT. Found the graveyard on its own: Yummly (Whirlpool) shut 2024, Mealime (Albertsons) shutting Oct 2026.
02-tax-prep-service.md — a real service I'm building, disguised. First verdict: BULLSHIT on the stated hours. It found that the "paperless" filing route requires a signed form physically dropped at a government office during business hours, and that the government is closing the wedge. I didn't know that.
03-one-page-site-builder.md — Carrd, described as if proposed today. BULLSHIT, because Carrd exists. Fair — and a limit: the skill judges an idea today; it cannot tell you what would have worked in 2016.
04-handmade-marketplace.md — Goimagine, disguised. BULLSHIT. It identified the real company, its US$400k raise, its US$40k/month burn, and its closure on 23 March 2026.
05-tax-prep-constraints-fixed.md — the same tax service with a paid runner and more hours. Verdict moved to BUSINESS, WRONG OWNER: wrapper test passed, hours now fit, but the owned audience is 7× too small for the target. That is the behaviour you want — the verdict moves with the constraints, and it names which one still fails.
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/auracura-idea-autopsy/)
- Source
- github · https://github.com/auracura/idea-autopsy
- Browse by field
- Open source
- Traction
- 5 ★
- added
- today
More like this
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.