6 skills · 101 founder interviews · data-backed
Startup advice
from 101 founders.
The advice draws on real patterns identified in 101 Starter Story founder interviews — not pitch-deck fairytales. Evaluate ideas, spot failure risks, and plan growth. Try it in your browser or install six skills in a coding agent.
Use it in Claude Code or Codex
▲ Claude Code · plugin install (recommended)
# two slash commands, no shell script
/plugin marketplace add yayashuxue/solo-founder-playbook
/plugin install solo-founder-playbook
▲ OpenAI Codex · one-line install
# same 6 skills, codex flavor
git clone --depth 1 https://github.com/yayashuxue/solo-founder-playbook.git ~/.codex/skills/solo-founder-playbook && bash ~/.codex/skills/solo-founder-playbook/install-codex.sh
101
founder interviews
6
terminal skills
80+
extracted patterns
11
indexed categories
/solo · six skills
SKILL · 01
/solo-analyze
Evaluate your idea
Match your idea against the real paths 101 founders took. What's likely to work, and what dies quietly.
IDEA · VALIDATION
SKILL · 02 · NEW
/solo-failures
Why startups die
Failure-mode analyst. 49.5% of founders openly discussed failure — see how startups actually die, not the post-hoc narrative.
DIAGNOSE · POST-MORTEM
SKILL · 03
/solo-growth
How to grow
Growth-strategy advisor. Which channels actually worked for similar founders — organic vs paid, content vs community.
GROWTH · DISTRIBUTION
SKILL · 04
/solo-patterns
Pattern browser
Browse and search the founder-pattern database. 11 categories, 80+ patterns, ordered by frequency. See what actually shipped.
RESEARCH · DATABASE
SKILL · 05
/solo-playbook
Step-by-step playbook
Ask "how do I [goal]?" → get a concrete, data-backed playbook distilled from what 101 founders actually did.
PLAN · EXECUTE
SKILL · 06
/solo-roast
Roast my plan
Garry Tan-style critique. Brutal but data-backed — kind, never sugarcoated. Hear the truth before you ship.
ROAST · REALITY CHECK
/solo-failures · 8 cards from one session
Why startups die.
One /solo-failures call produces 8 Xiaohongshu-style cards.
Click any tile → preview → jump to GitHub.
Click any tile → preview → jump to GitHub.
field notes · pdf
Earn w/ AI.
Cal AI 访谈拆解 / 0-1 增长笔记 — what actually worked when one AI app went from zero to seven figures.
A short, cardable PDF.
Influencer math, niche selection, RPM/CPM thresholds, the kill-rules.
No fluff, no hashtag advisor talk.
Influencer math, niche selection, RPM/CPM thresholds, the kill-rules.
No fluff, no hashtag advisor talk.
underlying dataset
101 founders. Counted, pattern-tagged, indexed.
Every recommendation maps back to a video count in the source data — no vibes, no advisor-speak, only frequencies actually observed.
49.5%
of interviews discussed failure — those experiences are included in the pattern library.
32
mentioned shipping fast and iterating — one of the more frequent practices in this sample.
18
built on and improved existing ideas — an idea does not have to start from scratch.
15
mentioned multiple failures before success — ask what they changed afterward.
14
emphasized early validation and charging — test demand with observable behavior.
12
discussed overwork and isolation — founder energy is a real operating constraint.
stop guessing
Install once. Type /solo. Get data.
30-second install — the skills live in your terminal. Use them when you're stuck, ignore them when you're not.