Claude Skills Founder

Startup strategy from your terminal. Not another dev tool.

Skills · Install · Usage · Examples · Philosophy · Contributing

[!NOTE] Most Claude Code skill packs are developer tools: code review, git workflows, security audits. This one is for the founder sitting in a repo at 2am, trying to figure out if their idea is worth building.

We built these at Emotix while running our own startup. They encode the frameworks, questions, and patterns we wish we had on day one.

See it work before you install anything. examples/ holds five ideas run end to end, with every file the runs produced and nothing edited, including the one the skills told us to kill. evals/ measures the skills against plain Claude and publishes where they make no difference.

Skills

Every skill takes plain-language input and returns scorecards, tables, and checklists you can act on.

Skills that build on each other

Each skill saves its output to a founder/ folder in your project and reads what earlier skills saved. Run /founder:competitor-matrix once, and /founder:pricing-strategy and /founder:pitch-deck reuse its sourced competitor prices instead of starting over.

founder/
├── facts.md               # what you told the skills: stage, metrics, team, the raise
├── validate-idea.md
├── competitor-matrix.md
├── pricing-strategy.md
└── ...

Commit the folder to keep a history of how your thinking changed, or add it to .gitignore to keep it private.

Facts about the outside world, such as competitor prices, funding rounds, market sizes, and benchmarks, come with a source link. When a skill can't find one, it says "Estimate" or "Not found" instead of guessing. Your own metrics, customer quotes, and testimonials are never invented: missing ones show up as [placeholders] for you to fill in.

Install

/plugin marketplace add emotixco/claude-skills-founder
/plugin install founder@emotix

git clone https://github.com/emotixco/claude-skills-founder.git ~/.claude/skills/founder

git -C ~/.claude/skills/founder pull

Usage

Open Claude Code and type any skill with your context:

> /founder:validate-idea An AI tool that analyzes pitch decks and gives investor-perspective feedback

> /founder:competitor-matrix Project management tools for solo founders (not teams)

> /founder:pricing-strategy B2B SaaS for restaurant inventory, targeting independent restaurants with 1-3 locations

> /founder:pitch-deck Pre-seed raise, $500K, AI competitor analysis for founders, 200 beta users, $2K MRR

Later, when the folder has something in it:

> /founder:status

Run a skill with no input and it asks for what it needs.

[!IMPORTANT] If an install step fails or a skill invents a number, that's a bug and we want to hear it. Open an issue with your claude --version and what you ran. A broken install is silent otherwise: people try it once and leave.

[!TIP] Be as specific as possible. Include your stage, metrics, constraints, and what you've already tried. Facts you give once are saved to founder/facts.md, so you don't have to repeat them.

Examples

examples/ has five ideas run end to end, with every file the runs produced and nothing edited: a UK physiotherapy SaaS, an office coffee subscription in Baku, a solo developer's puzzle game, a restaurant inventory tool, and one idea the skills told us to kill. Each one lists what it cost and which of its cited links are dead. Read it before installing to see what you get.

evals/ tests the plugin against plain Claude with claude plugin eval, and evals/benchmark.md carries the scheduled re-runs, including the cases where the plugin loses. The honest summary, measured on 2026-10-07 across the 10 cases the suite held then: on Opus 5, 8 of those 10 scored the same with and without the plugin, because the model already keeps the placeholders and reads the saved files on its own. The exception is what happens when a new number contradicts a saved one: only the plugin updates founder/facts.md, so the next session doesn't inherit the stale figure. On Haiku 4.5 the plugin scored higher on 3 of the 10, the skills often need calling by name, and the shared rules file goes unread, so most of these rules never reach a Haiku session. Full numbers, including per-run cost, are in evals/README.md.

What each skill produces

Philosophy

Contributing

Have a skill idea? Open an issue or submit a PR.

To add a new skill:

1. Create skills/<name>/SKILL.md
2. Add frontmatter with name, description, argument-hint, and allowed-tools
3. Add the "Before you start" block every skill uses: which founder/ files it reads, what it needs, and where it saves
4. Keep the expected output under 2000 words
5. Test it with 3 different inputs: claude --plugin-dir .
6. Run python3 .github/scripts/check_skills.py and claude plugin validate . before opening the PR. The first catches frontmatter the second accepts, such as an unquoted colon in a description
7. If the skill makes a promise (never invent X, always ask for Y), add a case under evals/ that checks it

Rules that apply to every skill, such as sourcing and writing style, live in shared/conventions.md. Change them there, not in each skill.

[!IMPORTANT] We're looking for skills that fill genuine gaps in the founder workflow, not developer tools repackaged with startup vocabulary.

License

MIT. Use these however you want.

Built by the Emotix team

We build AI agents that turn startup ideas into research, analysis, and product briefs.

Website · GitHub · Twitter