Skip to content

Founder Skills v0.1.0 — Your Fundraising Prep Room

Choose a tag to compare

View all tags
@yaniv-golan yaniv-golan released this 26 Feb 20:22
· 158 commits to main since this release
v0.1.0
ba7ed5d

Three AI coaching agents that help startup founders prepare for fundraising — built as a Claude Cowork plugin by lool ventures.

What's Included

Market Sizing Agent — Builds defensible TAM/SAM/SOM analysis that holds up under investor scrutiny.

  • Top-down, bottom-up, or combined methodology
  • External validation against analyst reports, government data, and industry stats
  • Sensitivity analysis with confidence-based range widening
  • 22-item self-check against common market sizing pitfalls
  • HTML report with funnel, tornado, and donut charts

Deck Review Agent — Reviews pitch decks against 2026 investor best practices, calibrated by funding stage.

  • 35 criteria scored across 7 categories (pass / fail / warn)
  • Stage-aware expectations for pre-seed, seed, and Series A
  • Slide-by-slide feedback grounded in named best-practice principles from Sequoia, YC, a16z, DocSend, and Carta
  • Overall readiness rating with highest-leverage recommendations
  • HTML report with gauge, radar, bar charts, and slide map

IC Simulation Agent — Simulates a realistic VC Investment Committee discussion.

  • Three independent partner voices: The Visionary (markets & timing), The Operator (execution & GTM), The Analyst (unit economics & financials)
  • 28 evaluation dimensions across 7 categories with conviction-based scoring
  • Portfolio conflict detection
  • Fund-specific mode with WebSearch-backed research on real funds
  • Imports prior market sizing and deck review artifacts for grounded analysis
  • HTML report with gauge, radar, partner cards, and conflict table

How It Works

Each agent follows a structured artifact pipeline: gather inputs, analyze, validate with deterministic scripts, compose a report, and optionally generate a self-contained HTML visualization. Every calculation is transparent and auditable — scripts output JSON, assumptions are categorized and sourced, and cross-artifact consistency is validated before the final report.

All three agents work standalone, but they're designed to build on each other: run market sizing first, then deck review, then IC simulation for the most grounded analysis.

Technical Details

  • 13 Python scripts — pure Python, no external dependencies, JSON in/out, portable CLI
  • 123 regression tests across 6 test modules
  • 10 reference documents grounding agent behavior in real investor methodology
  • Runs entirely within your Claude session — no data collected or transmitted
  • Apache 2.0 license

Installation

In Claude Cowork: Customize → Personal Plugins → + → Browse → add the lool-ventures/founder-skills marketplace → install founder-skills. Start a new session after installing.

Also works in Claude Code (claude plugin marketplace add lool-ventures/founder-skills).

Requirements

  • Claude Cowork (or Claude Code)
  • Python 3.12+
  • Works best with Opus-class models; Sonnet produces adequate but less distinct IC partner voices

What's Next

  • More skills coming — covering the full fundraising lifecycle before and after the pitch
  • Additional platform support as marketplace/discovery layers stabilize
  • Have a skill you'd like to see? Request it in GitHub Discussions
Assets 2
Loading
0 Join discussion