Pitch Generator
hytonylee/pitch-generator
Gera pitches multi-formato otimizados por estágio e tipo de investidor
Ver no GitHubSkill que gera pitches concisos e direcionados por audiência para produtos, projetos e campanhas de fundraising. Produz one-liners, parágrafos, pitches de investimento e oferece suporte especializado para pré-seed, incluindo ajustes de tração, tipos de investidor e anti-padrões comuns.
Instalação
npx convez-ai hytonylee/pitch-generator1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Preparar pitches para rodadas de fundraising (pré-seed, seed)
- Adaptar mensagens de produto para diferentes públicos e contextos
- Estruturar argumentos de investimento com evidências de tração apropriadas
- Identificar e evitar anti-padrões em apresentações de startups
Veredito

Skill robusto e bem pensado para founders e product managers. Integra benchmarking local (decks reais), ajustes por estágio de maturidade e calibração por tipo de investidor — vai além de templates genéricos. Implementação clara em JavaScript com suporte a contexto pré-seed que reflete realidades do mercado.
README
A tradução para o português deste README ainda está na fila.
Pitch Generator (Claude skill)
Generate concise, audience-targeted pitch drafts for products, projects, and ideas — with dedicated guidance for pre-seed fundraising.
What this skill does
This skill helps produce:
- A short "patterns used" note (when benchmark decks are available)
- A one-liner pitch
- A short paragraph pitch
- A concise investment-style pitch
It guides the assistant to ask for missing context (audience, goal, stage, traction, timing, and ask), then returns clear, specific messaging optimized for either live delivery or email.
Pre-seed stage support
When the user is at the pre-seed stage, the skill automatically adjusts:
- Inputs: asks about investor type (friends & family, angel, accelerator, institutional), use of funds / milestone roadmap, and instrument type (SAFE, convertible note).
- Traction framing: adds a Tier 5 evidence level for problem validation (customer discovery, waitlists, prototype demos) since most pre-seed companies lack revenue.
- Pitch priorities: weights founder-market fit, problem depth, and use-of-funds clarity over financial projections.
- Investor-type tailoring: adapts tone and detail level for friends & family vs. angels vs. accelerators vs. institutional pre-seed VCs.
- Anti-patterns: flags common pre-seed mistakes like premature financials, stacking confusing SAFE terms, and claiming traction that doesn't exist.
- Benchmark data: calibrates asks and framing against typical pre-seed ranges ($250K–$1M, SAFEs, 12–18 month runway targets).
Benchmark workflow (local-first)
- Local benchmark decks should be placed in
skills/pitch-generator/benchmarks. - When benchmark decks exist, the skill should:
- select the closest benchmark by audience/stage
- extract 1-2 concrete messaging patterns
- adapt those patterns to the user's context (without copying claims or wording)
- Prefer local files over external URLs for security and reliability.
Current local benchmark example:
skills/pitch-generator/benchmarks/2009_airbnb-pitch_deck.pdf
Sources
This skill draws on:
- Skillsmith Docs — skill authoring framework
- Carta: Pre-Seed Funding Guide — pre-seed stage definitions, investor types, readiness signals, typical round sizes, instrument guidance (SAFEs / convertible notes), dilution benchmarks, and common fundraising mistakes
Skill file
SKILL.md