Claude Code Skills
makappen-collab/claude-skills
Skills prontos para estender capacidades do Claude Code
Ver no GitHubConjunto curado de skills para Claude Code, focado em automação de tarefas específicas como avaliação de negócios e renderização segura de apresentações. Cada skill é auto-contido com instruções próprias e pode ser instalado globalmente ou por projeto.
Instalação
npx convez-ai makappen-collab/claude-skills1 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
- Avaliar viabilidade de ideias de negócio com rubrica estruturada
- Gerar decks PowerPoint em código com garantia de renderização correta
- Automatizar workflows repetitivos com agents especializados
Veredito

Coleção bem arquitetada de skills Python que resolve problemas concretos (validação de negócios, renderização de decks). Instalação simples e documentação clara, mas o repositório é auto-gerado e recomenda issues ao invés de PRs diretos—útil para quem já usa Claude Code e precisa estender funcionalidades específicas.
README
A tradução para o português deste README ainda está na fila.
Claude Code skills
A small, curated set of Claude Code skills,
shared publicly. Each skill is a self-contained folder with a SKILL.md that
tells Claude when to fire and how to behave.
This repo is auto-generated. The skills are mirrored from a private source on every update, so direct pushes here may be overwritten — open an issue for bugs or suggestions rather than a PR.
Install a skill
Copy any skill folder into your Claude Code skills directory:
# global (all projects)
cp -r skills/<skill-name> ~/.claude/skills/<skill-name>
# or per-project
cp -r skills/<skill-name> .claude/skills/<skill-name>
Start (or restart) Claude Code and it will pick the skill up.
Skills in this repo
solopreneur-business-assessor
Scores a concrete business idea cold against a fixed 9-axis rubric plus three gate tests (willingness-to-pay, AI-resilience, cost-of-entry), tuned for a solo founder building with AI agents at lifestyle scale. It spawns an isolated sub-agent that never sees the chat that produced the idea, so the verdict isn't swayed by your enthusiasm, then runs an adversarial pass that tries to flip the result before reporting. Output is a scorecard, a tiered verdict (Avoid / Hard / Viable / Strong), the single biggest reason, and what would have to be true to make it work.
Optional: drop a calibration.local.md in the skill folder with your own edge,
channels and target economics to tune axes 7-8 and the cost gate to you.
pptx-render-safe
Production discipline for PowerPoint decks built in code with python-pptx. It
encodes the trap where "shrink text to fit" (autofit) previews cleanly in
LibreOffice but silently overflows its boxes in real PowerPoint and Keynote, and
fixes it by baking explicit font sizes. Ships a reusable text-fitting helper
(scripts/fit_text.py) and a render-to-images verification workflow so you prove
the deck is clean before handing it over.
Optional: drop a brand.local.md in the skill folder pointing at your default
brand-template folder so the skill resolves your palette and fonts automatically.
Licence
MIT.