20 AI agent skills that cover the entire build-to-launch loop:

DEFINE THE JOB
1. bmad-spec – turn ideas into structured specs
2. domain-modeling – model core concepts and relationships
3. codebase-design – plan codebase structure
4. google-agents-cli-workflow – build agent workflows with Google Agents CLI
5. prompt-optimizer – improve and refine prompts

BUILD THE SYSTEM
6. google-agents-cli-adk-code – build agent code with Google's tooling
7. mastra – skills for building AI agent systems
8. typesafe-ai – build AI workflows with stronger type safety
9. fastify-best-practices – better patterns for Fastify backends
10. neon-postgres – work with Neon + Postgres

MAKE IT USABLE
11. prisma-database-setup – set up database layer with Prisma
12. clerk-orgs – add organization and auth workflows
13. emil-design-eng – improve product design and frontend execution
14. review-animations – review and improve UI animations
15. remotion-best-practices – build better programmatic video workflows

VERIFY THE REAL THING
16. playwright-cli – test real browser journeys
17. semgrep – scan codebase for security issues
18. sentry-fix-issues – find and fix production issues
19. instrument-product-analytics – track what users actually do
20. email-best-practices – improve product email flows

The loop: idea → spec → architecture → build → browser testing → catch failures → measure users → improve next version

Pick the biggest bottleneck in your project and install the skill that attacks it first.