Este plugin cambia por completo lo que creías saber sobre Claude Code. Anthropic lo lanzó discretamente: claude-code-setup, un plugin oficial que escanea tu proyecto y te dice exactamente qué te falta: hooks, skills, servidores MCP, subagentes y automatizaciones. Todo configurado automáticamente, paso a paso. El verdadero poder de Claude Code no viene solo de la herramienta. Viene del ecosistema que la rodea. Sin eso, estás funcionando en modo degradado sin darte cuenta. Instalación: /plugin install claude-code-setup@claude-plugins-official Guarda este post antes de olvidarlo.

May 17, 2026 · 6:55 PM UTC

12
204
1,565
136,088
Sort replies: Relevant Recent Liked
Replying to @SantiTorAI
read-only scanners list what's missing but skip the cost math. a 2k token setup still fires on every turn and runs $14+ per long opus session. scoring what's firing matters as much as what's missing ccmd.dev/r/e4u9gspq

ccmd — Score your CLAUDE.md in 5 seconds.

Polyglot agent-config analyzer. Free.

ccmd.dev
5
1,954
Replying to @SantiTorAI
No sabía que existía ese plugin. Lo probeé ahora y tiene razón con los MCP que me faltaban. Gracias por el dato.
2
1,421
Replying to @SantiTorAI
What’s interesting about plugins like this isn’t “auto setup.” It’s that AI coding environments are starting to look like operating systems. Hooks, Skills, MCP servers, Subagents, Automations, persistent rules. We’re moving from “chat with a model” to “configure an autonomous software organization.” And once that happens, infra quality starts mattering more than raw model IQ.
1,028
Replying to @SantiTorAI
My project man : github.com/wesleysimplicio/l…
1
1,318
Replying to @SantiTorAI
Been running this for a week now. The MCP server detection alone caught 3 servers I had configured but never actually hooked up. Biggest win: it flagged a custom skill that had been silently broken for weeks because the upstream API changed its response format. Worth it just for the health check.
2,047
Replying to @SantiTorAI
The scaffold recommendation is useful but the gap between "recommended" and "actually working with your codebase" is still manual work. Good start though — the hooks config alone saves an hour of setup.
1,319
Replying to @SantiTorAI
Why Japanese letter in your Claude Code - “週 (week)”
1,201
Replying to @SantiTorAI
Amazing share
274
Replying to @SantiTorAI
useful if it also marks sacred paths and files the agent should never touch. does it write those rules back into the repo?
1,898
Replying to @SantiTorAI
Scans your project and tells you exactly what's missing, this should have been the default from day one lol
1,349
Replying to @SantiTorAI
@SantiTorAI 'Anthropic lo lanzó discretamente: claude-code-setup…' La IA no va a reemplazar a nadie. Es herramienta.
1,201
Load more