Fonti — Claude Code, quello che avrei voluto sapere prima

The article is a curated list of resources for Claude Code, an AI coding tool by Anthropic. It covers foundational LLM concepts, official documentation for features like memory, subagents, and permission modes, as well as external tools and community frameworks like "Superpowers" and "caveman" to enhance functionality. The summary serves as a practical guide for users seeking to understand and optimize their use of Claude Code.

- LLMs explained briefly 3Blue1Brown : https://www.youtube.com/watch?v=LPZh9BOjkQs - Transformers, the tech behind LLMs 3Blue1Brown : https://www.youtube.com/watch?v=wjZofJX0v4M - Liu 2020, Nature Communications: https://www.nature.com/articles/s41467-020-15804-w - Anthropic glossary token, context window : https://docs.claude.com/en/docs/about-claude/glossary - Pricing input/output, cache, batch : https://platform.claude.com/docs/en/about-claude/pricing - Pro/Max usage limits session + weekly : https://support.claude.com/en/articles/11145838-use-claude-code-with-your-pro-or-max-plan - Memory CLAUDE.md, scoping utente/progetto, /memory : https://code.claude.com/docs/en/memory - Explore the .claude directory mappa canonica, file per file : https://code.claude.com/docs/en/claude-directory - MCP in Claude Code overview, scope, install : https://docs.claude.com/en/docs/claude-code/mcp - GitHub CLI gh : https://cli.github.com/ - Google Workspace CLI gws : https://github.com/googleworkspace/cli - Playwright MCP: https://github.com/microsoft/playwright-mcp - Skills overview Claude Code docs : https://code.claude.com/docs/en/skills - Annuncio ufficiale /insights Thariq Shihipar, Anthropic, feb 2026 : https://x.com/trq212/status/2019173731042750509 ancora non c'è documentazione ufficiale - Model configuration alias opusplan ,/effort : https://code.claude.com/docs/en/model-config - Permission modes default, acceptEdits, plan, auto, bypassPermissions, --dangerously-skip-permissions : https://code.claude.com/docs/en/permission-modes - Docker sandbox ufficiale per Claude Code sbx run claude , immaginedocker/sandbox-templates:claude-code lancia con--dangerously-skip-permissions di default : https://docs.docker.com/ai/sandboxes/agents/claude-code/ - "Claude Code dangerously-skip-permissions: Why It's Tempting, Why It's Dangerous": https://thomas-wiegold.com/blog/claude-code-dangerously-skip-permissions/ - Subagents Claude Code docs canonical : https://code.claude.com/docs/en/sub-agents - Agent teams Claude Code docs canonical, research preview : https://code.claude.com/docs/en/agent-teams - "How and when to use subagents in Claude Code" blog Anthropic, quando delegare vs parallelizzare : https://claude.com/blog/subagents-in-claude-code - Addy Osmani — "Claude Code Swarms" approfondimento sui pattern di coordinazione, delegate mode, plan-before-implement : https://addyosmani.com/blog/claude-code-agent-teams/ - caveman Claude Code skill/plugin che taglia ~65% degli output token : https://github.com/JuliusBrussee/caveman - Output styles meccanismo canonico Claude Code per personalizzare lo stile di risposta : https://code.claude.com/docs/en/output-styles - Repo ufficiale Jesse Vincent / Prime Radiant, ~14 SKILL.md : https://github.com/obra/superpowers - Pagina marketplace Anthropic descrizione canonica : https://claude.com/plugins/superpowers - Marc Nuri — "Superpowers: The Claude Code Skills Framework Shipped as Markdown" approfondimento su come funziona la methodology : https://blog.marcnuri.com/superpowers-claude-code-skills-framework