Claude Code Cheat Sheet

Original Article ↗ Hacker News Discussion ↗

Cheat Sheet Purpose & Features

  • Community response to a single-page Claude Code cheat sheet is largely positive; many find it helpful as a quick reference.
  • Sheet includes keyboard shortcuts, slash commands, workflows, skills, memory/CLAUDE.md, MCP setup, CLI flags, config files, and auto-detects OS for key mappings.
  • It auto-updates daily from the changelog and highlights new features; no signup, printable, mobile-friendly.

Accuracy, Omissions & Update Concerns

  • Several users report errors in keybindings (e.g., paste image shortcuts, external editor shortcut) and path notation.
  • Some important items were initially missing, such as /keybinds, --dangerously-skip-permissions, certain env vars (e.g., IS_SANDBOX=1), and project rule conventions.
  • The author indicates they’re fixing issues and incorporating feedback; automatic daily updates aim to reduce staleness, but some still question printing something that changes frequently.

Printability & Format Debates

  • Works as a one-page print on US Letter if browser headers/footers are disabled; some note an extra blank page.
  • Jokes about “Ctrl+P to print” on mobile and the practicality of printing fast-changing docs; several prefer a PDF or just in-app help.

UX, Discoverability & Need for Cheat Sheets

  • Some view the need for a cheat sheet as a UX red flag; others argue CLIs traditionally have man pages/cheat sheets and this is normal.
  • Mixed views on discoverability: some say Claude’s own help and /keybinds make most features easy to find; others remark that basic help shortcuts are oddly named or missing.
  • Debate over CLI vs GUI: some see the CLI as powerful but intense; others prefer IDE integrations or GUI tools like competing products.

Flags, Env Vars & Power-User Features

  • Power users emphasize lesser-known env vars and flags (e.g., IS_DEMO=1, IS_SANDBOX=1, --dangerously-skip-permissions) and want them documented.
  • Some worry that model “safety alignment” might under-document anything labeled “dangerous.”

Safety, Permissions & Sandboxing

  • Strong discussion around --dangerously-skip-permissions: some say it’s essential; others stress using containers or dedicated tools to sandbox agents.
  • Multiple anecdotes of AI agents in various tools damaging production systems highlight the need to treat agents as potentially unsafe, enforce task boundaries externally, and prefer copy–diff–apply workflows over live mutation.

Claude Code vs Alternatives & Agents

  • Several comments compare Claude Code favorably to other CLIs for extensibility (hooks, MCP, skills) but note performance issues (lag, flicker, sluggishness due to the tech stack).
  • Others still find competing tools better at pure coding quality, using Claude mainly for UX and extensibility.
  • Discussion branches into agentic workflows, self-improving skills, and applying Claude to complex automation (including trading), with both enthusiasm and skepticism about real-world payoff and “AGI” claims.