Here are my tips for getting the most out of Claude Code, including a custom status line script, cutting the system prompt in half, using Gemini CLI as Claude Code’s minion, and Claude Code running itself in a container.
Nice list. For tip 27, the killer feedback loop I've recently discovered for building Web apps is to use the Chrome DevTools MCP server. This gives Claude the power to visually inspect the app in a running dev server as well find and fix errors that only happen under real condition on the client side.
Tip 20 about orchestrating worker Claude Code in containers changed my workflow. I run my main instance locally with --dangerously-skip-permissions in Docker for autonomous nightshift tasks. The tmux self-supervision pattern you mention? Game-changer for monitoring long-running jobs. I've been doing this for two months on production systems—full autonomy with exponential backoff when GitHub Actions choke.
One thing I'd add: containerized autonomy only works if you trust your test coverage.
Tip 0 about the token usage status bar is brilliant. I've been obsessing over context consumption too and ended up taking a different angle. Instead of monitoring and compacting after the fact, there's a plugin called context-mode that intercepts tool output before it ever enters context. Commands run in subprocesses, output gets indexed locally, and only summaries come back. Wrote it up here: https://reading.sh/how-one-plugin-cuts-claude-codes-context-bloat-by-98-096355e68166?sk=e06168222a38a98d4bf6add2daa10973
Nice list. For tip 27, the killer feedback loop I've recently discovered for building Web apps is to use the Chrome DevTools MCP server. This gives Claude the power to visually inspect the app in a running dev server as well find and fix errors that only happen under real condition on the client side.
Awesome tips! I’m a very casual user, but intend to dig in deeper. This list gives me lots of things to try out. Thanks for sharing!
Very helpful recommendations, thank you
Tip 20 about orchestrating worker Claude Code in containers changed my workflow. I run my main instance locally with --dangerously-skip-permissions in Docker for autonomous nightshift tasks. The tmux self-supervision pattern you mention? Game-changer for monitoring long-running jobs. I've been doing this for two months on production systems—full autonomy with exponential backoff when GitHub Actions choke.
One thing I'd add: containerized autonomy only works if you trust your test coverage.
Tip 0 about the token usage status bar is brilliant. I've been obsessing over context consumption too and ended up taking a different angle. Instead of monitoring and compacting after the fact, there's a plugin called context-mode that intercepts tool output before it ever enters context. Commands run in subprocesses, output gets indexed locally, and only summaries come back. Wrote it up here: https://reading.sh/how-one-plugin-cuts-claude-codes-context-bloat-by-98-096355e68166?sk=e06168222a38a98d4bf6add2daa10973