From X:
- Many (1-5) sessions in different tabs, with system notifications (Optimize your terminal setup - Claude Code Docs)
- 5-10 session in the cloud Claude Code on the Web
- ~~send tasks from your terminal to run on the web with the
&prefix ~~ - teleport web sessions back to your terminal to continue locally.
- ~~send tasks from your terminal to run on the web with the
- Use Opus 4.5 (or I guess in general frontier models)
- Claude Code GitHub Actions to edit Claude.md during code review (i.e. Compounding Engineering)
-
Most sessions start in Plan mode (shift+tab twice). If my goal is to write a Pull Request, I will use Plan mode, and go back and forth with Claude until I like its plan. From there, I switch into auto-accept edits mode and Claude can usually 1-shot it. A good plan is really important!
- Slash commands for repeated prompting 7. Note that Custom slash commands have been merged into skills.1
-
I use a few subagents regularly: code-simplifier simplifies the code after Claude is done working, verify-app has detailed instructions for testing Claude Code end to end, and so on.2
- Use a PostToolUse3 hook to format Claude’s code
- Use /permissions to pre-allow common bash commands, and check into
.claude/settings.json - Using Tools
- Slack via MCP (MCP configuration is checked into our
.mcp.json) - BigQuery using bq CLI
- Slack via MCP (MCP configuration is checked into our
-
For very long-running tasks, I will either (a) prompt Claude to verify its work with a background agent when it’s done, (b) use an agent Stop hook to do that more deterministically, or (c) use the ralph-wiggum4 plugin. I will also use either
--permission-mode=dontAskor--dangerously-skip-permissionsin a sandbox to avoid permission prompts for the session, so Claude can cook without being blocked on me. - Give Claude a way to verify its work
- Claude Chrome extension: It opens a browser, tests the UI, and iterates until the code works and the UX feels good.