A Claude Code execution mode where the agent maps out the full approach before touching files, asking clarifying questions until it reaches high confidence. Triggered via /plan, shift-tab, or a claude.md rule like ‘do not make changes until you have 95% confidence’. Ultra Plan is a cloud-hosted variant (research preview) that offloads the planning to an Anthropic container running Opus 4.6 with three parallel exploration agents plus a critique agent, delivers a structured doc-style plan on the web with comment and emoji reactions, and teleports the approved plan back to the terminal for execution.
Key points
- Plan mode is the single biggest source of token savings by preventing wrong-path execution
- Ultra Plan runs on cloud Opus 4.6 with 3 parallel exploration + 1 critique agent
- Ultra Plan has a 30-minute cloud cap and requires the project to be synced to a Git repo
- Executing an Ultra Plan often runs faster locally because the plan is cleaner
- Ultra Plan not accessible from desktop app or VS Code extension — CLI only
- Comments and emoji reactions available on cloud-hosted plans for richer review
Related entities
Related concepts
Source references
- [src-004] Nate Herk cluster — Nate Herk — Claude Code cluster (21 videos)
– Videos referenced: T4fXb3sbJIo, 49V-5Ock8LU, tXtCK66fPj8
Keep reading from this thread
From 491 indexed pages and articles.
- Wiki concept Claude Code Scheduled Tasks Desktop-app-only feature that lets Claude Code run agents on a cron schedule with persistent storage and catchup. Related by claude
- Wiki concept Claude Code Dynamic Workflows A high-parallelism orchestration mode introduced alongside Claude Opus 4.8, where Claude generates an executable JavaScript workflow that Related by claude
- Insight Agentic Web Readiness A practical checklist for making websites understandable, navigable, and useful for AI agents and answer engines Readers have engaged with this next