Claude Code Week 15 (April 6–10, 2026): /ultraplan, Monitor Tool, /autofix-pr, /team-onboarding

Anthropic's Week 15 update (April 6–10, 2026, v2.1.92 → v2.1.101) brought four headline features to Claude Code: /ultraplan for cloud plan mode driven from the terminal, the new Monitor tool for streaming background events into the conversation, /autofix-pr to enable PR auto-fix without leaving the CLI, and /team-onboarding to generate a ramp-up guide from your local Claude Code usage. Together they push more long-running and asynchronous work out of the foreground turn and into the cloud or background.

TL;DR

• /ultraplan (research preview): runs plan mode in a Claude Code on the web session; your terminal stays free. Review in the browser, comment on sections, ask for revisions, then execute remotely or send the plan back to your CLI. v2.1.101 auto-creates a default cloud environment on first run.
• Monitor tool (v2.1.98): a built-in background watcher that streams events into the conversation as new transcript messages — pairs with /loop, which now self-paces.
• /autofix-pr: enables Claude Code on the web's PR auto-fix for the open PR on your current branch in one terminal command.
• /team-onboarding (v2.1.101): generates a teammate ramp-up guide from your local Claude Code usage.

Quick Answer

Update Claude Code to v2.1.101 or later, push your branch, and try /ultraplan for a non-trivial change you have not started yet. For long-running work — a training run, a flaky CI, a dev server you want watched — ask Claude to use the Monitor tool instead of polling. If you have an open PR, run /autofix-pr from its branch and walk away.

What Shipped in Week 15

/ultraplan: Cloud Plan Mode From the Terminal

/ultraplan runs plan mode in the cloud rather than blocking your local session. Claude drafts the plan inside a Claude Code on the web session while your terminal stays free for other work; when the plan is ready you review it in the browser, comment on individual sections, ask for revisions, and then either execute remotely or send the plan back to your CLI for local execution.

Run it in a session, or include the keyword in any prompt:

> /ultraplan migrate the auth service from sessions to JWTs

v2.1.101 removes the setup tax: the first run auto-creates a default cloud environment, so there is no web onboarding step before you can try it.

This is a different shape from the local plan mode most users know — it's closer in spirit to handing a senior engineer a brief and getting back a reviewed proposal you can act on. For background on plan mode mechanics, see our AI coding agents explainer.

Monitor Tool: Background Events Without Sleep Loops

The new built-in Monitor tool spawns a background watcher and streams its events into the conversation. Each event lands as a new transcript message that Claude reacts to immediately. Practical uses:

• Tail a training run and respond to loss spikes or crashes
• Babysit CI on a PR and react to failures the moment they appear
• Auto-fix a dev server crash the instant it happens

The point is what it replaces: you no longer need a Bash sleep loop holding the turn open while you wait for something to change.

Ask Claude in plain language:

> Tail server.log in the background and tell me the moment a 5xx shows up

Monitor pairs with /loop, which now self-paces: omit the interval and Claude schedules the next tick based on the task — or reaches for the Monitor tool to skip polling altogether.

> /loop check CI on my PR

/autofix-pr: PR Auto-Fix From Your Terminal

PR auto-fix landed on the web in Week 13. Week 15 lets you turn it on without leaving your terminal. /autofix-pr infers the open PR for your current branch and enables auto-fix for it on Claude Code on the web in one step. After that, Claude watches CI and review comments and pushes fixes until the PR is green — push your branch, run the command, walk away.

> /autofix-pr

/team-onboarding: Replay Your Setup for a New Teammate

/team-onboarding generates a teammate ramp-up guide from your local Claude Code usage. Run it inside a project you know well and hand the output to a new teammate so they can replay your setup instead of starting from defaults. It's a sharp answer to the perennial "how is this person actually using Claude Code day to day?" question — the kind of context that usually lives only in someone's head.

> /team-onboarding

Other Notable Wins in Week 15

A few smaller items worth knowing about:

• Focus view — Ctrl+O in flicker-free mode collapses the view to your last prompt, a one-line tool summary with diffstats, and Claude's final response.
• /agents tabbed layout with a Running tab showing live subagents and a ● N running count.
• /cost now shows a per-model and cache-hit breakdown for subscription users.
• /release-notes is now an interactive version picker.
• Default effort level is now high for API-key, Bedrock, Vertex, Foundry, Team, and Enterprise users (control with /effort).
• Guided Bedrock and Vertex AI setup wizards on the login screen.
• OS CA certificate store is now trusted by default, so enterprise TLS proxies work without extra setup (CLAUDE_CODE_CERT_STORE=bundled to opt out).
• Hardened Bash tool permissions: backslash-escaped flags, env-var prefixes, /dev/tcp redirects, and compound commands now prompt correctly.

How This Compares to Other Agents

For deeper context see our Claude Code vs Aider guide, the Cursor vs Claude Code comparison, and the Claude Code tool page.

Sources

• Claude Code "What's New" — Week 15 (official): code.claude.com/docs/en/whats-new/2026-w15
• Claude Code changelog v2.1.92 → v2.1.101: code.claude.com/docs/en/changelog
• Claude Code on the web — Auto-fix pull requests: code.claude.com/docs/en/claude-code-on-the-web

For current model guidance, see our Claude Opus 4.6 vs Opus 4.5 comparison and the beginner's guide to AI models.