Claude Code feature where Claude writes orchestration scripts that run tens to hundreds of parallel subagents with verification.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Dynamic workflows is a Claude Code feature from Anthropic that lets Claude plan a task dynamically, split it into subtasks, and fan the work out across tens to hundreds of parallel subagents in a single session. Results are checked before they are merged: independent agents attempt the problem from different angles, and other agents try to refute the findings until answers converge. Progress is saved as the run goes, so an interrupted job resumes instead of restarting, and runs can last hours or days.
Anthropic positions it for work too large for a single agent pass: codebase-wide bug hunts, security and optimization audits, large migrations and language ports, and stress-testing plans. Its featured example is the Bun port from Zig to Rust (about 750,000 lines of Rust, 99.8% of the existing test suite passing, eleven days from first commit to merge), as reported by Anthropic. No independent benchmarks were provided.
A workflow can be started by asking Claude to create one, or by enabling the Claude Code-specific ultracode setting in the effort menu, which sets effort to xhigh and lets Claude decide when to use a workflow. Auto mode is recommended. The first time a workflow triggers, Claude Code shows what will run and asks for confirmation. Workflows can consume substantially more tokens than a typical session, so starting with a scoped task is advised, and organization admins can disable them through managed settings.
It is generally available in the Claude Code CLI, Desktop and VS Code extension on Pro, Max, Ultra, Team and Enterprise plans, and via the Claude API, Amazon Bedrock, Vertex AI and Microsoft Foundry. It is on by default for Max, Ultra, Team, Enterprise and API use; Pro users enable it in /config.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Claude Code feature that plans dynamically, splits a prompt into subtasks, fans them out to parallel subagents, and has agents cross-check and refute each other until answers converge. The creator hasn't tested it yet.
Research preview feature where Claude writes orchestration scripts that run tens to hundreds of parallel sub-agents. It verifies results adversarially and targets bug hunts, large migrations and audits. It can use substantially more tokens, and a new 'ultra code' setting sets effort to extra high. No benchmarks were given.
Closest neighbours by meaning, not just by category.
Agent · Anthropic
Claude Code mode that fans out many parallel sub-agents for deep tasks such as full codebase reviews.
App · Anthropic
Anthropic's Claude assistant and agentic coding tool, with chat apps, Projects, skills, connectors and Claude Code.
App · Anthropic
Claude Code CLI view that lists and manages multiple parallel Claude Code sessions, showing which need input or are done.
App · Anthropic
In-app browser in the Claude Code desktop app with element annotation for requesting edits to web pages.
App · Anthropic
Claude Code feature that lets separate Claude Code sessions send each other messages and summaries to share context.
App · Anthropic
Anthropic's Claude Cowork agent running in the cloud on web and mobile, continuing tasks in the background.
