Claude Code Launches Agent View: Manage Multiple Claude Coding Agents from a Single Dashboard — The Era of Parallel Tasks Begins L1
Confidence: High
Key Points: Anthropic released Claude Code v2.1.139 on May 11 and launched the "Agent View" research preview. Agent View is a unified dashboard built into the Claude Code CLI that lets you see the status of all sessions on a single screen: which agents are waiting for your input, which are still working, and which have finished. How to access: press the left arrow key in any session, or type `claude agents` in the terminal. Existing sessions can be added to the dashboard with `/bg`, and new tasks can be launched directly in the background with `claude --bg [task]`. Agent View is available immediately on Pro, Max, Team, Enterprise, and Claude API plans, but requires Claude Code v2.1.139 or later.
Impact: For individual Claude Code developers: evolves from 'one window per agent' to 'running N agents simultaneously' — throughput multiplies. For engineering leads: team members can advance multiple PRs and refactoring tasks at the same time. For teams building their own multiagent tools: reduces the case for custom builds further. For competing tools like Cursor and Continue.dev: differentiation pressure increases.
Detailed Analysis
Trade-offs
Pros:
- Native CLI integration — no new tools to learn
- Available immediately on all paid plans — no barriers
- Combined with the May 6 Code with Claude doubled quota, parallel task costs are affordable
- Forms a complete agent toolchain alongside Dreaming, Multiagent Orchestration, and Outcomes
Cons:
- Running multiple agents simultaneously can easily exceed daily token budgets
- Manual switching and monitoring of each session increases cognitive load
- If agents edit the same repo in parallel, the probability of merge conflicts rises
- Research Preview — stability has not been validated in long-term production use
Quick Start (5-15 minutes)
- Upgrade to Claude Code v2.1.139 (`npm i -g @anthropic-ai/claude-code` or via brew)
- In an existing session, press the left arrow key or run `claude agents` to enter Agent View
- Try running a long task in the background with `claude --bg "refactor the src/payments module"`
- Assign each agent a distinct working directory (e.g., frontend / backend / docs) to avoid conflicts
Recommendation
Claude Code subscribers are strongly encouraged to upgrade immediately and try the feature. Teams currently building their own multiagent frameworks should re-evaluate switching to Anthropic's native solution. New users should get comfortable with single-agent workflows before moving to Agent View.
Sources: Claude Official Blog (Official) | Claude Code Documentation (Documentation) | TestingCatalog (News)