Async Workflows: Claude Code Background Agents & Parallel Execution

Problem: When Claude Code spawns a sub-agent for research or complex analysis, your entire session blocks. You wait while the sub-agent works, unable to continue the conversation.

Quick Win: When Claude spawns a sub-agent, press Ctrl+B to move it to the background:

You: Research authentication best practices for our Next.js app
Claude: I'll spawn a sub-agent to research this...
[Sub-agent starts]
You: [Press Ctrl+B]
You: While that runs, let's work on the database schema...

Your session continues. The sub-agent works independently and surfaces results when done.

Background Agents: True Parallel Execution

Claude Code now supports asynchronous agent execution. When the main agent spawns sub-agents, you can background them and keep working with Claude on other tasks.

The workflow:

1. Request: Ask Claude to handle a task that benefits from a sub-agent
2. Claude spawns: The main agent creates a sub-agent for the work
3. Background it: Press Ctrl+B while the sub-agent runs
4. Continue working: Keep chatting with the main agent on other tasks
5. Auto-resume: When the sub-agent completes, it wakes up the main agent with results

Check running agents anytime with /tasks:

/tasks

You'll see each background agent's status, token usage, and progress. Click any agent to inspect details.

When Background Agents Excel

Perfect for background execution:

• Research tasks requiring web searches
• Code analysis across large codebases
• Documentation generation
• Security audits and vulnerability scans
• Performance profiling reports

Keep in foreground:

• Tasks requiring your immediate input
• File modifications you need to review
• Anything with sequential dependencies on your current work

Background Bash Commands

The same pattern works for long-running shell commands. When Claude runs npm install, docker build, or ffmpeg and it takes a while, background it:

Claude: Running npm install...
[Command starts]
You: [Press Ctrl+B]
You: While that installs, can you review the API routes?

Monitor with /tasks, same as agents.

The --agent Flag: Debug Your Sub-Agents

New CLI flag lets you run Claude Code AS any sub-agent:

claude --agent plan

This launches Claude Code with the planning agent's configuration. Ask it questions, test its behavior, verify it works as expected before deploying in workflows.

Use cases:

• Debug custom sub-agents before production use
• Test agent instructions and tool access
• Run specialized agents manually for one-off tasks
• Understand built-in agent capabilities

Works with built-in agents (plan, explore, etc.) and your custom agents in .claude/agents/.

Also in This Release

Anthropic shipped several other features alongside background agents:

Instant compaction: The /compact command now executes immediately. Claude maintains a continuous session memory, so compaction just loads the summary into a fresh context. Learn more in our context management guide.

Session memory architecture: Every session now maintains a structured summary including status, completed work, discussion points, and a work log. See memory optimization for details.

Stats dashboard: Run /stats to see your usage patterns, favorite models, token counts, and streaks. Press Ctrl+S to copy for sharing.

Session naming: Use /rename to name sessions, then resume by name with claude --resume session-name. The /resume screen now groups forked sessions and adds keyboard shortcuts (P preview, R rename, B browse forks).

MCP quick toggle: Enable or disable MCP servers without editing config files. See our MCP basics guide for the commands.

Slack integration: Delegate tasks to Claude Code directly from Slack channels. Tag @Claude with bug reports or feature requests for team workflows.

Common Issues

Agent doesn't background: Press Ctrl+B while the agent is actively running, not after it completes.

Lost track of agents: Use /tasks to see all running background processes with their IDs.

Agent completed but no results: The AgentOutputTool automatically surfaces results. If missed, check /tasks for the completed agent's output.

Next Steps

• Foundation: Understand agent fundamentals before building complex workflows
• Design patterns: Learn sub-agent design for effective task delegation
• Distribution: Master task distribution for multi-agent orchestration
• Custom agents: Build specialized agents with custom agent patterns

Background agents transform Claude Code from a turn-based assistant into a parallel development environment. Launch research, continue building, get results when ready.