[FEATURE]:allow to deploy multipe agent

[wendy0005]

Feature hasn't been suggested before.

• I have verified this feature I'm about to request hasn't been suggested before.

Describe the enhancement you want to request

Summary

Allow running and managing multiple agents in the same OpenCode session.

Problem

Currently, it appears that only one agent can be active at a time. For larger tasks, I often need to split work into multiple independent subtasks, but I have to wait for one agent to finish before starting another.

Examples:

• One agent implements a feature.
• Another agent writes tests.
• Another agent updates documentation.
• Another agent reviews or refactors code.

This makes it difficult to fully utilize available compute and slows down development.

Proposed Solution

Support multiple concurrent agents within the same project/workspace.

Some possible capabilities:

• Launch multiple agents simultaneously.
• Give each agent its own task and context.
• View the status of all running agents.
• Pause, resume, or stop individual agents.
• Merge or cherry-pick results from each agent when they finish.
• Optionally limit the maximum number of concurrent agents.

Example Workflow

1. Agent A: Implement authentication.
2. Agent B: Write unit tests.
3. Agent C: Update documentation.
4. Agent D: Refactor related modules.

All agents work in parallel, and once they complete, I can review and merge their changes.

Benefits

• Faster development through parallel execution.
• Better utilization of available LLM resources.
• Reduced waiting time between independent tasks.
• Improved workflow for larger or more complex projects.
• Similar to having a small AI development team working concurrently.

Thanks for considering this feature!

[github-actions]

This issue might be a duplicate of existing issues. Please check:

• [FEATURE]: Multi-Agent Coding #9649: [FEATURE]: Multi-Agent Coding — requests multi-agent collaboration with agents handling different portions of work
• [FEATURE]:Support for multi-agent orchestration in isolated workspaces #17994: [FEATURE]: Support for multi-agent orchestration in isolated workspaces — requests spawning multiple agents in parallel with lifecycle management
• [FEATURE]: Ephemeral Sub-Agent Teams (parallel multi-agent orchestration) #19999: [FEATURE]: Ephemeral Sub-Agent Teams (parallel multi-agent orchestration) — proposes parallel sub-agent teams with concurrency, dependencies, and result merging
• [FEATURE]: Add 'opencode agents' command for background session management #27746: [FEATURE]: Add 'opencode agents' command for background session management — requests managing multiple background agent sessions
• [FEATURE]: Coordinator shell + fork mode #20368: [FEATURE]: Coordinator shell + fork mode — requests forking sessions to run multiple agents concurrently

Please review these issues and consider adding your thoughts there instead of opening a new one.

[abkzkg-oss]

Hi! Saw that single-agent execution is limiting your development speed. If you're looking to run independent agents in parallel and improve workflow and compute utilization, Mnemo could be worth a look.