What happens if an agent mission fails?

If a mission fails, DevFleet provides a structured report detailing the errors and steps taken, allowing you to review the state before retrying or adjusting the plan.

Do I need a separate server to run DevFleet?

Yes, this skill requires a running Claude DevFleet instance (typically on localhost:18801) connected to Claude Code via the Model Context Protocol (MCP).

How does Claude DevFleet handle code conflicts?

Each agent operates in its own isolated git worktree. Upon completion, changes are automatically merged into the main branch; if a conflict occurs, the changes remain in the worktree for manual resolution.

How many agents can run simultaneously?

By default, DevFleet supports 3 concurrent agents, but this capacity is configurable via the DEVFLEET_MAX_AGENTS environment variable.

Can I set dependencies between different missions?

Yes, DevFleet uses a Directed Acyclic Graph (DAG) structure allowing missions to wait for specific parent tasks to complete before auto-dispatching.

Claude DevFleet

Name: Claude DevFleet
Author: affaan-m

byaffaan-m

•

172,007

•

生産性とワークフロー

Orchestrates multiple parallel AI agents to execute complex development tasks through isolated git worktrees and automated mission management.

Claude DevFleet transforms individual AI coding sessions into a coordinated multi-agent system, allowing developers to decompose large projects into a Directed Acyclic Graph (DAG) of missions. It manages parallel execution across isolated git worktrees, handles automatic merges upon task completion, and provides comprehensive status monitoring via a central dashboard. This skill is essential for large-scale refactors, building full-stack features from scratch, or any scenario where complex workflows can be accelerated by offloading sub-tasks to specialized, concurrent AI agents.

主な機能

01Dependency-based planning (DAG) for complex sequential or parallel workflows

02Automated merging of completed agent tasks with manual conflict resolution fallbacks

03Real-time progress monitoring and centralized structured mission reporting

04Multi-agent task orchestration with automated mission dispatching

05Isolated execution environments using git worktrees to prevent code conflicts

06172,007 GitHub stars

ユースケース

01Executing massive codebase refactorings across multiple modules in parallel

02Automating complex project initialization where multiple boilerplate and configuration tasks must run concurrently

03Building full-stack features by simultaneously dispatching agents for backend, frontend, and testing

主な機能

01Dependency-based planning (DAG) for complex sequential or parallel workflows

02Automated merging of completed agent tasks with manual conflict resolution fallbacks

03Real-time progress monitoring and centralized structured mission reporting

04Multi-agent task orchestration with automated mission dispatching

05Isolated execution environments using git worktrees to prevent code conflicts

06172,007 GitHub stars

ユースケース

01Executing massive codebase refactorings across multiple modules in parallel

02Automating complex project initialization where multiple boilerplate and configuration tasks must run concurrently

03Building full-stack features by simultaneously dispatching agents for backend, frontend, and testing