How do I prevent timeout errors during delegation?

You should increase the MCP tool timeout for the codex-specialized-subagents server to 600 seconds in your config.toml file to allow sub-agents enough time to complete complex tasks.

When does Delegation Autopilot trigger automatically?

It triggers for multi-step requests (using words like 'and' or 'then'), cross-cutting changes affecting multiple files, or when code changes require simultaneous updates to tests and documentation.

Can sub-agents perform research without editing my files?

Yes, by setting the sandbox parameter to 'read-only', sub-agents can perform multi-step analysis and research without making any changes to your workspace.

Does this skill support recursive delegation?

No, for safety and performance, delegated sub-agents are prohibited from calling delegation tools themselves, preventing infinite loops.

Delegation Autopilot

Name: Delegation Autopilot
Author: leonardsellem

byleonardsellem

•

生产力与工作流

Automates the orchestration of complex coding tasks by delegating them to specialized, isolated sub-agents.

关于

Delegation Autopilot enhances Claude Code by automatically identifying multi-step or cross-cutting requests and routing them to specialized sub-agents. It streamlines the development process for complex changes—such as simultaneous updates to code, tests, and documentation—by decomposing tasks into manageable units, selecting the appropriate tools for each, and aggregating the results back to the parent agent. This ensures high-quality execution for research-heavy or architecturally significant changes without manual management of sub-tasks.

主要功能

Isolated sub-agent execution for focused code changes
Integrated result aggregation and artifact tracking
Support for both write-access and read-only research modes
Automatic multi-step task decomposition and routing
Automated repo-specific and global skill selection
31 GitHub stars

使用场景

Performing multi-step architectural research and code analysis using read-only sub-agents.
Refactoring cross-cutting concerns across multiple modules while simultaneously updating tests.
Implementing new features that require synchronized changes to logic, documentation, and API specifications.

关于

主要功能

Isolated sub-agent execution for focused code changes
Integrated result aggregation and artifact tracking
Support for both write-access and read-only research modes
Automatic multi-step task decomposition and routing
Automated repo-specific and global skill selection
31 GitHub stars

使用场景

Performing multi-step architectural research and code analysis using read-only sub-agents.
Refactoring cross-cutting concerns across multiple modules while simultaneously updating tests.
Implementing new features that require synchronized changes to logic, documentation, and API specifications.