Why is a worker handover necessary in Claude Code?

AI agents have finite turn limits (often 100 turns). A handover ensures that when one worker stops, the next one understands the exact state of the code, tests, and logic without having to re-discover everything from scratch.

Where are the handover files stored?

Handover files are saved in the .orchestrator/ directory of your repository, typically named handover-[ISSUE_NUMBER].md for easy retrieval by the next worker.

What happens if a worker reaches 100 turns without a handover?

The session will automatically exit. Without the handover file created by this skill, the replacement worker may struggle to understand recent uncommitted changes or the 'why' behind specific implementation choices.

When should the handover process begin?

According to the skill guidelines, you should evaluate the need for a handover at 85 turns, begin preparation at 90 turns, and complete the file by turn 95 to ensure a clean exit.

Worker Context Handover

Name: Worker Context Handover
Author: troykelly

bytroykelly

生产力与工作流

Standardizes context transfer between AI agents by generating structured handover files when turn limits are reached.

关于

This skill manages the seamless transition between sequential AI workers by creating detailed handover documentation when turn limits are approached. It captures critical metadata, current branch status, test results, architectural decisions, and specific next steps, ensuring that replacement workers can resume complex, long-running tasks with full context and zero friction. By formalizing how progress is recorded and verified, it prevents work duplication and context loss in autonomous development workflows.

主要功能

Automated turn limit monitoring and transition notifications
Verification protocols for incoming workers to acknowledge and resume work
Clear task breakdown including completed, in-progress, and remaining items
0 GitHub stars
Standardized Markdown handover file generation with metadata and state tracking
Comprehensive documentation of architectural decisions and failed approaches

使用场景

Providing a clear audit trail of decisions made by autonomous agents during issue resolution
Preserving technical context and 'lessons learned' across different orchestration sessions
Managing long-running software development tasks that exceed a single AI session turn limit

关于

主要功能

Automated turn limit monitoring and transition notifications
Verification protocols for incoming workers to acknowledge and resume work
Clear task breakdown including completed, in-progress, and remaining items
0 GitHub stars
Standardized Markdown handover file generation with metadata and state tracking
Comprehensive documentation of architectural decisions and failed approaches

使用场景

Providing a clear audit trail of decisions made by autonomous agents during issue resolution
Preserving technical context and 'lessons learned' across different orchestration sessions
Managing long-running software development tasks that exceed a single AI session turn limit