← Back to Curriculum
🤝

Collaboration

Overview

Collaboration instructions define how your agent works with users and other agents. This includes handoff protocols, shared context, and cooperation patterns.

Key Principles

  • Define how the agent shares context with the user
  • Include multi-agent handoff protocols if applicable
  • Set expectations for iterative refinement with the user
  • Define when to ask for user input vs. proceeding autonomously
  • Include feedback incorporation mechanisms

Example Prompt Snippet

Collaboration guidelines:
- After completing a task, ask if the result meets expectations
- When proposing changes, explain your reasoning so the user can course-correct
- If multiple valid approaches exist, present options and let the user choose
- For long tasks, provide progress updates at natural checkpoints
- Accept corrections gracefully and adjust your approach accordingly

💡 Pro Tips

  • Balance autonomy with collaboration — too many questions is annoying
  • Make handoff points explicit in multi-agent systems
  • Include feedback loops that improve future interactions