Context engineering

Handoff

A handoff is passing work from one session or agent to the next by summarising the current state, so the successor can continue without relearning everything. It is the antidote to a dead or overflowing window.

James Phoenix
Understanding Data Updated July 2, 2026

A handoff is how work survives the end of a session. Context windows fill up, conversations get long and confused, and sometimes you just want a fresh start. A handoff captures where things stand so the next session, or the next agent, can continue without relearning everything from scratch.

Why you need one

The model remembers nothing between sessions, so if you clear the context or the window overflows, whatever the agent worked out is gone. A handoff is the deliberate save point: you write down the state before it disappears.

It is the healthier alternative to letting a window degrade. Rather than pushing a bloated session further and hoping compaction keeps the right details, you summarise on purpose and start clean.

A good handoff usually captures:

  • What is done and what is confirmed working.
  • What is next and what is blocked.
  • The key decisions and why they were made.

That summary, written down, is the handoff artifact the successor reads.

Note
The best moment for a handoff is before you need one. Once a window is confused, the summary it produces is confused too. Hand off while the agent still has a clear picture.

Related terms

Building with AI agents?

This dictionary is part of how I think about agentic engineering. If you want the same thinking applied to your codebase, that is what I do.

See how I can help