When to use it

Use session-handoff when you are pausing a task, moving to a fresh conversation, or returning to work that already has a handoff. It is most useful when the next session needs to understand both the remaining work and the decisions that led there.

How it works

Figure 01

Resume from a checked record of the work.

  1. Pause

    Capture progress, decisions, and Git anchors.

  2. Compare

    Check the saved anchors against the live repository.

  3. Resume

    Surface drift before using the saved plan.

Result.ai/memory/HANDOFF.md
The saved handoff connects two sessions. Its Git anchors help identify when the repository has changed since the record was written.

The skill writes a structured handoff to .ai/memory/HANDOFF.md. It records the goal, plan status, decisions and their rationale, approaches that failed, next steps, and Git anchors identifying the working state.

On resume, it checks those anchors against the live repository. If the branch, files, or history have moved, it reports the mismatch before relying on the saved plan.

An example request

Ask your agent

“Create a handoff for the search migration. Include what is implemented, why we rejected dual writes, the failing check, and the next useful step.”

This illustrates a request you can adapt; it is not a transcript of a completed run.

The next session gets a concrete starting point: what to inspect, what remains unresolved, and which earlier decisions still need to hold.

Install and use

The skill’s identifier is session-handoff. It ships in the session-memory plugin. In Claude Code, first add the Overclock marketplace, then install the package:

/plugin install session-memory@overclock

For a standalone installation, keep the skill directory and its bundled resources together. The repository also includes per-skill Codex metadata. Read the full skill instructions for its exact workflow and supporting files.

What to keep in mind

A handoff preserves working context; it is not a full conversation archive. Repository changes can still require you to revise the plan.

The skill uses plain Markdown under .ai/memory/ and never commits it automatically. Review the saved content as you would any project file.

  • Lessons learned — Turn a correction into a reusable lesson, and keep repeated lessons up to date.
  • Solutions — Save a verified fix so the same problem is easier to solve next time.