Handoff skill

Compact Codex and Claude Code sessions into concise handoff documents for fresh LLM agent sessions.

Install

npx -y skills add realraelrr/handoff-skill

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

Copied from the file, not written here

Compact the current conversation into a handoff document for another agent to pick up.

1.4 KB, 292 tokens by cl100k_base, as published. Nobody here has run it

Write a concise handoff document that lets a fresh agent continue the current work without rereading the whole conversation.

Save it under the current project when possible:

In a git repo, use .state/handoffs/ under the repo root.
Outside a git repo, use the system temp directory.
Use mktemp for uniqueness and keep a .md extension. Avoid templates like handoff-XXXXXX.md; on macOS/BSD, create handoff-XXXXXX first, then rename it to add .md.

Use .state/ as agent operational workspace. Do not treat handoff files as product source or commit them by default.

The handoff must include:

Goal
Current state, including done, remaining, and only the key files/artifacts needed to continue
What worked
What didn't work or remains blocked
Verification, with exact commands and pass/fail signals
Next steps, starting with the next concrete action

Keep it concise. Reference existing plans, PRDs, ADRs, issues, commits, diffs, or other artifacts by path or URL instead of duplicating them.

After writing, read the handoff once to verify it is complete enough for a new agent to act from it, then report the saved path.

If the user passed arguments, treat them as a description of what the next session will focus on and tailor the doc accordingly.

3.6 KB alongside SKILL.md