agentsclimarketplace

Session handoff

Skill dominikwozniak/claude-kit/skills/session-handoff

🧰 Opinionated Claude Code starter kit β€” bootstrap, git, handoff, pre-commit

Install
npx -y skills add dominikwozniak/claude-kit --skill session-handoff

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

  • 1 stars1 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.

What its author says it does

Copied from the file, not written here

Compact the current conversation into a handoff document so another agent (fresh Claude Code session, Codex, etc.) can continue the work. Trigger phrases: "session handoff", "handoff", "summarise for next session", "prepare context for another agent", "wrap up so someone can pick this up".

SKILL.md

2.5 KB, 554 tokens by cl100k_base, as published. Nobody here has run it

Session Handoff

Write a continuation document so the next agent β€” fresh CC session, Codex, or a teammate's agent β€” can resume without re-reading the whole transcript.

Output location

Save to .agent/handoffs/<YYYYMMDD-HHMM>.md (project-local, gitignored). Create the directory if missing:

mkdir -p .agent/handoffs

Filename uses local time. Example: .agent/handoffs/20260522-1430.md.

What to include

  • Goal β€” one sentence: what we're trying to accomplish (use the user's argument if provided)
  • Current state β€” what's been done, what's working, what's broken
  • Open questions β€” decisions the next agent needs to make or surface to the user
  • Next steps β€” ordered list of concrete actions to take next
  • Suggested skills β€” Claude Code skills the next agent should invoke (e.g., git-workflow, spec-driven-development, debugging-and-error-recovery)
  • Pointers β€” references to existing artifacts (specs, plans, ADRs, PRs, issues) by path or URL. Don't duplicate their content
  • Gotchas / context β€” non-obvious things the next agent needs to know (env quirks, dependency versions, recent failures, etc.)

What to leave out

  • Don't re-summarise content already in committed files, PR descriptions, specs, plans, or ADRs. Reference them by path
  • Redact secrets, API keys, PII, internal URLs that shouldn't be shared
  • Skip narrative play-by-play of the conversation. Focus on actionable state

Document template

# Handoff β€” <YYYY-MM-DD HH:MM>

## Goal

<one sentence>

## Current state

- <what's done>
- <what's in progress>
- <what's blocked>

## Open questions

- <decision needed>

## Next steps

1. <action>
2. <action>
3. <action>

## Suggested skills

- `<skill-name>` β€” <why>

## Pointers

- spec: `.agent/spec-<name>.md`
- PR: <url or path>
- relevant files: `<path:line>`

## Gotchas

- <non-obvious context>

After writing

Tell the user:

Handoff saved to .agent/handoffs/<filename>. Open a new Claude Code session and run:

Read the handoff at .agent/handoffs/<filename> and continue from there.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.