# Handover

> Prepare a handover for continuing work in a new session. Use when the user asks to "hand over", "prepare handover", "wrap up for a new session", or "context dump".

- **Type:** Skill
- **Install:** `agentstack add skill-gjoranv-claude-plan-skills-handover`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [gjoranv](https://agentstack.voostack.com/s/gjoranv)
- **Installs:** 0
- **Category:** [Developer Tools](https://agentstack.voostack.com/c/developer-tools)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [gjoranv](https://github.com/gjoranv)
- **Source:** https://github.com/gjoranv/claude-plan-skills/tree/main/handover

## Install

```sh
agentstack add skill-gjoranv-claude-plan-skills-handover
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

Prepare everything needed for a new conversation to continue the current work seamlessly.

1. **Gather state**:
   - Current repo and branch
   - Uncommitted changes (`git status`, `git diff --stat`)
   - Recent commits on this branch
   - The plan issue (from argument, or ask the user)
2. **Summarize the session**: What was done, key decisions made, what's still in progress, any blockers or open questions.
3. **Offer to persist state** before handing over:
   - Update the plan issue (`/gh-update-plan`)
   - Commit uncommitted changes
   Ask the user which of these to do. Skip any already done.
4. **Capture learnings**: Review the conversation for surprises, gotchas, corrections, and non-obvious decisions. Propose where each belongs:
   - **Instructions file** (CLAUDE.md, AGENTS.md, or similar): conventions, build quirks
   - **Repo docs**: architecture insights, design rationale
   - **Agent memory**: cross-project patterns, user preferences

   User approves, edits, or drops each suggestion.
5. **Generate a handover prompt** that the user can paste into a new session. Include:
   - The plan issue reference (e.g. `/gh-read-plan owner/repo#12`)
   - The repo and branch to work in
   - A 3-5 sentence summary of where things left off
   - The specific next step to work on
   - Any gotchas or context the new session needs (e.g. "the tests pass locally but CI needs a config change first")

Rules:
- Keep the handover prompt concise. It should be copy-pasteable in one go.
- Do not include implementation details the plan issue already captures.
- If there's no plan issue, still generate a useful handover prompt with the repo, branch, and summary.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [gjoranv](https://github.com/gjoranv)
- **Source:** [gjoranv/claude-plan-skills](https://github.com/gjoranv/claude-plan-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-gjoranv-claude-plan-skills-handover
- Seller: https://agentstack.voostack.com/s/gjoranv
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.
