claude-handoff

Launches a background Claude agent seeded with a conversation handoff summary.

Updated Aug 22, 2026
One-click install
npx skills add https://github.com/MSC72m/DevForge --skill claude-handoff-msc72m
Or copy as Structured Prompt for Agentβ–Ό
Please help me install this Agent Skill.
Skill: claude-handoff
Source: https://github.com/MSC72m/DevForge/tree/main/skills/claude-handoff
Command: npx skills add https://github.com/MSC72m/DevForge --skill claude-handoff-msc72m

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Long coding sessions lose context when they end, forcing the next agent to rediscover decisions, file paths, and progress. This Skill captures the current conversation state and immediately continues the work in a fresh background agent. ## Core Features & Use Cases - Handoff Summary Generation: Writes a concise summary of the current conversation, referencing existing artifacts (specs, plans, ADRs, commits) by path instead of duplicating them. - Background Agent Launch: Starts a new agent via claude --bg --name "<descriptive name>" in the current working directory, returning control immediately while work continues. - Sensitive Data Redaction: Strips API keys, passwords, and personally identifiable information from the summary before it becomes the new agent's prompt. - Use Case: You have been debugging a login issue for hours and need to stop. Invoke the handoff, and a named background agent picks up the investigation with full context and suggested skills to call. ## Quick Start Ask the assistant to hand off the current conversation to a background agent focused on fixing the login bug.

Frequently Asked Questions about claude-handoff

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I continue a Claude Code session in a new agent?β–Ό

Invoke the handoff to write a summary of the current conversation, then a background agent launches automatically with that summary as its prompt via claude --bg. It starts in the current working directory and returns immediately.

How do I manage background agents after a handoff?β–Ό

Use the claude agents command to view and manage running background agents. Each handoff passes a descriptive --name flag, so agents appear with readable names in the job list, session picker, and terminal title.

Does the handoff summary include sensitive information?β–Ό

No. The Skill explicitly redacts API keys, passwords, and personally identifiable information before the summary becomes the new agent's prompt, since the prompt content is visible in the agent's session.

Can I specify what the next session should focus on?β–Ό

Yes. Pass arguments describing the next session's purpose, and the summary is tailored accordingly. The argument hint prompts you for what the next session will be used for.

Why does the handoff reference files instead of copying content?β–Ό

Duplicating specs, plans, ADRs, issues, commits, and diffs would bloat the prompt and risk stale copies. Referencing artifacts by path or URL keeps the summary concise and lets the next agent read current versions.