subagent-handoff

Community

Align subagents to user outcomes, safely.

Authorjonaseck2
Version1.0.0
Installs0

System Documentation

What problem does it solve?

It prevents subagent work from being marked complete when it does not actually move toward the plan’s user-observable outcome, even if it looks spec-correct or “tests pass.”

Core Features & Use Cases

  • User-goal-first prompting: Forces every subagent prompt to start from the plan’s User goal and to stop/escalate if the task fails that goal.
  • Verification-before-done: Requires the dispatcher to perform explicit user-observable checks (UI/MCP/file/gazetteer resolution) before marking tasks done.
  • Consistent reviewer/implementer reporting: Standardizes reports into three required answers to reduce omission and optimism.
  • Git worktree guardrails: Mandates rebase + branch verification to avoid CWD drift and accidental commits to the wrong branch.
  • Project-local MCP/tool alignment: Ensures outcome checks match the project’s built-in MCP + genealogy-research context.

Quick Start

Use the subagent-handoff skill templates to dispatch implementer, spec-reviewer, and code-quality-reviewer subagents for a task, then run an explicit user-observable verification step before marking the task done.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: subagent-handoff
Download link: https://github.com/jonaseck2/slaktforskning/archive/main.zip#subagent-handoff

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.