# Studio Handoff

> Use when pausing, transferring, or reactivating game-studio work and a durable handoff must capture branch, goal, scope, files, commands, Verified Snapshot Unverified BLOCKED facts, failures, decisions, next actions, and a reactivation prompt.

- Skill: `hoatv2211/studio-handoff` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add hoatv2211/studio-handoff`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hoatv2211/studio-handoff/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- License: MIT
- Author: hoatv2211 (https://skillmd.com/u/hoatv2211)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/hoatv2211/studio-handoff

---

# Studio Handoff

## Overview
Create a restartable state transfer that distinguishes verified work from snapshots, hypotheses, and blockers.

## When to use
Use when a session ends, ownership changes, work becomes blocked, a subsystem crosses teams, or another agent must continue without repeating discovery.

## When NOT to use
Do not create a handoff as a substitute for running available verification or resolving a small task that can be completed safely now.

## Required inputs and context discovery
Collect repository/path, branch, goal, owned and excluded scope, current Git state, files touched, commands run, artifacts, decisions, failures, open hypotheses, dependencies, and next actions.

## Safety and risk level
Writing a handoff is low-risk. Preserve existing project handoff conventions, never expose credentials, and never rewrite unrelated history or other sessions’ ownership.

## Workflow
1. Re-read the task packet, applicable instructions, and current Git status.
   Completion criterion: the handoff reflects current state rather than memory.
2. List files touched and commands run with exact results.
   Completion criterion: every PASS claim maps to fresh evidence.
3. Separate Verified, Snapshot, Unverified, and BLOCKED facts.
   Completion criterion: no hypothesis appears under Verified.
4. Record failures, decisions, ownership boundaries, and do-not-touch paths.
   Completion criterion: the next session can avoid conflicting or unsafe work.
5. Provide three to seven ordered next actions and a reactivation prompt.
   Completion criterion: another session can start with one bounded command or inspection.

## Evidence and output contract
Produce `HANDOFF.md` with repository/path, branch, goal, scope ownership/do-not-touch, files touched, commands/tests, Verified results, Snapshot assumptions, Unverified hypotheses, BLOCKED items, failures, decisions, next actions, and Reactivation prompt.

## Handoff contract
The output is the handoff contract. Keep exact paths, symbols, ports, error strings, and commands intact; sanitize secrets and raw credentials.

## Pitfalls and anti-rationalization
- Do not say “tests pass” without the command and fresh output.
- Do not omit dirty or untracked files.
- Do not bury blockers in prose.
- Do not prescribe next actions that require unavailable permissions without labeling them BLOCKED.

## Verification checklist
- [ ] Repository, branch, goal, and scope are current.
- [ ] Files and commands are exact.
- [ ] Verified, Snapshot, Unverified, and BLOCKED are separate.
- [ ] Failures and decisions are explicit.
- [ ] Three to seven next actions and a Reactivation prompt exist.

## References and scripts
Follow the project's existing `HANDOFF.md` and `.agents/CONTRACT.md` when present; otherwise use the handoff contract in `AGENTS.md`.