Skip to main content
skill CX Codex

create-handoff — Claude Code skill: hand work off to a new session

The skill writes a YAML handoff document so the next agent session picks up your task without losing context.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add parcadei/Continuous-Claude-v3 --skill create-handoff -a claude-code
npx skills add parcadei/Continuous-Claude-v3 --skill create_handoff -a claude-code
Open on GitHub

What create-handoff does

create-handoff compacts a session into a single YAML file. The agent works through steps: it reads the most recently modified folder under thoughts/shared/handoffs/ to get the session name, builds a path like YYYY-MM-DD_HH-MM_description.yaml, and fills in the fields. With no existing handoffs, the folder is general.

The format is strict. goal: and now: are required because the statusline parser reads them, and renaming them breaks it. The rest covers what got done with file references, decisions and their rationale, findings, what worked, what failed, and the next steps. The author suggests pointing at code as /path/to/file.ext:line instead of pasting large diffs.

Once the file is written, the skill asks how the session went through AskUserQuestion (SUCCEEDED, PARTIAL_PLUS, PARTIAL_MINUS, FAILED) and indexes the document with uv run python scripts/core/artifact_index.py. That step needs uv and an opc directory at the project root — without them the outcome marking fails.

What it helps with

  • hand an unfinished task to an agent in a fresh session
  • record decisions and the reasoning behind them
  • note which approaches worked and which did not
  • mark the session outcome in the database
  • list next steps with file references

Example prompts for the agent

  • create a handoff for this session
  • create-handoff: capture what we did and what comes next
  • write a handoff document for the next session

How to install create-handoff

1Skills CLI (Vercel)

npx skills add parcadei/Continuous-Claude-v3 --skill create-handoff -a claude-code
npx skills add parcadei/Continuous-Claude-v3 --skill create_handoff -a claude-code

2Manually: put the files in place

~/.claude/skills/create_handoff/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/create_handoff/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about create-handoff

How do I install create-handoff in Claude?

Use the skills CLI: npx skills add parcadei/Continuous-Claude-v3 --skill create-handoff -a claude-code. The skill file then lives at ~/.claude/skills/create_handoff/SKILL.md.

What does create-handoff do?

It writes a YAML document covering the goal, the current step, completed tasks with files, decisions, findings and next steps. Another agent session reads it to continue the work.

What is needed to mark the session outcome?

The skill asks for the outcome via AskUserQuestion, then indexes the file with uv run python scripts/core/artifact_index.py from the opc directory at the project root. It needs uv and that directory; the database is auto-detected — PostgreSQL if configured, SQLite otherwise.

Source text

SKILL.md · MIT
---
name: create-handoff
description: Create handoff document for transferring work to another session
---

# Create Handoff

You are tasked with writing a handoff document to hand off your work to another agent in a new session. You will create a handoff document that is thorough, but also **concise**. The goal is to compact and summarize your context without losing any of the key details of what you're working on.


## Process
### 1. Filepath & Metadata
Use the following information to understand how to create your document:

**First, determine the session name from existing handoffs:**
```bash
ls -td thoughts/shared/handoffs/*/ 2>/dev/null | head -1 | xargs basename
```

This returns the most recently modified handoff folder name (e.g., `open-source-release`). Use this as the handoff folder name.

If no handoffs exist, use `general` as the folder name.

**Create your file under:** `thoughts/shared/handoffs/{session-name}/YYYY-MM-DD_HH-MM_description.yaml`, where:
- `{session-name}` is from existing handoffs (e.g., `open-source-release`) or `general` if none exist
- `YYYY-MM-DD` is today's date
- `HH-MM` is the current time in 24-hour format (no seconds needed)
- `description` is a brief kebab-case description

**Examples:**
- `thoughts/shared/handoffs/open-source-release/2026-01-08_16-30_memory-system-fix.yaml`
- `thoughts/shared/handoffs/general/2026-01-08_16-30_bug-investigation.yaml`

### 2. Write YAML handoff (~400 tokens vs ~2000 for markdown)

**CRITICAL: Use EXACTLY this YAML format. Do NOT deviate or use alternative field names.**

The `goal:` and `now:` fields are shown in the statusline - they MUST be named exactly this.

```yaml
---
session: {session-name from ledger}
date: YYYY-MM-DD
status: complete|partial|blocked
outcome: SUCCEEDED|PARTIAL_PLUS|PARTIAL_MINUS|FAILED
---

goal: {What this session accomplished - shown in statusline}
now: {What next session should do first - shown in statusline}
test: {Command to verify this work, e.g., pytest tests/test_foo.py}

done_this_session:
  - task: {First completed task}
    files: [{file1.py}, {file2.py}]
  - task: {Second completed task}
    files: [{file3.py}]

blockers: [{any blocking issues}]

questions: [{unresolved questions for next session}]

decisions:
  - {decision_name}: {rationale}

findings:
  - {key_finding}: {details}

worked: [{approaches that worked}]
failed: [{approaches that failed and why}]

next:
  - {First next step}
  - {Second next step}

files:
  created: [{new files}]
  modified: [{changed files}]
```

**Field guide:**
- `goal:` + `now:` - REQUIRED, shown in statusline
- `done_this_session:` - What was accomplished with file references
- `decisions:` - Important choices and rationale
- `findings:` - Key learnings
- `worked:` / `failed:` - What to repeat vs avoid
- `next:` - Action items for next session

**DO NOT use alternative field names like `session_goal`, `objective`, `focus`, `current`, etc.**
**The statusline parser looks for EXACTLY `goal:` and `now:` - nothing else works.**
---

### 3. Mark Session Outcome (REQUIRED)

**IMPORTANT:** Before responding to the user, you MUST ask about the session outcome.

Use the AskUserQuestion tool with these exact options:

```
Question: "How did this session go?"
Options:
  - SUCCEEDED: Task completed successfully
  - PARTIAL_PLUS: Mostly done, minor issues remain
  - PARTIAL_MINUS: Some progress, major issues remain
  - FAILED: Task abandoned or blocked
```

After the user responds, index and mark the outcome:
```bash
# Mark the most recent handoff (works with PostgreSQL or SQLite)
# Use git root to find project, then opc/scripts/core/
PROJECT_ROOT=$(git rev-parse --show-toplevel 2>/dev/null || echo "${CLAUDE_PROJECT_DIR:-.}")

# First, index the handoff into the database
cd "$PROJECT_ROOT/opc" && uv run python scripts/core/artifact_index.py --file thoughts/shared/handoffs/{session_name}/{filename}.yaml

# Then mark the outcome
cd "$PROJECT_ROOT/opc" && uv run python scripts/core/artifact_mark.py --latest --outcome <USER_CHOICE>
```

**IMPORTANT:** Replace `{session_name}` and `{filename}` with the actual values from step 1.

These commands auto-detect the database (PostgreSQL if configured, SQLite fallback).

**Note:** If indexing fails, the marking step will show "Database marking was not available" - this is acceptable for the first handoff but indicates the indexing step was skipped.

### 4. Confirm completion

After marking the outcome, respond to the user:

```
Handoff created! Outcome marked as [OUTCOME].

Resume in a new session with:
/resume_handoff path/to/handoff.yaml
```

---
##.  Additional Notes & Instructions
- **more information, not less**. This is a guideline that defines the minimum of what a handoff should be. Always feel free to include more information if necessary.
- **be thorough and precise**. include both top-level objectives, and lower-level details as necessary.
- **avoid excessive code snippets**. While a brief snippet to describe some key change is important, avoid large code blocks or diffs; do not include one unless it's necessary (e.g. pertains to an error you're debugging). Prefer using `/path/to/file.ext:line` references that an agent can follow later when it's ready, e.g. `packages/dashboard/src/app/dashboard/page.tsx:12-24`

© parcadei · MIT · parcadei/Continuous-Claude-v3

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set69

planning-agent A subagent that turns conversation context into an implementation plan, researches the codebase and writes a handoff before returning. CX Codex +1 518installs ast-grep-find Runs AST-based code search and refactoring using language syntax trees. Helps locate and replace function calls, imports, and variables without matching text inside strings or comments. CX Codex +1 518installs repoprompt Explores codebases through the RepoPrompt CLI without reading entire files. Useful for fetching code signatures and line slices with lower token use. CX Codex +1 515installs continuity-ledger The skill writes a YAML handoff with goals, decisions and next steps so a new session picks up right after /clear. CX Codex +1 509installs debug Investigates application logs, SQLite database state, and git history without editing source files. Useful for pinpointing issues during manual testing. CX Codex +1 507installs compound-learnings Transforms ephemeral session learnings into permanent rules, skills, and hooks. It turns repeated coding patterns into lasting project capabilities. CX Codex +1 505installs hook-developer A reference for Claude hooks: input and output schemas, registration in settings.json, testing patterns. Handy when you are writing a hook and forgot the exact format. CX Codex +1 504installs tldr-code A skill that gives the agent tldr CLI commands for file trees, call graphs, data flow and dead code. Cheaper on tokens than reading whole files. CX Codex +1 503installs

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor