handoff
Skill39 starsChanged 37 days agoMITView on GitHub
When the human types `/handoff` (or "next session", "what''s next", "handoff note"), produce a copy-paste markdown block for the next session.
192 other projects use a version of this.
Adopted by
Most recently taken up by
- xiaoxiaosrm/dsh-mattpocock-skillstheir own version, first seen Nov 8, 2026
- ethosagent/ethostheir own version, first seen Nov 8, 2026
- dsifry/metaswarmtheir own version, first seen Nov 8, 2026
- glebis/claude-skillstheir own version, first seen Nov 8, 2026
- domengabrovsek/agent-configtheir own version, first seen Nov 8, 2026
- domengabrovsek/claudetheir own version, first seen Nov 8, 2026
Used with
Projects with it also use decision-mapping, setup-ts-deep-modules and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
FabienDehopreSkilldecision-mapping
Turn a loose idea into a sequenced map of investigation tickets, then drive them to resolution one at a time.
Also in 8 projects that use this
liufeng1976Skillsetup-ts-deep-modules
Wire dependency-cruiser into a TypeScript repo so each package is a deep module — implementation hidden in subfolders, reachable only through its entry-point files. User-invoked.
Also in 42 projects that use this
AWSSkillto-questionnaire
Turn a decision you can't fully answer into a questionnaire for someone else to fill in.
Also in 93 projects that use this
GoogleSkillask-matt
Ask which skill or flow fits your situation. A router over the skills in this repo.
Also in 87 projects that use this
RoentekSkillloop-me
Grill me about specs for the workflows I want to build, within this workspace.
Also in 51 projects that use this
LesaniSkillwriting-beats
Shape an article as a journey of beats, choose-your-own-adventure style. The user picks a starting beat from the raw material, you write only that beat, then offer options for where to pivot next, beat by beat, until the article reaches a natural end. Use when the user has raw material and wants to assemble it as a narrative rather than an argument.
Also in 54 projects that use this
Other versions192 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 192
- mattpocock/skills★ 309ksame name
- QuivrHQ/quivr★ 40ksame name
- The-Vibe-Company/quivr★ 40ksame name
- GreyDGL/PentestGPT★ 16ksame name
- ykdojo/claude-code-tips★ 10ksame name
- ArcReel/ArcReel★ 6.3ksame name
- vinvcn/mattpocock-skills-zh-CN★ 5.5ksame name
- withcoral/coral★ 4.9ksame name
- epicenter-md/epicenter★ 4.8ksame name
- EpicenterHQ/epicenter★ 4.8ksame name
182 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository41 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- build-itSkill
- challenge-implementation-1Skill
- challenge-implementation-2Skill
- challenge-implementation-3Skill
- challenge-implementation-4Skill
- challenge-implementation-5Skill
- challenge-plan-1Skill
- challenge-plan-2Skill
- challenge-plan-3Skill
- challenge-plan-4Skill
- challenge-plan-5Skill
- construct-the-planSkill
- frame-itSkill
- handoffSkill
- quizSkill
- security-scanSkill
- soloSkill
- start-upSkill
- test-itSkill
- wrap-upSkill
- build-itSkill
- challenge-implementation-1Skill
- challenge-implementation-2Skill
- challenge-implementation-3Skill
- challenge-implementation-4Skill
- challenge-implementation-5Skill
- challenge-plan-1Skill
- challenge-plan-2Skill
- challenge-plan-3Skill
- challenge-plan-4Skill
- challenge-plan-5Skill
- construct-the-planSkill
- frame-itSkill
- quizSkill
- security-scanSkill
- soloSkill
- start-upSkill
- test-itSkill
- wrap-upSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 173 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 18
If the pull request is ready and the work is meant to ship, merge it using the repository's merge policy.
- Distinctive: line 19
If the pull request is superseded, abandoned, or should not ship, close it without merging.
- Distinctive: line 75
When in doubt, flag as needs-the-human.
- Very specific: line 81
If nothing's left over and nothing qualifies, output exactly: Nothing to hand off this time.
- Distinctive: line 89
After resolving in-scope pull requests, handoff modifies no product or documentation files.
“Common” is measured: kept by at least 6 in 10 of the 173 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
When the human types /handoff (or "next session", "what's next", "handoff note"), produce a copy-paste markdown block for the next session. Top half is human-scannable; bottom half is pure agent instructions for next session's start-up. This is the closing half of the pair that makes sessions continuous: handoff writes the note, start-up reads it.
Precondition: the background-work barrier
Handoff may not emit while any background task spawned this conversation is unresolved: await finite work, act on landed-but-unhandled results, retire session daemons, or (narrowly, never for a verification task) abandon with the reason written down. Right after wrap-up this is a quick re-check that nothing new was spawned; standalone, the full barrier applies first. See AGENTS.md § The background-work barrier.
Pull request authority
Invoking handoff is explicit permission to resolve pull requests created by the work being handed off. Do this before writing the handoff note:
- Inspect the exact head commit, base branch, required reviews, and checks.
- If the pull request is ready and the work is meant to ship, merge it using the repository's merge policy. Never bypass a failed check or required review.
- If the pull request is superseded, abandoned, or should not ship, close it without merging.
- Prove the final GitHub state, clean up the resolved branch, synchronize the base branch, and verify any resulting base-branch check.
This authority is narrow. It covers pull requests created by this session or explicitly named in the work being handed off. It does not cover unrelated pull requests, product or provider changes, spending, deployment, or any other external effect. If an in-scope pull request cannot be resolved safely, leave it open and state the exact blocker in the handoff.
Voice
The output IS chat to the human: follow AGENTS.md § Voice, including in every fill-in. Pipe the draft through node scripts/voice-gate.js before sending.
Locked template
Output is ONE fenced markdown code block. Inside it:
## Handoff
{One paragraph for the human: what the session that just ended focused on, what's left over, and the recommended focus for next time. About 3-5 sentences of plain English. No file paths in the prose, no unexplained acronyms, no marker tags.}
- {Key context bullet 1}
- {Key context bullet 2}
- {Up to 4-5 bullets max}
---
**For the agent:**
Start with `/start-up`. Draw up a plan focused on **{recommended focus, one phrase}** using construct-the-plan. If open questions remain after the draft, run `/frame-it` and get the human's answers. Then STOP and wait: the human queues the challenge rounds, `/build-it`, `/test-it`, the implementation challenges, and `/wrap-up`; never start them unprompted.
**Specific items to pick up (agent-doable without the human):**
- {Item 1 with file path or scope}
- {Item 2}
**Needs the human's input, do NOT pick up unsupervised:**
- {Item, or "None"}
**Out of scope:**
- {Item, or "None"}The horizontal rule separates the human-scannable top from the agent instructions. The whole block is wrapped in a markdown code fence so it pastes cleanly.
Sources to pull from
- This session's chat: deferred decisions and in-flight items.
- The handoff note from last session, if one is still relevant.
docs/plans/filtered to plans marked IN_PROGRESS.- Whatever standing to-do or issues surface your project keeps.
Missing or unreadable sources: skip, continue, and add a one-line Skipped sources: [list] footnote.
Sorting items into sections
- Deferred markers from the conversation go to needs-the-human.
- Plans marked PROPOSED go to needs-the-human (they await review); plans marked IN_PROGRESS are agent-doable.
- Lint warnings, drift findings, doc updates: agent-doable.
- Design questions, product tuning, anything taste-shaped: needs-the-human.
- When in doubt, flag as needs-the-human.
Cap the agent-doable list at the top 5 by relevance and say where the rest live.
Empty case
If nothing's left over and nothing qualifies, output exactly: Nothing to hand off this time. Self-contained session.
Has an opinion
The recommended focus picks ONE thing and justifies it in a sentence. Don't offer a menu.
Repository boundary
After resolving in-scope pull requests, handoff modifies no product or documentation files. It produces the block and stops. The Git and GitHub state changes described above are the only writes this command authorizes.
Rationale (recorded so future edits don't drift it)
The two-half shape exists because the note serves two readers with opposite needs: the human skims the top in ten seconds to remember where things stand, and the next session's agent executes the bottom verbatim. Mixing the two produces a note neither reader trusts. The pull request step belongs before the note because handing off a ready or obsolete pull request just makes the next session repeat the end of this one. The explicit STOP instruction in the agent half is load-bearing: without it, next session's agent tends to run the whole chain uninvited, and the human loses the checkpoints the process exists for.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

