handoff
Skill1 starsChanged 4 months agoMITView on GitHub
Write a context-saving handoff doc for long-running work before a /clear, so the next chat resumes from the destination (current state + what's left) instead of re-deriving the whole journey from the transcript. Use when the user says "write a handoff", "handoff doc", "save context", "we're going to /clear", or when a session has accumulated a lot of state and is wrapping up.
191 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 versions191 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 191
- 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
181 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 repository9 other files it gives its agents.
- commitpushSkill
- discordembedSkill
- discordmsgSkill
- endSkill
- prettyhtmlSkill
- prettymsgSkill
- startSkill
- wrapupSkill
- wtfSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.
0
Common
Most of the 172 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 73
Read $ARGUMENTS (a topic label) and the conversation → identify the project/topic, its current state, and what's left.
- Distinctive: line 84
Doc-only: Never bumps versions, commits, deploys, or changes code — it only writes the handoff.
- Very specific: line 91
No skill or tool can make /clear itself seed a message — that's a hard Claude Code platform limit (/clear is user-typed only, never …
“Common” is measured: kept by at least 6 in 10 of the 172 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Write a Handoff Doc (context-saving for /clear)
Long work often spans many sessions on one thing. The chat fills with the journey (debugging, dead
ends, retries) which is almost all token-waste once the work lands. This skill captures the
destination — current state + what's left — into a short doc, so the user can /clear and the
next chat resumes from the doc, not the transcript.
Where handoff docs go
- Folder: a
handoffs/folder — in the current git repo root if we're in one, else the home directory. Override with theHANDOFF_DIRenv var.BASE="$(git rev-parse --show-toplevel 2>/dev/null || echo "$HOME")" DIR="${HANDOFF_DIR:-$BASE/handoffs}" mkdir -p "$DIR" - Untracked by default. These are transient context, not shipped docs. If you're in a git repo,
add
handoffs/to.gitignoreso they don't get committed. The destination doc a handoff points to (a README / plan / spec) stays tracked wherever it already lives. - Naming:
<Project>-<topic>-Handoff-<YYYY-MM-DD>.md(e.g.Auth-Migration-Session-Cookies-Handoff-2026-07-03.md). One per session/topic; don't overwrite an older one — use a new date. Use today's date from the Claude Code environment'scurrentDate.
The one rule: destination, not journey
The reader is a fresh agent with zero context. Give it where the work IS and what's LEFT, anchored to the destination doc — the project's README, plan, or spec (put a pointer to it at the very top if one exists). Cut the blow-by-blow: no "first I tried X, then Y." Keep failed approaches only as one-line "don't re-try this because…" warnings.
Template
# <Project> — <topic> Handoff (<YYYY-MM-DD>)
> **DESTINATION DOC (read FIRST):** `<path to the README / plan / spec, if any>`. This handoff is the
> bridge from the work NOW to that destination — state + remaining work, not a journey log.
## Subject
- What this is: the feature/bug/task, the repo + branch, key files, current state.
- **Status now:** committed? pushed? deployed? (say plainly — often UNCOMMITTED).
## DONE & working (build on these)
- Bullet the changes that are in and working, with the file + the WHY (one line each).
## BROKEN / UNVALIDATED / MISSING ← the heart of the doc
- What's not proven yet and the exact next step to prove it (the command to run, the thing to check).
- Known baselines/numbers/error messages worth keeping.
## REVERTED / deferred (do NOT blindly re-add)
- What was backed out and the reason, so the next agent doesn't repeat the mistake.
## Files touched (uncommitted)
- Path list so the next agent / a future commit knows the blast radius.
## Build / run / test workflow specifics
- The exact commands for THIS work: how to build, run, test, and see the result. Branch, env vars,
any setup the next agent needs so they don't rediscover it.
## Next-session task list (scoped)
1. … 2. … 3. … (ordered and concrete; mark anything that's someone else's domain to unblock)Steps
- Read
$ARGUMENTS(a topic label) and the conversation → identify the project/topic, its current state, and what's left. - Resolve
$DIR(above) and pick the dated filename. - Fill the template faithfully — destination not journey; be honest about what's unproven. Don't inflate "unvalidated" into "done".
- Write the doc to
$DIR/<name>.mdwith the Write tool. - Report the path and tell the user it's safe to
/clear— the next chat should read this doc (and the destination doc it points to) first.
Rules
- Doc-only. Never bumps versions, commits, deploys, or changes code — it only writes the handoff.
- Destination, not journey — the single most important rule. A handoff that re-tells the debugging story has failed; one that says "here's where it is, here's the next probe" has succeeded.
- Faithful — represent uncertainty and open bugs honestly; the next agent trusts this doc.
Companion: auto-load-on-/clear hook (optional)
No skill or tool can make /clear itself seed a message — that's a hard Claude Code platform limit
(/clear is user-typed only, never programmatic, and always starts empty). The closest substitute is
a SessionStart hook matching clear that fires right after /clear and injects a pointer to the
newest relevant handoff doc as context, so the fresh session reads it before doing anything else,
without waiting for the user to run /start.
Sketch — a script that prints the newest handoff for the current branch, wired in settings.json
under hooks.SessionStart with matcher: "clear":
#!/usr/bin/env bash
# ~/.claude/hooks/handoff-on-clear.sh — no-ops silently outside a git repo.
BRANCH="$(git rev-parse --abbrev-ref HEAD 2>/dev/null)" || exit 0
BASE="$(git rev-parse --show-toplevel 2>/dev/null || echo "$HOME")"
DIR="${HANDOFF_DIR:-$BASE/handoffs}" # same default as the skill above
[ -d "$DIR" ] || exit 0
# best-effort: prefer a handoff whose name matches the branch's topic, else the newest overall
DOC="$(ls -t "$DIR"/*"${BRANCH##*/}"*.md 2>/dev/null | head -1)"
[ -n "$DOC" ] || DOC="$(ls -t "$DIR"/*.md 2>/dev/null | head -1)"
[ -n "$DOC" ] && echo "Resuming? Read the handoff first: $DOC (confirm it matches the user's intent)."It's a nudge, not authority — the injected text should tell the resuming chat to confirm the
picked handoff actually matches what the user wants before proceeding. /start remains the
authoritative orientation step (clean-tree check, contamination flag, sync with origin); the hook just
removes the "remember to type /start" step for the common case.
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.

