handoff
Skill0 starsChanged 2 months agoMITView on GitHub
Capture a lightweight session handoff in the project's Memories directory. Use when the user explicitly asks to preserve current context for a fresh agent; record recalled signal without analysis and identify the exact next action.
191 other projects use a version of this, 77 of them edited.
Adopted by
Most recently taken up by
- xiaoxiaosrm/dsh-mattpocock-skillsedited, 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 · 77
- mattpocock/skills★ 309k16% same
- QuivrHQ/quivr★ 40k16% same
- The-Vibe-Company/quivr★ 40k16% same
- GreyDGL/PentestGPT★ 16k17% same
- ArcReel/ArcReel★ 6.3k16% same
- withcoral/coral★ 4.9k16% same
- bestofjs/bestofjs★ 3.1k16% same
- compozy/compozy★ 2.9k16% same
- bholmesdev/hubble.md★ 1.5k17% same
- CelestoAI/SmolVM★ 1.2k16% same
67 more not listed
Wrote their own version · 114
- ykdojo/claude-code-tips★ 10ksame name
- vinvcn/mattpocock-skills-zh-CN★ 5.5ksame name
- epicenter-md/epicenter★ 4.8ksame name
- EpicenterHQ/epicenter★ 4.8ksame name
- davidondrej/skills★ 4.3ksame name
- feiskyer/claude-code-settings★ 1.7ksame name
- skforecast/skforecast★ 1.5ksame name
- openclaw/agent-skills★ 1.1ksame name
- suboss87/FDEOps★ 940same name
- JordanCoin/codemap★ 710same name
104 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.
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. 0 are common, 7 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 172 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
If arguments were passed with /handoff (e.g., /handoff prototype the window communication layer), treat them as a description of what the …
- Distinctive: line 30
Ask yourself one question: can I clearly recall the beginning of this session — the original task, the first decisions, the early user …
- Very specific: line 37
Add an # Early-session gaps section listing what you know is missing (task origin?
- Distinctive: line 38
Do NOT fabricate early details to fill the gaps.
- Very specific: line 40
Why this matters: If this is your sole handoff and context_depth: deep, the recall agent will know to ask the user to fill the gaps.
- Very specific: line 46
Write the file to <project-root>/Memories/handoff-<YYYY-MM-DD-HHMM>.md.
- Very specific: line 48
If Memories/ doesn't exist, create it.
- Distinctive: line 105
Every section is optional except Task and Next action.
- Distinctive: line 107
Do not analyze, evaluate, or tag your claims.
- Distinctive: line 108
Do not redact or summarize further than one line per item.
- Very specific: line 109
No code blocks unless essential.
- Distinctive: line 110
No confidence tags, no provenance tags, no quality gate.
- Distinctive: line 111
Do not duplicate content already captured in other artifacts (PRDs, ADRs, issues, existing handoffs).
“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
You are about to be replaced. A fresh agent will arrive cold and read everything in Memories/ to reconstruct the session.
The only thing you need to do: write down what you can recall right now into the structure below. Do not filter. Do not judge relevance. Do not assess your own confidence or context health. The recall agent handles all of that — your job is just to dump signal into the file.
If arguments were passed with /handoff (e.g., /handoff prototype the window communication layer), treat them as a description of what the next session will focus on. Tailor the handoff — include what's relevant to that task, skip what isn't.
Step 1 — Gather metadata (run commands)
git rev-parse --show-toplevel
git branch --show-current
git rev-parse HEAD
git status --short
ls Context/Plans/*.md 2>/dev/nullStep 2 — Assess your recall boundary
Ask yourself one question: can I clearly recall the beginning of this session — the original task, the first decisions, the early user constraints?
- Yes, I can recall early turns with specificity → set
context_depth: shallowin the frontmatter. Your memory is intact. - No, early turns are fuzzy or lost → set
context_depth: deep. This tells the recall agent you're capturing from a degraded window — it will flag potential gaps to the user.
If context_depth: deep:
- Focus on what you can recall — recent events, current state, next action — without guessing at the fuzzy parts.
- Add an
# Early-session gapssection listing what you know is missing (task origin? early decisions? user constraints?). Be explicit: "I cannot recall whether the user approved approach X or Y in the early session." - Do NOT fabricate early details to fill the gaps. It is better to leave a gap the recall agent can surface than to record a wrong claim as fact.
Why this matters: If this is your sole handoff and context_depth: deep, the recall agent will know to ask the user to fill the gaps. If earlier handoffs exist, those gaps are already covered. The system degrades gracefully either way.
Step 3 — Write the handoff
Write the file to <project-root>/Memories/handoff-<YYYY-MM-DD-HHMM>.md.
If Memories/ doesn't exist, create it. If the filename collides, append -2.
Use this structure:
---
date: <ISO timestamp>
branch: <branch name>
commit: <commit sha>
context_depth: <shallow | deep>
plan: <Context/Plans/<Feature-Name>.md this session worked against — omit if none>
---
# Task
<one line — what are we working on?>
# Constraints & preferences
- <every constraint the user stated — verbatim if possible>
- <north star — what they optimize for: speed, correctness, minimality…>
- <vetoes this session + reason>
# Early-session gaps
<Only if context_depth: deep. List what you know is missing.>
# Recent events
<what happened recently, in chronological order. One line each. Include:
- Decisions made + rationale
- Discoveries + what was learned
- Surprises or contradictions
- Errors encountered + fix
- User approvals, vetoes, pivots>
# Decisions not yet in the plan
- <a decision made this session that the plan file does not reflect yet — one line, naming the work item it affects>
# Open questions
- <pending decisions awaiting the user>
- <things we know we don't know>
# Files changed
- `path/to/file.ts` — <what changed and why, in one line>
# Dead ends & false leads
- Tried <X>, failed because <Y> — don't repeat
# Next action
<concrete next step>
# Suggested skills
- <skills the next agent should invoke (e.g., "grill", "prototype", "prd")>
- <reason: why this skill fits the next action>
# Pointers
- <other artifacts the next agent should open: ADRs, issues, research notes>Rules:
- Every section is optional except Task and Next action. If you can't recall anything for a section, omit it — never write "N/A."
Decisions not yet in the planis recall, not analysis: list what was decided and not yet written into the plan, nothing more. If the plan already reflects every decision, omit the section./recallsurfaces it;/promotewrites it.- Do not analyze, evaluate, or tag your claims. Just write them down.
- Do not redact or summarize further than one line per item. Err on the side of including.
- No code blocks unless essential. Prefer
path:linereferences. - No confidence tags, no provenance tags, no quality gate. The recall agent handles those.
- Do not duplicate content already captured in other artifacts (PRDs, ADRs, issues, existing handoffs). Reference them by path or URL instead.
- Redact any sensitive information: API keys, passwords, or PII.
Resume with /recall.
Workflow
/handoff can be invoked at any point in any session. Use it to:
- Checkpoint progress mid-grill or mid-blueprint
- Pass a high-fidelity question to
/prototype - Pass prototype learnings back to a planning session
- Preserve context before starting implementation in a fresh window
The next agent resumes with /recall.
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.

