agentleFS
rosslevinskyrosslevinsky/portable-agent-skills/skills/handoff/SKILL.md

handoff

Skill12 starsChanged 43 days agoApache-2.0View on GitHub

Write a restart prompt for the current work to a file, so a fresh session can pick the work up after the context is cleared. Records the goal, what is done and what is next, the decisions already made, the traps found and the uncommitted work, then prints one line to paste after clearing. Use when the user invokes /handoff, or says 'give me a restart prompt so I can clear context', 'write a handoff', 'I want to clear context and continue'.

191 other projects use a version of this.

Adopted by

Most recently taken up by

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.
  • FabienDehopreSkill

    decision-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

  • liufeng1976Skill

    setup-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

  • AWSSkill

    to-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

  • GoogleSkill

    ask-matt

    Ask which skill or flow fits your situation. A router over the skills in this repo.

    Also in 87 projects that use this

  • RoentekSkill

    loop-me

    Grill me about specs for the workflows I want to build, within this workspace.

    Also in 51 projects that use this

  • LesaniSkill

    writing-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

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 1 is distinctive, and 3 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.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Write what a fresh session needs to continue this work into a file, and end with the one line the user pastes after /clear.

    line 8
  2. Distinctive:

    Do not stage, commit or push anything.

    line 54
  3. Very specific:

    When the file was written, end with one short sentence telling the user to paste the next line after /clear, then exactly this line, with …

    line 58
  4. Very specific:

    If the file cannot be written, say why in one sentence and print the whole prompt in one fenced block instead.

    line 65

“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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Handoff

Write what a fresh session needs to continue this work into a file, and end with the one line the user pastes after /clear.

Classification: Degraded — where the runtime cannot write outside the workspace (a sandbox that forbids it), the prompt is printed in one fenced block instead of saved, and the user pastes that block.

Step 1 — Choose the file

  • <project> is the folder name of the repository root (git rev-parse --show-toplevel), or of the working directory outside a repository.
  • The file is ~/.handoff/<project>.md, where ~ is the user's home folder (%USERPROFILE% on Windows). Resolve it to an absolute path, and create the folder if it is missing.
  • Before every write, read the first line of the candidate file if it exists. Overwrite it only when that line is exactly Project: <this project's absolute path>. Any other first line — another path, an empty line, anything unrecognized — means the file is not this project's: never overwrite it. Prefix the name with the next parent folder (<parent>-<project>.md, then <grandparent>-<parent>-<project>.md) and check again, until the name is free or this project's own. Stop at the filesystem root or a drive (a name never contains / or C:): if the parent folders run out first, write nothing and print the prompt as Step 3's fallback does.

Step 2 — Write the prompt

The first line is exactly Project: <absolute path of the project>. Then write, for a reader who has none of this conversation:

  • Written — the date and time, the branch, and the last commit (git log -1 --oneline).
  • Goal — what the work is for.
  • Done — what is finished, and how it was verified.
  • Next — the next concrete steps, in order.
  • Decisions already made, and why — so the next session does not reopen them.
  • Traps found — what failed or misled, and what to do instead.
  • Files to read first — paths, including the project's CLAUDE.md / AGENTS.md where either exists. Describe a file by its path and what it is for; never restate its contents, which the next session reads for itself and a paraphrase gets wrong.
  • Uncommitted work — git status --short, and which of those changes are this work's.
  • Open questions for the user.
  • A closing instruction: before acting on any of it, check the repository's current state (git status, git log -1), because the file may be older than the tree.

Be concrete: name files, commands and commit hashes. The next session can read the code; it cannot read this conversation. Never write a secret, key, token, password or credential into the file, which sits in the home folder as plain text: name where a credential is kept, never its value.

Do not stage, commit or push anything. The file lives outside the repository.

Step 3 — Report

When the file was written, end with one short sentence telling the user to paste the next line after /clear, then exactly this line, with the file's absolute path in it:

Read <absolute path> and continue from it.

If the file cannot be written, say why in one sentence and print the whole prompt in one fenced block instead. The user pastes that block after /clear.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.