agentleFS
augments-labsaugments-labs/sdlc-skills/skills/common/handoff/SKILL.md

handoff

Skill1 starsChanged 3 months agoMITView on GitHub

Writes a handoff so a fresh session or another agent can continue unfinished work. Use when a session is ending mid-work, the context is getting full, the user is heading out, or someone else is taking over. Skip resuming from a handoff an earlier session wrote: reading it needs no skill. Skip finished tasks and bounded worker assignments whose coordinator stays responsible.

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

9 rules. 0 are common, 7 are distinctive, and 2 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.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write the state the next session resumes from, not a summary of the conversation, and write it only where the user or a standing …

    line 8
  2. Distinctive:

    Skip: when this session is the one resuming: a handoff is read, not invoked.

    line 16
  3. Distinctive:

    Skip: when the task is finished and self-contained.

    line 18
  4. Distinctive:

    If an instruction or a standing authority names the destination, write there.

    line 24
  5. Very specific:

    Read references/handoff-example.md when the first draft is written and needs comparing against a filled example.

    line 46
  6. Distinctive:

    Never write an assumption as approval.

    line 58
  7. Distinctive:

    Leave no weak result out.

    line 61
  8. Distinctive:

    Before writing, remove every key, token, password, and piece of personal data.

    line 68
  9. Very specific:

    After writing a note to a file, record the first 7 characters of git hash-object of that file in its record (the ledger: add to it, never …

    line 70

“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 the state the next session resumes from, not a summary of the conversation, and write it only where the user or a standing instruction says it may go.

When to use

  • A conversation is ending mid-work, or responsibility for continuing it passes to a fresh session or a different agent.
  • Skip when this session is the one resuming: a handoff is read, not invoked.
  • Skip when the task is finished and self-contained.
  • Skip bounded worker assignments while the coordinator retains the task; those use the dispatch packet and return contract instead.

Step 1: Settle the destination

  1. If an instruction or a standing authority names the destination, write there. Do not ask.
  2. Otherwise ask one question and end the turn:

    Where should the handoff go?
    
    1. .sdlc-skills/handoffs/ in this project — readable by anyone who can read this checkout; committed only when the reader works from another checkout and the content is safe to share, and then readable by everyone with repository access
    2. A path you name
    3. This reply only — nothing written
    
    Recommendation: {{least-disclosing option that reaches the recipient}} — {{one sentence}}.
  3. "Wherever is easiest", approval of the content, or silence names no destination. Ask again.
  4. For a scratch path outside the workspace, fill Storage controls: data class, readers, lifetime, exact cleanup target and owner, and whether cleanup is still pending. Write nothing into the repository.

Step 2: Fill assets/handoff-template.md before writing

Read references/handoff-example.md when the first draft is written and needs comparing against a filled example.

  1. Fill every section: identity, goal, state identity, decisions and authority, evidence, gotchas and permissions, resume first action, suggested skills, references.
  2. In Handoff identity, name the predecessor record and add this note after it as its successor. If more than one successor already ends the predecessor's lineage, or the predecessor file's ledger holds no hash for it, or one the file no longer matches, stop and resolve the lineage.
  3. In Decisions and authority, give for each decision the direct answer or the standing default that authorized it. List every open decision as open. Never write an assumption as approval.
  4. In Evidence, give for each check: the command, where it ran, the tree, when, and the result. If a result is stale or the check never ran, say so. Leave no weak result out.
  5. Suggested skills lists possibilities only. It is not proof that any was invoked, and not a sequence to follow.
  6. References holds paths or URLs to specs, plans, ADRs, issues, commits. Copy nothing in.
  7. Resume first action is one concrete step, which the next session verifies before executing it.
  8. Before writing, remove every key, token, password, and piece of personal data.
  9. After writing a note to a file, record the first 7 characters of git hash-object of that file in its record (the ledger: add to it, never rewrite it; it sits outside the note, at the note's path with .ledger.md in place of .md), never in the note.

Gotchas

  • Session state can hold credentials, customer data, or unreviewed findings. A store the session invents discloses it to readers nobody chose.
  • A hash written inside the note changes the bytes it names, so the note never matches it and every later handoff stops on its lineage.

Common mistakes

  • Treating the handoff as a durable project-lessons store; it transfers current state once.

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.