agentleFS
gigio1023gigio1023/agent-skills/skills/minimal/session-handoff/SKILL.md

session-handoff

Skill2 starsChanged 10 days agoMITView on GitHub

Use only when the user asks to write a handoff, continuation prompt, or session transfer file so another agent, model, or harness can continue the current work, or names session-handoff. Writes one short file under `.handoff/` that the successor executes: what was being done, what comes next, and the questions to put to the user. NOT when the user gives the path of an existing handoff file or says to continue from one: that asks you to carry out the file's task, not to write or improve a handoff. NOT for bounding a weaker executor on an approved plan (small-model-handoff), writing or executing plans (work-plan), or reading another session's transcript (read-agent-sessions). Never activate because a session has grown long or context is running low.

139 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use ascii-diagram, research-fleet and 7 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.
  • SethGammonSkill

    ascii-diagram

    Generate perfectly aligned ASCII diagrams — architecture, flow, sequence, box-and-arrow. Uses a programmatic character-grid approach so alignment is guaranteed by math, not token prediction. Includes post-render verification.

    Also in 4 projects that use this

  • SethGammonSkill

    research-fleet

    Parallel research using Fleet wave mechanics. Spawns multiple scout agents, each investigating a different angle of the same question. Findings are compressed between waves. Produces a unified research brief from multiple independent perspectives.

    Also in 3 projects that use this

  • vicoa-aiSkill

    live-preview

    Start a local app in the current user project, expose it through a tunnel, and return preview details in structured Markdown. Use when the user asks for a preview URL, public URL, shareable localhost, mobile/web preview, or tunnel setup.

    Also in 4 projects that use this

  • botlearn-aiSkill

    doc-gen

    Also in 4 projects that use this

  • chidekinaSkill

    test-gen

    Generate test cases from type constraints and specs. Creates boundary tests, edge cases, and fixtures automatically. Pact-lang inspired. Triggers on: generate tests, test from spec, create fixtures, boundary tests.

    Also in 4 projects that use this

  • team-attentionSkill

    tribunal

    This skill should be used when the user says "/tribunal", "tribunal", "review this", "3-way review", "risk-value-feasibility check", or wants multi-perspective adversarial review. Also triggered by: "트리뷰널", "리뷰 해줘", "3관점 리뷰", "위험성 검토". Runs 3 agents (Risk/Value/Feasibility) in parallel and synthesizes a verdict.

    Also in 3 projects that use this

  • +1 more from SethGammon

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 116 edited versions we read keep these.

  • 1

    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
  1. Distinctive:

    Write one file that a successor given only its path recognizes as its task and starts executing.

    line 19
  2. Very specific:

    If a plan in .plans/ or another project record governs the work, point to it by path, carry only what it lacks, and name work-plan when …

    line 30

“Common” is measured: kept by at least 6 in 10 of the 116 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

Session Handoff

Write one file that a successor given only its path recognizes as its task and starts executing. If you were handed a handoff file instead, you are the successor: do its task and leave the file alone.

This request authorizes reading what the handoff needs, writing the file, and the local Git exclude entry in step 1, nothing more. Grants the user gave for the task are recorded for the successor, not used here.

Writing the File

  1. Write to .handoff/<YYYY-MM-DD>-<HHMM>-<task-slug>.md at the project root, in local time with a short ASCII kebab-case slug. For several repositories, use their common root and label each. The file carries quoted user words, session ids, absolute paths, and grants, so keep it out of version control: before writing, run git check-ignore -q .handoff/ from that root. Exit 0 (already ignored) or 128 (no repository) needs nothing more. On exit 1, append .handoff/ to the local exclude file at git rev-parse --git-path info/exclude, which is never committed, and mention it in the delivery; if that file cannot be written, tell the user the handoff will be visible to Git before writing it.
  2. Fill assets/handoff.template.md. Keep its bold opening instruction word for word and delete rows with nothing to say.
  3. The user's current explicit words decide intent and scope; live files, version control, and fresh tool output decide status. Quote the user's decisive sentences verbatim with their date. Point to paths, commands, and commits instead of pasting them, and leave out secrets.
  4. Subagents, background jobs, pending tool calls, and in-session grants end with this session. Record what each left on disk, and list as still running only handles the successor can check from outside, with the checking command. Record the origin harness and session id so read-agent-sessions, where installed, can recover the transcript.
  5. Ask only about unknowns that would change the work and that only the user can settle, each with its consequence, concrete options with a recommended default, and what to do if unanswered. A fact that inspection or a cheap check can settle goes into Next Actions instead.
  6. If a plan in .plans/ or another project record governs the work, point to it by path, carry only what it lacks, and name work-plan when continuing means executing it.

The file is ready when a reader with nothing else can run the first action, sees evidence for each completed claim, and knows what to ask the user.

Earlier Handoffs

The newest .handoff/ file with the task's slug is current. Keep its still-valid decisions, name it on the new file's Supersedes line, and add only this line at the top of the old file: Superseded by <new path>. Open that file instead. A legacy root handoff.md is superseded the same way.

Delivery

Report the path, the recorded status, the first action, and any exclude entry added. Read references/source-notes.md only when maintaining this skill.

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.