agentleFS
chrisalehmanchrisalehman/bionic/skills/handoff/SKILL.md

handoff

Skill9 starsChanged 30 days agoMITView on GitHub

Use when the work has to continue in another session: /bionic:handoff, "handoff", "hand off", "hand this off", "I''m stepping away", "write the handoff", "before I /clear", or a copy of the work going to a fork while this session goes on. Works with or without a bionic run. Composes the sections only this session knows, runs the handoff verb with them, and ends on the file''s last line, the way the next session resumes.

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

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

  • 0

    Common

    Most of the 172 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run the verb with the sections on its stdin, each under its heading, the purpose quoted when there is one:

    line 31
  2. Distinctive:

    Never write the handoff file by hand, and never edit it: State is computed, and the rest passes through the verb's redaction.

    line 74

“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

The handoff is one document a fresh session continues from without asking what was going on. Assume this session can be killed the moment the file is written. One verb writes the file: it computes the State (branch, head, uncommitted files, what runs, the agents, and the plan, step, open tasks and Patrol when a run is bound), redacts secrets, and adds the last line. This skill composes the rest and writes nothing else.

Typed as /bionic:handoff with no argument, it hands off whatever this session is doing. With an argument, the argument names what the next session is for (/bionic:handoff review the parser change): the document is tailored to that purpose, and the verb puts it in the title and, with no run bound, in the file name.

<plugin-root> below is the plugin's root, two directories above this skill's own.

The steps

  1. Compose the sections from this session, tailored to the purpose when one was given. Plain and specific, one line per item:

    • Summary, exactly four lines: Doing: what is in motion and where it stands; Learned: what the next session would otherwise relearn; Open: the question that decides what comes next; First: the first thing to do.
    • Goal: the person's own words, quoted verbatim, with who said them and when. Never a summary of them.
    • Findings: one fact per line, each ending on its source: a path with its line, a log, a commit or a URL. A finding you cannot source ends on unverified; the verb marks any line with no source that way.
    • Decisions: each ruling with who made it, when, and their words in quotes. A call you made yourself says so: (the agent's call).
    • Ruled out: each approach dropped, with the evidence that dropped it.
    • Open: each unknown, labelled hypothesis: or question:.
    • Suggested skills: the skills the next session should reach for, and when.
    • Next: the commands to run, in order, the first one verbatim in backticks.

    Name a settled artifact (a spec, a plan, a log, a design) by its path; never paste it. A pasted block is left out by the verb. Leave a section out when it has nothing: the verb prints - none. under it. Never write State or Resume: they are the verb's.

  2. Run the verb with the sections on its stdin, each under its heading, the purpose quoted when there is one:

    bash <plugin-root>/hooks/session-poker.sh handoff ['<purpose>'] <<'HANDOFF'
    Summary
    - Doing: <text>
    - Learned: <text>
    - Open: <text>
    - First: <text>
    
    Goal
    - "<the person's words>" (<who>, <when>)
    
    Findings
    - <fact> — <source>
    
    Decisions
    - <who> ruled "<their words>" (<when>)
    
    Ruled out
    - <approach> — <evidence>
    
    Open
    - hypothesis: <text>
    
    Suggested skills
    - <skill>, <when>
    
    Next
    - `<the first command>`
    HANDOFF

    A refusal is the answer: say it in one line and stop.

  3. On success the verb prints one line, the file's last line, and nothing else. It is the one thing the next session needs, and it is your whole closing message: the line in a fenced block, with nothing after it. With a run bound the next session types it to rebind the run, arm its Patrol at a scale that has one and adopt the running agents. Typing it is the consent: the resume engages the session itself when the slash command did not, and says so in its first line. The resume prints about a dozen lines (bound, the step, the agents still open, what landed since, what is owed now, the Patrol) and the path of one recap file holding the rest. With no run the line is Read <path> whole, then do Next 1. and goes in the same fence. A second line, saying the plan's pointer was not written, is a warning about the plan, not the file: say it in one sentence above the fence. The last thing you write, with a run bound, is:

    /bionic:canonical-sdlc <path>

The rules

  • Never write the handoff file by hand, and never edit it: State is computed, and the rest passes through the verb's redaction.
  • The document lives under .bionic/docs/record/handoffs/, which git never tracks; the verb creates the folder and its ignore file when the project has none.
  • Nothing here changes a plan, a record or a run beyond the file and, with a run bound, the plan's ## Handoff pointer the verb writes.

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.