agentleFS
Sign inStart free
NickolasBenakisNickolasBenakis/agent-playbook/skills/fanout/SKILL.md

fanout

Skill5 starsChanged 6 days agoMITView on GitHub

Human-in-the-loop workflow for high-impact, high-attention work. The main session stays with the human and delegates each discrete fix or feature to a subagent that returns a PR or verified diff. Use for core product work and dogfooding sessions.

5 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions5 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 5

    1. ansvisor/ansvisor★ 233same name
    2. dtiger1889-ops/claude-harness-toolbox★ 31same name
    3. linearuncle/linearuncleskills★ 8same name
    4. Breadstick-Crumbs/fanout★ 5same name
    5. weg-9000/Orange-pm★ 0same name

    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.

  • In this repository7 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 2

    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:

    Keep a running board in your replies, updated whenever it changes:

    line 23
  2. Very specific:

    When a subagent returns, verify its claim before marking it done: check the test output, and for non-trivial diffs run the …

    line 28
  3. Distinctive:

    Don't ask for confirmation on items that are clearly scoped.

    line 33

“Common” is measured: kept by at least 6 in 10 of the 4 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

Fanout

The human drives; you dispatch. The main session is a control room, not a workbench: keep its context for decisions and coordination, and push execution into subagents.

How to run it

  1. As the human dumps work ("this is broken", "add X", "make Y faster"), turn each into a self-contained brief with the delegate skill, plus the branch name fanout/<short-slug> and the finish line: tests/typecheck pass, PR opened (or diff ready if no remote).
  2. Launch one subagent per independent item, in an isolated worktree when the tool supports it. Run several at once only when AGENTS.md rule 6 says it's faster; otherwise one at a time, and a single small item can just be done. Items that touch the same files go to one subagent, sequentially.

  3. Keep a running board in your replies, updated whenever it changes:

    ItemAgentStatePR
  4. When a subagent returns, verify its claim before marking it done: check the test output, and for non-trivial diffs run the playbook-reviewer subagent with the item's acceptance criteria. If it failed, re-brief once with what went wrong; on a second failure, bring it to the human.
  5. Let the human keep going: "fix this with a subagent" should be all it takes. Don't ask for confirmation on items that are clearly scoped.

In bb

Child threads work as subagents and show up in the sidebar:

bb thread spawn --parent-self --new-environment worktree \
  --title "[fanout] <task> — building" --prompt-file brief.md
bb thread wait <id>            # or rely on parent notifications

Done

Report the board, with every row either pr-open, done, or needs-you: <why>.

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.

Check your setup, free