fan-out
Skill0 starsChanged 46 days agoMITView on GitHub
Split work that has many independent units across parallel subagents, verify each one's evidence, and consolidate into one result. Use for audits across many services, files, or capabilities; migrations that touch many places the same way; several unrelated failures; or when the user asks to parallelise.
8 other projects use a version of this.
Adopted by
Most recently taken up by
- Suge8/friedbuntheir own version, first seen Nov 8, 2026
- keli-wen/rondotheir own version, first seen Oct 25, 2026
- keli-wen/dev-skillstheir own version, first seen Oct 13, 2026
- me2resh/apexyardtheir own version, first seen Oct 8, 2026
- alexanderpino/skillstheir own version, first seen Sep 15, 2026
- darrenCWJ/agent-skills-scaffoldtheir own version, first seen Aug 9, 2026
Other versions8 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 8
- me2resh/apexyard★ 542same name
- keli-wen/dev-skills★ 12same name
- keli-wen/rondo★ 12same name
- Suge8/friedbun★ 12same name
- alexanderpino/skills★ 1same name
- darrenCWJ/agent-skills-scaffold★ 0same name
- haliphax-ai/skills★ 0same name
- reaatech/mcp-gateway★ 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 repository14 other files it gives its agents.
- architectureSkill
- debugSkill
- finishSkill
- merge-conflictsSkill
- reviewSkill
- spec-adoptSkill
- spec-amendSkill
- spec-boundariesSkill
- spec-implementSkill
- spec-initSkill
- spec-reviewSkill
- spec-runSkill
- spec-shapeSkill
- tddSkill
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
1
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
- Distinctive: line 31
Read-only work (audits, investigations): general-purpose agents told to read and not edit.
“Common” is measured: kept by at least 6 in 10 of the 6 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.
Files
Fan out, verify, consolidate. Parallel agents are fast and fresh-eyed; they also make confident mistakes. The value is in the verification step.
Procedure
- Split into units that share nothing. One per service, directory, capability (with a spec layer the registry already draws those lines), or failure domain. A unit fits one fresh context. Units that share state, a file, or a root cause are one unit — splitting them buys conflicts, not speed.
- Write one brief per unit, the same shape for all:
- scope: the paths, and "change nothing outside them";
- the question or change, stated once, with the exact criteria;
- the output contract: a table row per finding with file:line evidence, or for a change, the commits and the test command's output;
- "do not spawn subagents". Briefs go to files when they are long; the dispatch names the path.
- Dispatch every unit in one message so they run concurrently.
- Read-only work (audits, investigations): general-purpose agents told to read and not edit.
- Changes:
spec-system:implementer, each withisolation: "worktree", its own branch name, and the base commit. The same change in many places is still one task per place, reviewed as one diff when the change is mechanical. - Size the fan-out to the work, not the other way round.
- Verify before you consolidate.
- Discard findings without file:line evidence.
- Spot-check a sample of each agent's claims yourself; agents repeat the same systematic error across units.
- For changes: merge each branch (spec-system:merge-conflicts when needed), then run the full suite once on the merged result. Per-unit green does not prove the whole is green.
- Consolidate into one table: unit, finding or change, evidence, confidence. Mark what you could not confirm and say where you looked. With a spec layer, audit findings about behaviour cite the statement they contradict and go to the anomaly list — an audit records, it does not fix.
Rules
- Subagents do not spawn subagents. One level of fan-out.
- The consolidated result is yours: a finding you did not verify is labelled unverified, not dropped silently and not passed on as fact.
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
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.