BjornMelin/dev-skills / plugins/claude-core/skills/fan-out
fan-out
VerifiedDelegates work to bounded subagents and synthesizes their evidence. Use when asked to delegate, fan out, parallelize, or spawn subagents, or when a task splits into 2-5 independent subtasks such as multi-area code search or parallel review. You stay the planner and read every result.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.3 KB |
| references/agent-templates.md | markdown | 2.0 KB |
references/agent-templates.md
View on GitHub# Fan-out reference: role stubs & Workflow skeleton
Load this only when you need a concrete custom role or a multi-stage Workflow. For ad-hoc
delegation, the inline spawn contract in `SKILL.md` is enough — no file needed.
## Custom `.claude/agents/*.md` role stubs
Authoring these is `claude-subagent-creator`'s job; these are illustrative shapes only.
```markdown
---
name: repo-explorer
description: Read-only fan-out search across a subsystem; returns located code, not reviews.
tools: Read, Grep, Glob, Bash
model: sonnet
---
You are a read-only explorer. Search the assigned scope, return file:line evidence and a short map.
Never edit files. Do not exceed the assigned scope.
```
```markdown
---
name: reviewer
description: Adversarial read-only reviewer; returns confirmed findings with severity + evidence.
tools: Read, Grep, Glob, Bash
model: opus
---
You are a skeptical reviewer. Default to refuting a finding unless the evidence is concrete.
Return: finding, severity, file:line, why-real, confidence. No edits.
```
```markdown
---
name: implementation-worker
description: Scoped edit-worker that owns a disjoint file set; fixes then verifies.
tools: Read, Edit, Write, Bash
model: opus
---
Edit ONLY the files named in your task. You are not alone in this repo — never touch or revert
files outside your set. After editing, run the named verification and report the result.
```
## Workflow skeleton (dependent / DAG batches)
When stages depend on each other, prefer the dynamic Workflow tool over hand-rolled batches:
```
phase('Find')
const found = await parallel(AREAS.map(a => () =>
agent(`Search ${a} for X. Return file:line evidence.`, { schema: FINDINGS })))
phase('Verify')
const verified = await pipeline(found.flat().filter(Boolean),
f => agent(`Adversarially verify: ${f.title}`, { schema: VERDICT }))
return verified.filter(v => v?.isReal)
```
Rules: `pipeline` by default (no barrier between stages); `parallel` only when a stage truly needs
all prior results at once; always `.filter(Boolean)` returns (a dropped agent resolves to `null`).
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |