agentleFS
ZachArticulateVZachArticulateV/designer-pro-and-seo/skills/parallel-build/SKILL.md

parallel-build

Skill10 starsChanged 26 days agoMITView on GitHub

Spin up multiple website or page variants in parallel — each in its own git worktree or sibling folder, built by a sub-agent from the same brief and shared design tokens but a different inspiration source — then cherry-pick and merge the best sections into one final. Captures side-by-side comparison screenshots when Playwright is connected; otherwise renders a structured comparison summary. Trigger when the user says "parallel build", "build three versions", "spin up variants", "make 3 sites at once", "side-by-side build", or "different design flavors to compare".

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

parallel-build

Family: build-and-qa Status: Stable

Purpose

The multi-variant build primitive: instead of one prompt → one site, spawn N sub-agents (default 3), each in its own worktree/folder, each given a different inspiration source but the same brief and the same design tokens. The user gets a side-by-side comparison and cherry-picks the best sections into a merged final.

Triggers

  • "parallel build" / "build three versions" / "spin up variants"
  • "make 3 sites at once" / "side-by-side build"
  • "different design flavors to compare"

Inputs

  • Target (business/page/client) + the brief
  • N variants (default 3) and N inspiration sources (one per variant; html-extract can supply them)
  • Output mode: git worktrees | sibling folders

Steps

  1. Capture the brief and the N inspiration sources.
  2. Generate shared tokens once with design-system-gen so all variants share a coherent palette/type/effects — consistency across variants, distinctiveness via inspiration + layout.
  3. Create N output locations — git worktrees (one per variant) or, if git isn't available, sibling folders variant-1/ … variant-N/.
  4. Dispatch N build sub-agents in parallel behind the shared-token barrier — each runs the design-build skill with the same brief + the one shared token set from step 2, differing only by inspiration source. (design-build is an orchestrator, so each variant runs it as a skill inside its own sub-agent — not as a dispatched-leaf agent; the barrier guarantees every variant shares one palette/type/effects system.) One-shot each (iterate only if needed).
  5. Collect outputs and render a side-by-side comparison. When Playwright is connected, fan out one design-visual-qa agent per variant for screenshot capture and cross-variant diffs; otherwise render a structured comparison summary.
  6. Cherry-pick + merge. The user picks sections from each; assemble the final and run qa-gate before delivery.

Outputs

  • N variant builds in their output locations
  • A side-by-side comparison artifact
  • The merged final (after the user picks)

Dependencies

  • design-system-gen (required) — generates the shared tokens once so every variant shares one palette/type/effects system
  • design-build (required) — runs the per-variant build inside each sub-agent
  • html-extract (optional — adds auto-sourced inspiration per variant; free path: user supplies the inspiration sources)
  • design-visual-qa (optional — adds side-by-side comparison screenshots via Playwright; free path: structured comparison summary)
  • qa-gate (optional — adds a structured PASS/FAIL gate on the merged final; free path: manual review before handoff)
  • Sub-agent dispatch + git worktrees (optional — falls back to sibling folders variant-1/ … variant-N/)

Notes

Shared tokens + divergent inspiration is the formula: variants feel like one brand exploring directions, not three unrelated sites.

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.