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
- Ultimate-Alchemy/claude-skillstheir own version, first seen Oct 15, 2026
- MatanelGordon/skillstheir own version, first seen Aug 21, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- MatanelGordon/skills★ 0same name
- Ultimate-Alchemy/claude-skills★ 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 repository46 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- blast-promptSkill
- client-outreachSkill
- content-draftSkill
- copywritingSkill
- csv-to-reportSkill
- design-accessibilitySkill
- design-buildSkill
- design-croSkill
- design-dimensionsSkill
- design-motionSkill
- design-researchSkill
- design-system-genSkill
- design-system-persistSkill
- design-tokens-emitSkill
- design-visual-qaSkill
- html-extractSkill
- portable-html-portSkill
- qa-gateSkill
- route-codex-reviewSkill
- route-gemini-contextSkill
- route-three-brainSkill
- seo-auditSkill
- seo-backlinksSkill
- seo-clusterSkill
- seo-competitor-pagesSkill
- seo-content-briefSkill
- seo-contentSkill
- seo-dataforseoSkill
- seo-driftSkill
- seo-ecommerceSkill
- seo-firecrawlSkill
- seo-geoSkill
- seo-googleSkill
- seo-hreflangSkill
- seo-image-auditSkill
- seo-image-genSkill
- seo-local-unifiedSkill
- seo-pageSkill
- seo-programmaticSkill
- seo-schemaSkill
- seo-sitemapSkill
- seo-strategySkill
- seo-sxoSkill
- seo-technicalSkill
The file
Files
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-extractcan supply them) - Output mode: git worktrees | sibling folders
Steps
- Capture the brief and the N inspiration sources.
- Generate shared tokens once with
design-system-genso all variants share a coherent palette/type/effects — consistency across variants, distinctiveness via inspiration + layout. - Create N output locations — git worktrees (one per variant) or, if git isn't
available, sibling folders
variant-1/ … variant-N/. - Dispatch N build sub-agents in parallel behind the shared-token barrier — each
runs the
design-buildskill with the same brief + the one shared token set from step 2, differing only by inspiration source. (design-buildis 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). - Collect outputs and render a side-by-side comparison. When Playwright is
connected, fan out one
design-visual-qaagent per variant for screenshot capture and cross-variant diffs; otherwise render a structured comparison summary. - Cherry-pick + merge. The user picks sections from each; assemble the final and
run
qa-gatebefore 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 systemdesign-build(required) — runs the per-variant build inside each sub-agenthtml-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.
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.

