Feature-Sliced Design Skill - Setup Guide¶

Source: feature-sliced/skills GitHub: feature-sliced/skills (80 stars) Skills: 1 skill (feature-sliced-design) · 16.6K installs Category: Frontend Architecture First Seen: Mar 9, 2026 (catalogued August 15, 2026 midday sweep) Quality Tier: 🟢 Production (all three security audits pass - Gen Agent Trust Hub, Socket, Snyk)

This is the official agent skill for Feature-Sliced Design (FSD) v2.1, the frontend architecture methodology documented at fsd.how. Its core principle is "start simple, extract when needed": place code in pages/ first, accept duplication, and extract to shared/app/features/entities layers only when code is actually reused and boundaries are focused.

Installation¶

npx skills add feature-sliced/skills --skill feature-sliced-design

Prerequisites¶

Requirement	Details
Node.js + npx	For the skill installer
Frontend project	Any JS/TS app where structure decisions are being made

What It Provides¶

Capability	Notes
Layer system	pages/, shared/, app/, features/, entities/ with clear extraction rules
Pragmatic doctrine	Not all layers required; start with shared, pages, app
Extraction criteria	Extract only on actual reuse, stable boundaries, focused responsibility
Anti-overengineering guard	Empty layer folders "just in case" are explicitly discouraged

Quick Start¶

npx skills add feature-sliced/skills --skill feature-sliced-design
Ask: "restructure this frontend project following FSD v2.1, starting from pages/ and extracting only what is genuinely shared"

CorpusIQ Use Cases¶

Use Case	How
Frontend architecture standard	A consistent structure doctrine for agent-built frontends
Refactor guidance	Rules for when to extract vs leave duplication
Client project scaffolding	Architecture decisions on new builds
Review rubric	Check structural drift against FSD principles

Integration with CorpusIQ¶

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects

Limitations / Verification¶

Single-skill cluster; all three security audits pass (verified on skill page)
Methodology skill - no runtime dependencies, but also no tooling; it is pure procedural knowledge
Strictness is adjustable per project scale and team context

npx skills add feature-sliced/skills   # verify install works

Powered by CorpusIQ