Skip to content

[Hackathon] Stark Spiders – Frontend Intelligence MCP (Gavel) - #327

Open
nidhi-nayana wants to merge 1 commit into
nitrocloudofficial:mainfrom
nidhi-nayana:add-sample-stark-spiders-gavel
Open

[Hackathon] Stark Spiders – Frontend Intelligence MCP (Gavel)#327
nidhi-nayana wants to merge 1 commit into
nitrocloudofficial:mainfrom
nidhi-nayana:add-sample-stark-spiders-gavel

Conversation

@nidhi-nayana

Copy link
Copy Markdown

⚖️ GAVEL: Frontend Intelligence MCP Server

Core Motto: "The rules decide. The LLM only phrases."

📌 Overview

Gavel is an AI Frontend Architect MCP Server built with NitroStack. When pointed at a real web application codebase, Gavel inspects actual project files (package.json, lockfiles, tailwind.config.js, component trees), gathers user intent through intelligent caching, evaluates 6 modern UI and animation libraries against deterministic engineering rules, calculates transparent confidence scores with visible math, generates actionable coding specs with extracted design tokens, and validates recommendations using automated before-and-after Lighthouse performance benchmarks—rendering live interactive NitroStack widgets inside AI chat interfaces.

🛠️ Key Features & Tools Included (10 MCP Tools):

  • elicitIntent: Elicits and caches human business goals in .gavel-context.
  • analyzeProject: Audits project structure, framework versions, and component depth.
  • inspectDependencies: Audits existing packages to prevent library overlap.
  • inspectDesignLanguage: Extracts hex colors, font scales, and spacing tokens.
  • recommendLibraries: Evaluates candidate rules with deterministic math and Groq rationale synthesis.
  • compareLibraries: Direct 1-v-1 architectural trade-off comparisons.
  • estimateBundleImpact: Computes minified/gzipped KB size and tree-shaking metrics.
  • generateDesignSpec: Synthesizes winner + extracted design tokens into actionable specs.
  • runLighthouse: Runs automated before/after Chrome Lighthouse performance audits.
  • compareMetrics: Measures Core Web Vitals performance delta between audit runs.

🧪 Verification

  • All 44 Vitest tests passing across 5 test suites.
  • TypeScript strictly typed (npm run typecheck passes cleanly).
  • Production build verified (npm run build succeeds).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant