Next.js Performance Audit & Fix Plan

    1

    Executable audit of your actual Next.js codebase: a bundled read-only scanner maps performance findings to files and lines, ranked by impact, then a staged fix plan and a before/after VERIFY pass.

    $12

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Next.js Performance Audit & Fix Plan

    Example session with this skill installed

    Audit the Next.js app in ./web. App Router, Next 14, hosted on Vercel. LCP is our problem. Audit only, no patches.

    • Read your context and instructions
    • Compiled the next js performance
    • Generated the document

    Next.js Performance Audit - fixture project (real scanner run on a teaching fixture)
    Router: app | Files scanned: 4 | Findings: 6 (high 3, medium 2, low 1)

    1. [HIGH] wide-client-boundary - app/layout.tsx:1 - root layout is a client component; the whole segment ships as client JS. Direction: move interactivity to a small leaf client component.
    2. [HIGH] barrel-import - app/layout.tsx:2 - root import of lodash in the root layout lands in every route's bundle. Direction: deep import or optimizePackageImports.
    3. [HIGH] barrel-import - app/components/Chart.tsx:1 - root import of @mui/material.
    4. [MEDIUM] raw-img-tag - app/page.tsx:12 - hero <img> bypasses next/image (also the likely LCP element).
    5. [MEDIUM] missing-optimize-package-imports - next.config.js.
    6. [LOW] possible-waterfall - app/page.tsx:6 - 3 consecutive awaits (heuristic, low confidence; check dependencies).

    No savings figures claimed - sizes are measured in VERIFY, not promised in AUDIT.

    layout.tsx

    PDF · document

    Generated

    Example file from a real run - the skill writes it into your workspace.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Best-practice guides tell your agent what good code looks like; this skill inspects the code you actually have. It opens with a 5-question Intake (router, versions, hosting, main pain, desired output) that changes which checks run and which reference applies, then runs a bundled Python scanner (standard library only, fully offline, read-only) over your Next.js project. The scanner detects barrel imports from heavy packages, overly wide 'use client' boundaries, raw tags, missing optimizePackageImports, eager heavy imports, and possible sequential-await waterfalls (labeled low-confidence heuristics). AUDIT mode delivers a findings report mapped to files and lines, ranked by impact. FIX PLAN mode turns findings into staged, minimal patch plans with risk, order, verification method, and rollback for each item - patches are only ever applied after your explicit per-batch approval. VERIFY mode re-runs the scanner and compares build output before/after, reporting an honest delta: what cleared, what remains, what is new, and what the evidence does and does not prove. Covers App Router and Pages Router, Vercel and self-hosted. No unmeasured savings claims, ever: sizes and timings appear only when measured.

    Static analysis only: runtime waterfalls inside shared helpers, third-party script cost, slow backends, and field Core Web Vitals cannot be seen or claimed. The sequential-await heuristic cannot distinguish dependent from independent awaits and will sometimes be wrong. Bundle findings are import-level indications until a build comparison in VERIFY measures them.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 25 days ago

    Requires

    Python 3
    Local Next.js project

    Permissions

    Terminal / Shell

    File Scopes

    Authorized
    project
    folder
    only

    The scanner runs fully offline with the Python standard library and only reads the project. No network calls, no API keys, no credentials. AUDIT and VERIFY never modify the project; patches require explicit per-batch approval.

    Listed25 days ago

    What's inside

    Frequently Asked Questions

    Guides list rules; this skill audits your actual codebase with a bundled scanner, maps findings to files and lines, and verifies fixes with a before/after pass.

    AUDIT and VERIFY are read-only (proven by checksum tests). Patches are drafted in FIX PLAN and applied only after your explicit per-batch approval.

    Yes. The scanner uses only the Python 3 standard library and makes no network calls.

    Yes. The Intake switches check families and references: _app weight, per-page bundles, and getServerSideProps data shape instead of RSC boundaries.

    No. It guarantees a mapped, ranked audit and honest measured deltas. Static findings are indications, and the skill never claims unmeasured savings.

    Yes. It exits 1 when any finding is medium or higher and 0 otherwise, and can emit JSON.

    Runtime behavior, third-party script cost, slow APIs, and field Web Vitals are invisible to static analysis; the report names these in its limitations section.

    No. It is an independent Suki Dev skill, written from scratch, not affiliated with or endorsed by Vercel.

    New in Workflow Automation