Synapse

Component-level frontend craft toolkit with 20 iterative design commands (/critique, /audit, /normalize, /polish, /harden, /animate, /colorize, /distill, /typeset, /arrange, /delight, /clarify, /optimize). Use to refine, critique, or polish EXISTING UI components and pages — spacing, typography, motion, states, micro-interactions — or to run a specific refinement command. For generating a whole new page from a brief use hallmark; for a data-backed design system use ui-ux-pro-max.

Anti-AI-slop design skill with four explicit modes — build (new pages), audit (score existing UI against 57 slop-test gates), redesign (rebuild with a different fingerprint while preserving copy and IA), and study (extract design DNA from a URL or screenshot into a portable design.md). Use when the user says audit, redesign, or study, invokes Hallmark by name, or wants a greenfield page that does not look AI-generated. Includes 20 built-in themes and a pre-emit self-critique pass.

Premium redesign of existing sites and apps (taste-skill family): audits the current design, removes generic AI patterns, and applies high-end standards without breaking functionality; works with any CSS framework or vanilla CSS. Complements hallmark's redesign mode; use for upgrading an existing site's visual quality in place.

Aesthetic-direction design skill: reads a brief, infers the right visual style and direction, and ships interfaces that do not look templated. Umbrella for concrete style variants — minimalist, brutalist, soft/premium-agency. Use when the user wants a particular vibe or aesthetic, or wants a landing page or portfolio to feel handcrafted and expensive. For explicit build/audit/redesign/study modes use hallmark; for data-backed tokens use ui-ux-pro-max.

Data-backed design-system generator: given a product type or industry, a searchable local database returns concrete recommendations — pattern, style, color palette, typography, effects and anti-patterns — spanning 84 styles, 192 palettes, 74 font pairings, 25 chart types and 22 stacks (React, Next.js, Vue, Svelte, SwiftUI, Flutter, Tailwind, shadcn and more). Use when you need concrete data-driven design tokens and recommendations for a specific product, not just design principles. For principle-based anti-slop craft use hallmark or taste-skill.

Component-level frontend craft toolkit with 20 iterative design commands (/critique, /audit, /normalize, /polish, /harden, /animate, /colorize, /distill, /typeset, /arrange, /delight, /clarify, /optimize). Use to refine, critique, or polish EXISTING UI components and pages — spacing, typography, motion, states, micro-interactions — or to run a specific refinement command. For generating a whole new page from a brief use hallmark; for a data-backed design system use ui-ux-pro-max.

Data-backed design-system generator: given a product type or industry, a searchable local database returns concrete recommendations — pattern, style, color palette, typography, effects and anti-patterns — spanning 84 styles, 192 palettes, 74 font pairings, 25 chart types and 22 stacks (React, Next.js, Vue, Svelte, SwiftUI, Flutter, Tailwind, shadcn and more). Use when you need concrete data-driven design tokens and recommendations for a specific product, not just design principles. For principle-based anti-slop craft use hallmark or taste-skill.

Anti-AI-slop design skill with four explicit modes — build (new pages), audit (score existing UI against 57 slop-test gates), redesign (rebuild with a different fingerprint while preserving copy and IA), and study (extract design DNA from a URL or screenshot into a portable design.md). Use when the user says audit, redesign, or study, invokes Hallmark by name, or wants a greenfield page that does not look AI-generated. Includes 20 built-in themes and a pre-emit self-critique pass.

Aesthetic-direction design skill: reads a brief, infers the right visual style and direction, and ships interfaces that do not look templated. Umbrella for concrete style variants — minimalist, brutalist, soft/premium-agency. Use when the user wants a particular vibe or aesthetic, or wants a landing page or portfolio to feel handcrafted and expensive. For explicit build/audit/redesign/study modes use hallmark; for data-backed tokens use ui-ux-pro-max.

Premium brand-kit image generation skill for creating high-end brand-guidelines boards, logo systems, identity decks, and visual-world presentations. Trained for minimalist, cinematic, editorial, dark-tech, luxury, cultural, security, gaming, developer-tool, and consumer-app brand systems. Optimized for intentional logo concepting, refined composition, sparse typography, strong symbolic meaning, premium mockups, art-directed imagery, and flexible grid layouts.

Raw mechanical interfaces fusing Swiss typographic print with military terminal aesthetics. Rigid grids, extreme type scale contrast, utilitarian color, analog degradation effects. For data-heavy dashboards, portfolios, or editorial sites that need to feel like declassified blueprints.

Elite website image-to-code skill for Codex. For visually important web tasks, it must first generate the design image(s) itself, deeply analyze them, then implement the website to match them as closely as possible. In Codex, it must prefer large, readable, section-specific images instead of tiny compressed boards, generate fresh standalone images for sections or detail views instead of cropping old ones, avoid lazy under-generation, avoid cards-inside-cards-inside-cards UI, and keep the hero clean, spacious, readable, and visible on a small laptop.

Clean editorial-style interfaces. Warm monochrome palette, typographic contrast, flat bento grids, muted pastels. No gradients, no heavy shadows.

Overrides default LLM truncation behavior. Enforces complete code generation, bans placeholder patterns, and handles token-limit splits cleanly. Apply to any task requiring exhaustive, unabridged output.

Teaches the AI to design like a high-end agency. Defines the exact fonts, spacing, shadows, card structures, and animations that make a website feel expensive. Blocks all the common defaults that make AI designs look cheap or generic.

Full-site SEO audit orchestrator. Crawls the site, detects business type, fans out to the seo-* specialist skills, then aggregates an evidence-backed 0-100 health score and a prioritized action plan (Critical → Low). Use when user says "audit", "full SEO check", "analyze my site", "website health check", or "SEO report". Triggers on: seo audit, full seo check, site health score, action plan, crawl my site, technical+content+schema audit.

Off-page/backlink audit: link-profile analysis, referring-domain quality, anchor-text distribution, toxic-link detection, disavow guidance, expired-domain heritage risk, parasite-SEO exposure, and digital-PR link earning. Use when auditing or building backlinks, assessing a link profile, judging link toxicity, or planning link acquisition. Triggers on: backlinks, link profile, referring domains, anchor text, toxic links, disavow, link gap, link building, digital PR, expired domain, parasite SEO, off-page SEO.

Build and audit competitor comparison, "alternatives to X", and category roundup pages — competitive content gap analysis, SERP-feature targeting, page-level teardown, feature matrices, JSON-LD, and a 0-100 prioritization rubric. Use when user says "comparison page", "vs page", "alternatives page", "competitor comparison", "content gap", "SERP analysis", "X vs Y". Triggers on: versus, compare competitors, alternative to, comparison table, gap analysis, SERP features, competitive teardown.

Content and translation drift detection: find pages decaying over time, machine-translation quality drift on i18n sites, staleness signals, and prioritize what to refresh. Methodology, thresholds, a 0-100 freshness score, and refresh cadence tied to Next.js + next-intl + hreflang. Use when content is going stale, traffic is decaying on old pages, translations drifted from source, deciding what to refresh, or planning a content-decay audit. Triggers on: content drift, content decay, stale content, translation drift, machine translation quality, i18n parity drift, refresh priority, content freshness, content rot, decaying pages, next-intl drift, hreflang parity.

Content quality, E-E-A-T analysis, search-intent mapping, content-brief generation, and AI-citation readiness for Next.js sites. Use when the user says "content quality", "E-E-A-T", "content analysis", "content audit", "readability check", "thin content", "content brief", "content outline", "search intent", or "helpful content". Triggers on: E-E-A-T, content audit, content brief, search intent, citability, helpful content, thin content, AI citation, GEO content.

OPTIONAL reference for wiring a programmatic crawler / scraper / mass-Lighthouse runner for SEO at scale (vendor-agnostic, via MCP or CLI). Use when you need to crawl a whole site, map all URLs, scrape JS-rendered pages, detect broken links/orphans, or run Lighthouse across every route. Triggers on: "crawl site", "map site", "full crawl", "find all pages", "broken links", "orphan pages", "scrape JS-rendered", "site-wide Lighthouse", "mass PageSpeed", "crawl at scale", "scraping for SEO".

E-commerce SEO for product and category pages — Product/Offer/AggregateRating JSON-LD, faceted navigation & crawl control, variant/canonical handling, out-of-stock and pagination, category page optimization, and merchant feed basics. Includes Next.js 15 App Router + Payload code and a 0-100 audit rubric. Use when working on ecommerce SEO, product SEO, shop/store pages, product schema, category pages, faceted filters, merchant feed, or shopping visibility. Triggers on: ecommerce SEO, product SEO, product schema, Product JSON-LD, Offer, AggregateRating, faceted navigation, canonical variants, out of stock SEO, pagination SEO, category page SEO, merchant feed, shopping.

Generative Engine Optimization (GEO) — audit and implementation for AI search surfaces: Google AI Overviews and AI Mode, ChatGPT search, Perplexity, Bing Copilot. Covers AI-crawler accessibility, question-based citability scoring, passage extraction, server-side rendering checks for AI bots, llms.txt (with the primary-source caveat on real support), agent-friendly page checks, and a 0-100 scoring rubric. Use when user says "AI Overviews", "AI Mode", "SGE", "GEO", "AI search", "LLM optimization", "Perplexity", "AI citations", "ChatGPT search", "agent-friendly", or "AI visibility". Triggers on: GEO, AEO, AI Overviews, AI Mode, generative engine optimization, llms.txt, AI crawler, citability, Perplexity, ChatGPT search.

Ground-truth reference for what Google Search actually says — Search Essentials, spam policies, E-E-A-T/YMYL, Core Web Vitals thresholds, structured-data eligibility, algorithm-update awareness (core/helpful-content/spam) and how to react, plus a falsifiability methodology for SEO claims. Use when verifying an SEO claim against primary sources, diagnosing a ranking drop after an update, or deciding whether a practice is allowed. Triggers on: "is this allowed by Google", "Google guidelines", "Search Essentials", "spam policy", "E-E-A-T", "YMYL", "helpful content", "core update", "ranking drop", "manual action", "penalty recovery", "Core Web Vitals", "LCP INP CLS", "structured data eligibility", "rich results retired", "what does Google say".

Hreflang and international SEO audit, validation, and generation for Next.js + next-intl. Detects return-tag/self-ref/x-default errors, validates language and region codes, scores content parity, and emits correct hreflang via metadata, sitemap, or HTTP headers. Use when working on hreflang, i18n SEO, international SEO, multi-language, multi-region sites, or alternate language tags. Triggers on: hreflang, i18n SEO, international SEO, multi-language, multi-region, language tags, x-default, alternate locale, next-intl SEO.

Image SEO and performance for Next.js. Covers alt text, descriptive filenames, next/image responsive output, AVIF/WebP, lazy loading and LCP priority, CLS prevention, IPTC/XMP metadata, AI-generated image labeling (DigitalSourceType TrainedAlgorithmicMedia), OG/social images, and image sitemaps. Use when working on image optimization, alt text, image audit, or image SEO. Triggers on: image SEO, alt text, next/image, image optimization, image audit, WebP, AVIF, image metadata, OG image, image sitemap.

Local SEO for brick-and-mortar, service-area, and multi-location businesses — Google Business Profile signals, NAP consistency, citations, reviews, local pack & Maps ranking factors, LocalBusiness JSON-LD, and location-page quality. Use when working on local search, map pack, store locator, or location pages. Triggers on: local SEO, Google Business Profile, GBP, map pack, local pack, NAP, citations, local rankings, service area, multi-location, LocalBusiness schema, geo-grid, store locator.

Audit-driven SEO planning that turns audit findings into a sequenced, owned roadmap with effort/impact scoring, leading indicators, and falsifiable success criteria. Industry-specific architecture templates, competitive analysis, and content strategy for Next.js sites. Use when user says "SEO plan", "SEO strategy", "SEO roadmap", "content strategy", "keyword strategy", "content calendar", "site architecture", or "turn this audit into a plan". Triggers on: SEO plan, SEO roadmap, SEO strategy, audit to roadmap, prioritize SEO fixes, content calendar, site architecture, effort impact SEO.

OPTIONAL reference for wiring third-party rank-tracking, keyword, backlink, SERP, and AI-visibility DATA APIs into an SEO workflow via an MCP server. Vendor-agnostic: explains what each data class provides, when to reach for live data vs. first-party signals, how to wire an MCP server, and how to keep API spend under control. Not required by any other seo-* skill. Use when user says they have (or want) a rank-tracking / keyword / backlink data API, or asks how to get live SERP positions, search volume, keyword difficulty, referring-domain counts, or AI share-of-voice. Triggers on: rank tracking, keyword data API, search volume, keyword difficulty, backlink API, referring domains, live SERP, SERP API, AI visibility, share of voice, LLM mentions, rank data MCP, DataForSEO, Ahrefs, SE Ranking, Semrush, Moz.

Detect, validate, and generate Schema.org structured data (JSON-LD) for rich-result eligibility, with Next.js 15 App Router injection patterns. Use when adding, auditing, or fixing structured data. Triggers on: schema, structured data, rich results, JSON-LD, markup, schema.org, rich snippets, Organization schema, Product schema, BreadcrumbList, deprecated schema.

Advanced XML sitemaps for Next.js App Router — index/shard files, news/image/video sitemaps, lastmod hygiene, 50k sharding, IndexNow ping, plus an audit rubric. Use when user says "sitemap", "sitemap index", "news sitemap", "image sitemap", "video sitemap", "large site sitemap", "IndexNow", "lastmod", or "sitemap audit". Triggers on: sitemap, sitemap.xml, sitemap index, news sitemap, image sitemap, video sitemap, lastmod, sharding, IndexNow, robots sitemap reference.

Umbrella SEO & GEO knowledge base for Next.js 15 App Router — robots/sitemap/metadata/JSON-LD code, Core Web Vitals (INP), primary-source Google guidance, and a router to specialist seo-* skills. Use when implementing SEO, setting up metadata/structured data, optimizing Core Web Vitals, preparing for AI search, or deciding which SEO sub-skill to load. Triggers on: SEO, GEO, metadata, sitemap, robots.txt, JSON-LD, structured data, Core Web Vitals, INP, LCP, CLS, E-E-A-T, AI Overviews, hreflang, technical SEO.

Search Experience Optimization — the SEO×UX×CRO overlap. Reads the SERP backwards to detect page-type mismatch, derives user stories from intent signals, scores a page from multiple persona perspectives, and fixes engagement/page-experience problems (dwell, pogo-sticking, CWV) that block ranking even on technically perfect pages. Includes a 0-100 SXO gap rubric with falsifiability checks and Next.js patterns. Use when the user says "SXO", "search experience", "page type mismatch", "intent mismatch", "why isn't my page ranking", "pogo-sticking", "dwell time", "engagement signals", "conversion-aware content", "persona scoring", or "SERP analysis". Triggers on: SXO, search experience optimization, page-type mismatch, intent mismatch, pogo-sticking, dwell time, engagement signals, persona scoring, SERP backwards analysis, conversion-aware SEO, page experience.

Technical SEO audit + implementation across 9 categories: crawlability, indexability, security headers, URL structure, mobile, Core Web Vitals (LCP/INP/CLS), structured data, JavaScript rendering (CSR vs SSR), and IndexNow. Includes Next.js robots.ts/headers/metadata patterns and a 0-100 scoring rubric. Use when working on technical SEO, crawl/index issues, or CWV. Triggers on: technical SEO, crawlability, indexability, robots.txt, robots.ts, sitemap, canonical, security headers, HSTS, CSP, Core Web Vitals, LCP, INP, CLS, JS rendering, SSR, IndexNow, AI crawlers, GPTBot, ClaudeBot.

Build topical authority with pillar/cluster (hub-and-spoke) content architecture, internal-linking strategy, content-gap analysis, and cannibalization avoidance — with Next.js App Router internal-linking patterns and a coverage-scoring rubric. Use when planning content architecture, building topical authority, designing a pillar page with supporting cluster pages, fixing keyword cannibalization, or auditing internal links. Triggers on: topic cluster, content cluster, topical authority, pillar page, hub and spoke, internal linking, content gap, keyword cannibalization, semantic clustering, content architecture, orphan pages.

Audit websites for SEO, performance, security, technical, content, and 15 other issue cateories with 230+ rules using the squirrelscan CLI. Returns LLM-optimized reports with health scores, broken links, meta tag analysis, and actionable recommendations. Use to discover and asses website or webapp issues and health.

Use this before any creative work — creating features, building components, adding functionality, or modifying behavior. Use when user asks to design, plan, or think through a feature before implementing. Explores intent, requirements, and design through collaborative dialogue.

Captures a validated learning into the Memory Graph (SQLite). Invoke when: a bug is resolved non-obviously, a pattern is discovered, the user corrects a mistake, or a solution took 2+ attempts. Do NOT invoke for typos, obvious errors, or one-time config changes.

Use at the START of any coding session or before touching code in a project. Triggers on: "lavora su X", "modifica X", "aggiungi feature a X", "fix in X", any request that implies working inside a specific project folder. Ensures the CodeGraph index is fresh and loads the 5-layer Cortex briefing. Without this, you are working blind — no impact analysis, no memory, no context.

Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies. Triggers on: "multiple failures", "parallel work", "independent problems", any multi-domain issue. Dispatches one agent per independent problem domain for concurrent investigation.