CyranoB

Deep web research skill that searches the web and fetches full page content to produce comprehensive, well-sourced research reports. Use this skill whenever the user asks to research a topic, investigate a question, do a deep dive, compare products or services, look something up online, find information about a product/company/technology, or gather current information from the web. Trigger on phrases like "can you find out about X", "look up X", "what do you know about X (current/recent)", "tell me about X", "research X for me", "help me understand X", "what are people saying about X", "I'm choosing between X and Y", or "search the web for X" when current web info would be valuable. Also trigger when the user needs recent news, pricing, reviews, or real-world adoption data that may have changed since training. Don't trigger for purely conceptual questions that Claude can answer from training data without needing live web sources — unless the user explicitly asks to search the web.

ALWAYS use this skill when a user wants to understand a competitive landscape or position a product against competitors. Two modes: (1) Market landscape — user has a product idea or wants to map a market: "who are the competitors in X space", "I want to build a Y — what's out there", "market study for Z". (2) Competitive positioning — user has a product and wants strategic comparison: "how do we compare to competitors", "what are we missing vs the competition", "competitive analysis of our product". Trigger on: "competitive landscape", "market study", "who are the competitors", "competitive analysis", "I want to build X — what exists", "what are the alternatives to X", "how does X compare to the market", "gap analysis", "market map". Do NOT trigger for personal buying decisions (use tech-advisor), general web research, news monitoring, or fact-checking.

Verify claims and statements by searching for supporting and contradicting evidence, then deliver a clear verdict with confidence level and sources. Use this skill when the user asks to fact-check something, verify a claim, check if something is true, or questions the accuracy of a statement. Trigger on phrases like "is it true that X", "fact check this", "verify this claim", "is X really Y", "I heard that X — is that accurate?", "can you confirm that X", "someone told me X", or "this article says X, is that right?". Don't trigger for general research questions or product comparisons — only when the user has a specific claim they want verified.

Search for recent news and developments on a topic, organize them chronologically, and deliver a concise briefing. Use this skill when the user wants to catch up on recent events, news, or developments around a topic. Trigger on phrases like "what's new with X", "recent news about X", "any updates on X", "what happened with X lately", "catch me up on X", "news roundup for X", "what did I miss about X", "latest developments in X", or "has anything changed with X recently". Also trigger when the user mentions a time frame like "this week", "this month", "since January", or "in the last few days" combined with wanting information. Don't trigger for general research, product comparisons, or fact-checking — only when recency is the point.

ALWAYS use this skill when a user needs help choosing between tech products or evaluating whether a technology is ready for adoption. Searches the web for current data and produces structured, evidence-based recommendations with comparison tables and sourced evidence. Two modes: (1) Product comparison — "X vs Y", choosing between devices, editors, services, frameworks, laptops, or any tech products. (2) Maturity assessment — "is X production ready", "should we adopt X". Trigger on: "MacBook vs ThinkPad", "Supabase vs Firebase", "is Bun stable enough", "help me pick a message broker", "which iPad for note-taking", "our CTO wants us to evaluate X", or any question about choosing tech or evaluating readiness. Do NOT trigger for general web research, news, fact-checking, competitive intel, or debugging.

Build and deploy polished landing pages and one-page websites to Vercel from a text description. Use this skill when a user wants any single-page web presence — landing pages, product pages, startup sites, business websites, portfolio sites, event pages, coming soon pages, or promotional pages. Triggers when someone describes a business, product, or service and wants a live website, even if they say "website" or "site." Common triggers: "build me a landing page," "I need a website for my business," "create a page for my product," "make a website for my startup," or any business description (like "I have a piano teaching business" or "I run a photography studio") paired with wanting a web presence. Generates a single self-contained HTML file with distinctive design, deployed to Vercel instantly — no auth required. Do NOT use for multi-page apps, e-commerce, dashboards, docs sites, blogs, or debugging existing code.

Turn a video file into a premium scroll-sequence website — the Apple-style experience where scrolling scrubs through video frames with choreographed text animations. Use this skill whenever the user provides a video file (MP4, MOV, WebM, etc.) and wants it turned into a scroll-sequence site, scroll-driven landing page, or scrollytelling experience. Also triggers when the user mentions "scroll sequence," "scrollytelling," "scroll-driven animation," "video frames on scroll," "Apple-style scroll animation," "frame-by-frame scroll," or wants a product reveal site built from video footage. If the user has a product video, demo reel, or any footage they want presented as a polished scroll-sequence web experience rather than just an embedded video player, this is the skill to use.

Deploy any static site to AWS S3 + CloudFront with HTTPS, global CDN, and no size limit. Use this skill when the user wants to deploy to AWS, host on S3, deploy to CloudFront, put a site on AWS, or needs to deploy assets that exceed Vercel's 4.5MB limit (e.g., full-resolution scroll-sequence sites). Triggers on "deploy to AWS," "host on S3," "deploy to CloudFront," "put this on AWS," "deploy with no size limit," or any request to host a static site on AWS infrastructure.

Deploy any static site directory to Google Cloud (Firebase Hosting) using your local CLI. Use this skill whenever the user wants to deploy, host, or publish a static website to GCP/Firebase. Triggers on "deploy this to GCP," "host on Google Cloud," "publish to Firebase," "deploy to GCP/Firebase," or any request to host a local static site on Google infrastructure. Works with single HTML files, multi-file projects, and asset-heavy directories. Requires the Firebase CLI (`firebase-tools`) to be installed and authenticated locally.

Deploy any static site directory to Vercel instantly — no authentication required. Use this skill whenever the user wants to deploy, host, publish, or put live any static website, HTML project, or directory containing an index.html. Triggers on "deploy this to Vercel," "put this live," "host this site," "publish this page," "deploy this directory," "I want to see this online," or any request to make a local static site accessible via a public URL. Works with single HTML files, multi-file projects (HTML + CSS + JS), and asset-heavy directories (images, fonts, frames). Returns a live preview URL and a claim URL to transfer ownership.

Score text for AI-generated writing patterns without rewriting it. Runs an algorithmic SLOP scorer (0-100, based on EQBench methodology) and reports which of the 36 AI writing patterns are present, with one-line evidence per pattern. Verdict only — no rewrite, no audit, no humanization. Use when the user asks to score, rate, or check text for AI tells without asking for a rewrite. Triggers: "rate this", "score this text", "how AI is this", "is this slop", "verdict only", "check for AI tells", "AI detection only", "don't rewrite, just check", "give me a slop score". Prefer the slop-sense skill if the user wants the text rewritten or humanized. Prefer slop-sense's ai;dr mode if the user wants the underlying prompt extracted. Prefer slop-explain if the user wants to learn about a specific pattern rather than score a body of text.

Detect, score, rewrite, and explain AI writing patterns (36 known tells, EQBench SLOP methodology). Four modes: (1) Rewrite — humanize text, remove AI tells. Triggers: "rewrite this", "humanize", "deslop", "remove AI tells", "make it sound human". (2) Verdict only — score 0-100 and list patterns, no rewrite. Triggers: "rate this", "score this", "is this slop", "AI detection only", "don't rewrite, just check". (3) Pattern deep-dive — explain one of the 36 patterns. Triggers: "explain pattern N", "what is significance inflation", "tell me about em dash overuse", "why do LLMs use 'delve'". (4) ai;dr — extract the probable prompt behind AI text. Triggers: "ai;dr", "what was this AI asked to write", "extract the prompt".

Detect and remove AI-generated writing patterns from text. Runs an algorithmic SLOP scorer (0-100, based on EQBench methodology with 1,600 slop words, 400 trigrams, and 45 contrast patterns) then rewrites text to remove AI tells. Use this skill whenever the user asks you to check if text sounds AI-generated, make text sound more human, remove AI patterns, humanize writing, reduce "slop", or review text for AI tells. Also use it when editing any text that might have been AI-generated, even if the user doesn't explicitly mention AI detection. If the text reads like it came from an LLM, suggest using this skill. Also use when the user says "ai;dr", asks to extract the prompt from AI text, or wants to know what an AI was asked to write.

Educational deep-dives on the 36 AI writing patterns. Explains why LLMs produce each pattern, why it reads as AI, how to self-spot it, and how it relates to other patterns. One pattern per invocation. Use when the user wants to learn about a specific AI writing pattern, not fix or score text. Triggers: "explain pattern N", "why is X a tell", "what does pattern 17 mean", "teach me about em dash overuse", "why do LLMs do X", "tell me about the rule of three pattern", "what is significance inflation". Prefer slop-sense if the user wants text rewritten or humanized. Prefer slop-check if the user wants a score or verdict on a body of text. This skill is for learning, not for fixing.

Use when the user asks to review code, lint files, fix linting errors, audit a project, run static analysis, check types, inspect complexity, find circular dependencies, review architecture, scan for vulnerabilities or secrets, run pre-commit checks, or set up ESLint, Biome, ruff, pyright, tsc, madge, depcycle, Semgrep, or detect-secrets.

Landing page outline generator for small and local businesses. Conducts a short interview to learn about the business, then produces a structured section-by-section landing page outline. Use when someone wants to create a website outline, landing page plan, or homepage structure for their small business. Triggers on requests like "create a landing page for my business," "I need a website outline," "help me plan my business website," "landing page for my plumbing company," "website outline for my salon," "I need a homepage for my car wash," "help me with my small business website," "create a page for my babysitting service," "page for my piano teaching business," "website for my photography business," "landing page for my tutoring services," or "I need a website for my DJ business."

Interactive startup ideation and planning pipeline that walks users through generating business ideas, target personas, elevator pitch, customer avatar, diary entry, feature list, pricing strategy, and landing page outline. Use when the user wants to brainstorm startup ideas, develop a product concept, plan a new business, create a go-to-market strategy, or work through any stage of early-stage startup planning. Also triggers on requests like "help me come up with a business idea," "I need a startup plan," "generate a pitch for my idea," or "help me think through my product."