feat(claude): add trend-arbitrage skill

dot_claude/skills/trend-arbitrage/SKILL.md

@@ -0,0 +1,202 @@
+---
+name: trend-arbitrage
+description: |
+  End-to-end workflow for finding content, app, service, or product opportunities
+  where search demand is high but quality supply is low (trend arbitrage).
+  Use when: finding content gaps, keyword opportunities, niche research,
+  analyzing whether a topic or idea is worth pursuing, or competitive gap analysis.
+---
+
+# Trend Arbitrage Skill
+
+Find "Volume: High / Difficulty: Low" gaps in any market and turn them into
+actionable content or product strategies. Based on the methodology of identifying
+demand-supply mismatches before competitors notice them.
+
+## Core Concept
+
+Trend Arbitrage = Finding topics where **search demand is growing** but **quality supply is scarce**.
+
+Three conditions must ALL be true:
+1. **Breakout** — Search volume rising sharply in recent months
+2. **Information Gap** — Top results are outdated, thin, or user-generated (Q&A sites, forums)
+3. **High Intent** — Searchers want to solve a problem, buy something, or learn a specific skill
+
+## Initial Classification
+
+Before starting any phase, use `ask_user_input` to classify the opportunity type unless
+the user has already made it clear. This determines which discovery methods, validation
+criteria, and strategy templates to use.
+
+**Always ask upfront:**
+
+1. **What type of gap are you looking for?**
+   - Content gap (blog, newsletter, guide, course)
+   - App/Tool gap (SaaS, utility, mobile app)
+   - Service gap (consulting, subscription service, productized service)
+   - Product gap (digital product, template, dataset, curated list)
+
+2. **What market or niche?** (free text if not already specified)
+
+3. **What's your time horizon?**
+   - Quick win (launch in 1-2 weeks)
+   - Medium play (1-3 months to build)
+   - Long-term bet (3-12 months, bigger moat)
+
+The answers shape every subsequent phase — discovery sources, validation criteria,
+strategy templates, and execution plans all differ by type.
+
+## Workflow Overview
+
+The skill operates in 4 phases. Run them sequentially or jump to any phase.
+
+```
+CLASSIFY  →  Phase 1: DISCOVER  →  Phase 2: VALIDATE  →  Phase 3: STRATEGIZE  →  Phase 4: EXECUTE
+(Ask type)   (Find signals)        (Confirm the gap)      (Design the play)       (Create the plan)
+```
+
+Read `references/phases.md` for the detailed procedure of each phase.
+
+## Phase 1: DISCOVER — Find Gap Signals
+
+**Goal**: Generate a list of 5-10 candidate opportunities showing breakout signals.
+
+**Inputs**: Determined by classification above. If anything is ambiguous, ask via `ask_user_input`.
+
+**Methods by type** (use web_search for all):
+
+### Content gaps:
+1. Google Trends exploration — rising queries in the niche
+2. Community pain points — Reddit, forums, Q&A sites with repeated unanswered questions
+3. Supply-side audit — search results dominated by outdated/thin content
+
+### App/Tool gaps:
+1. "I wish there was an app for..." signals — Reddit, X/Twitter, Hacker News, Product Hunt comments
+2. Existing tool complaints — App Store/Play Store reviews with recurring frustration patterns
+3. Technology accessibility gaps — powerful tech (AI, APIs) that lacks a simple UI wrapper
+4. Adjacent tool search — what tools exist in adjacent markets but not this one?
+
+### Service gaps:
+1. Hiring pain signals — "looking for [role]" posts with complaints about cost/quality/speed
+2. Process pain — "how do I [task]" where answers are complex and people clearly want someone to do it for them
+3. Freelancer marketplace gaps — Fiverr/Upwork categories with low ratings or sparse supply
+
+### Product gaps:
+1. Curation demand — "best [resources] for [niche]" with no definitive answer
+2. Template/framework demand — "how to [process]" where a reusable template would save time
+3. Data gaps — frequently asked questions requiring research that nobody has packaged
+
+**Output**: A ranked table of candidates:
+
+```
+| # | Opportunity | Type | Signal Source | Trend Direction | Initial Assessment |
+|---|-------------|------|--------------|-----------------|-------------------|
+| 1 | ...         | App  | ...          | ↑↑ Breakout     | Promising         |
+```
+
+## Phase 2: VALIDATE — Confirm the Gap Exists
+
+**Goal**: For top 3 candidates from Phase 1, produce a quantified gap score.
+
+Validation criteria shift by opportunity type:
+
+### For Content gaps:
+
+**Demand Score (1-5)**: Based on search volume and community interest
+**Supply Score (1-5, inverted)**: Based on quality/freshness of existing content
+**Intent Score (1-5)**: From casual browsing (1) to purchase-ready (5)
+
+### For App/Tool gaps:
+
+**Demand Score (1-5)**: Based on how many people are asking for this tool, workaround complexity
+**Supply Score (1-5, inverted)**: Based on existing tools' quality, UX, and pricing gaps
+**Feasibility Score (1-5)**: Can the user realistically build this? (replaces Intent for apps)
+
+### For Service gaps:
+
+**Demand Score (1-5)**: Based on hiring posts, freelancer marketplace activity, complaint volume
+**Supply Score (1-5, inverted)**: Based on existing service providers' quality, speed, pricing
+**Margin Score (1-5)**: Is the willingness-to-pay high enough for sustainable solo delivery?
+
+### For Product gaps:
+
+**Demand Score (1-5)**: Based on how often people search for / ask about this resource
+**Supply Score (1-5, inverted)**: Does a good version of this product already exist?
+**Packaging Score (1-5)**: How easily can scattered information be packaged into a sellable unit?
+
+**Arbitrage Score** = Metric1 × Metric2 × Metric3 (max 125)
+
+Produce a validation report. Read `references/validation-template.md` for the format.
+
+**Thresholds**:
+- Score ≥ 60: Strong GO — prioritize immediately
+- Score 30-59: Conditional — worth pursuing if you have domain expertise
+- Score < 30: PASS — not enough edge
+
+## Phase 3: STRATEGIZE — Design the Play
+
+**Goal**: For each GO topic, create a concrete strategy.
+
+Determine the optimal **format** based on the gap type:
+
+| Gap Type | Best Format | Example |
+|----------|-------------|---------|
+| No comprehensive guide exists | Long-form article/guide | "Complete Guide to X" |
+| Info is scattered across forums | Curated resource list | "Internet Pipes" style |
+| Existing tools are too complex | Simple tool/template | PhotoAI model |
+| No localized version exists | Localized adaptation | JapanDrop-style newsletter |
+| Information changes frequently | Newsletter/subscription | Milk Road model |
+| People need ongoing help | Community or service | Designjoy model |
+
+Strategy output must include:
+1. **Positioning statement**: "For [audience] who [pain point], this is [format] that [unique value]"
+2. **Content angle**: What specific angle differentiates from existing content
+3. **Monetization path**: How this becomes revenue (ads, product, affiliate, service, subscription)
+4. **Competitive moat**: Why copycats can't easily replicate (first-mover, curation quality, trust, community)
+5. **Effort estimate**: Time to create MVP content/product
+6. **Success metrics**: What to measure in the first 30/90 days
+
+## Phase 4: EXECUTE — Create the Action Plan
+
+**Goal**: Produce a concrete, time-boxed execution plan the user can start today.
+
+Output a week-by-week plan:
+
+**Week 1: Foundation**
+- Keyword list (primary + long-tail)
+- Content outline or product spec
+- Distribution channel selection
+
+**Week 2-3: Creation**
+- Content/product creation milestones
+- Draft → Review → Publish pipeline
+
+**Week 4: Launch & Measure**
+- Distribution plan (SEO, social, community seeding)
+- Metrics tracking setup
+- Iteration triggers (when to double down vs pivot)
+
+If the user's project context is known (e.g., JapanDrop, Tuck), tailor the execution plan
+to fit their existing infrastructure and tech stack.
+
+## Usage Modes
+
+**Quick scan**: User says "find opportunities in [niche]" → Classify, then run Phase 1 only
+**Full analysis**: User says "analyze [specific topic/idea]" → Classify, then run Phase 2 + 3
+**End-to-end**: User says "find and plan a new project" → Classify, then run all 4 phases
+**Validation only**: User says "is [topic/app idea] worth pursuing?" → Classify, then Phase 2 only
+**Recon**: User says "what gaps exist in [market]?" → Classify all 4 types, run Phase 1 across types
+
+## Important Principles
+
+1. **Data over intuition** — Always ground recommendations in searchable evidence, not guesses
+2. **Be brutally honest** — If a gap doesn't exist or is closing fast, say so clearly
+3. **Speed matters** — Trend windows close. Bias toward actionable speed over perfect analysis
+4. **Localization is an edge** — For Japanese market content targeting English speakers (or vice versa), cross-language gaps are often the richest arbitrage opportunities
+5. **AI content saturation** — In 2025-2026, generic AI-written content floods mid-tier keywords. Focus on gaps requiring human curation, original research, or unique perspective
+
+## References
+
+- `references/phases.md` — Detailed procedures for each phase
+- `references/validation-template.md` — Gap validation report template
+- `references/examples.md` — Real-world arbitrage case studies for pattern matching

dot_claude/skills/trend-arbitrage/references/examples.md

@@ -0,0 +1,171 @@
+# Trend Arbitrage Case Studies
+
+Reference examples for pattern matching when evaluating opportunities.
+These illustrate different gap types and how they were exploited.
+
+---
+
+## Case 1: Steph Smith — Remote Work Content (2019)
+
+**Gap type:** Freshness + Comprehensiveness
+**Timeline:** Detected 2019, exploded 2020
+
+**Discovery signal:**
+- "Remote work" search trend showing steady upward trajectory for 18+ months
+- No pandemic yet — just a gradual cultural shift
+
+**Supply gap:**
+- Existing content was shallow listicles ("10 Remote Work Tools")
+- No deep content addressing emotional/psychological aspects
+- Related searches revealed pain points: "remote work loneliness", "remote work career growth"
+
+**Strategy:**
+- Created deep, empathy-driven guides addressing the EMOTIONAL side of remote work
+- Positioned as the human voice in a sea of tool lists
+- Compiled insights into paid ebook "Doing Content Right"
+
+**Result:**
+- $130K+ in book sales from personal site alone (no Amazon)
+- Built authority that led to a16z podcast host role
+
+**Pattern to recognize:**
+- Steady upward trend (not spike)
+- Emotional/human angle missing from existing content
+- Related searches reveal deeper unmet needs than surface query
+
+---
+
+## Case 2: Steph Smith — Internet Pipes (2024)
+
+**Gap type:** Curation
+**Timeline:** 2024
+
+**Discovery signal:**
+- AI flood creating "information overload" fatigue
+- Growing demand for trusted, curated information sources
+
+**Supply gap:**
+- Plenty of "news" but no curated "signal vs noise" resource
+- People willing to pay for someone else's filtering work
+
+**Strategy:**
+- Packaged her personal collection of high-quality RSS feeds/URLs
+- Sold as a simple downloadable list
+- Value = curation judgment, not original content
+
+**Result:**
+- 1,400+ copies sold
+
+**Pattern to recognize:**
+- When information overload is the problem, curation IS the product
+- "Simple" products can command premium prices if they save time/effort
+- Personal taste and judgment as a moat
+
+---
+
+## Case 3: Milk Road — Crypto Newsletter (2022)
+
+**Gap type:** Format/Accessibility
+**Timeline:** 10 months to acquisition
+
+**Discovery signal:**
+- Crypto market booming, massive search/interest volume
+- Existing crypto content was dense, jargon-heavy, intimidating
+
+**Supply gap:**
+- NO entertaining, beginner-friendly daily crypto news
+- Gap was in FORMAT, not in topic coverage
+
+**Strategy:**
+- "5-minute daily crypto newsletter you actually enjoy reading"
+- Humor + simplicity in an overly serious market
+- Email format = owned audience (not algorithm dependent)
+
+**Result:**
+- 250K+ subscribers in 10 months
+- Acquired for 8-figure sum
+
+**Pattern to recognize:**
+- The gap can be in FORMAT, not topic
+- Sometimes a saturated topic has a massive accessibility gap
+- Newsletter format enables rapid audience building + high acquisition value
+
+---
+
+## Case 4: Pieter Levels — PhotoAI (2023)
+
+**Gap type:** Tool/Usability
+**Timeline:** Early mover in AI photo generation
+
+**Discovery signal:**
+- Stable Diffusion launched, massive interest in AI image generation
+- Technical barrier to entry was extremely high for average users
+
+**Supply gap:**
+- AI image generation tools required technical setup (Python, command line, GPUs)
+- No "upload and get results" simple tool existed
+
+**Strategy:**
+- Built simplest possible UI: upload photos → get AI-generated professional photos
+- Targeted clear use case: professional headshots, dating profile photos
+- Zero marketing budget — product-led growth via social sharing
+
+**Result:**
+- $132K MRR (Monthly Recurring Revenue)
+- Solo founder, no employees
+
+**Pattern to recognize:**
+- When new technology is powerful but inaccessible → usability wrapper opportunity
+- First to simplify wins, even if the underlying tech is open source
+- Specific use case > general tool
+
+---
+
+## Case 5: Designjoy — Design Subscription (2022)
+
+**Gap type:** Service model (process pain)
+**Timeline:** Grew to $1M ARR as solo founder
+
+**Discovery signal:**
+- "Hire designer" and "design agency" searches showing frustration signals
+- Community complaints about slow agencies, unreliable freelancers, expensive firms
+
+**Supply gap:**
+- Not a content gap — a SERVICE DELIVERY gap
+- Existing options: expensive agencies (slow), freelancers (unreliable), cheap marketplaces (low quality)
+- Nobody offered: fixed price + unlimited requests + no meetings + fast delivery
+
+**Strategy:**
+- Monthly subscription for unlimited design requests
+- Asynchronous communication only (no meetings)
+- Solo operator keeping overhead at zero
+
+**Result:**
+- $130K monthly revenue
+- Zero employees
+
+**Pattern to recognize:**
+- Arbitrage can apply to services, not just content
+- "Process pain" (meetings, hiring, managing) can be bigger than "quality pain"
+- Subscription model with async delivery = highly scalable for solo operators
+
+---
+
+## Common Patterns Across All Cases
+
+1. **None of them competed on "quality" in a crowded space.** They found empty spaces.
+2. **All of them validated demand BEFORE creating.** Data first, creation second.
+3. **Speed of execution mattered more than perfection.** First mover advantage was critical.
+4. **The "product" was often surprisingly simple.** URL lists, wrapper UIs, curated newsletters.
+5. **Moats came from positioning and trust, not technical complexity.**
+6. **Zero or minimal marketing spend.** The product-market fit did the marketing.
+
+---
+
+## Anti-Patterns (What NOT to Do)
+
+1. **"I'll write better content about [saturated topic]"** — Quality alone rarely wins
+2. **"Let me spend 6 months perfecting this"** — The window will close
+3. **"AI can write this for me"** — If AI can easily write it, the gap will fill fast
+4. **"This topic is popular so it must be good"** — Popular ≠ opportunity (check supply!)
+5. **"Nobody has done this so nobody wants it"** — Could be true; validate demand first