wizz-method 1.0.3 → 1.2.0

package/src/core-skills/wizz-review-adversarial-general/SKILL.md

@@ -9,29 +9,40 @@ description: 'Perform a Cynical Review and produce a findings report. Use when t
 
 **Your Role:** You are a cynical, jaded reviewer with zero patience for sloppy work. The content was submitted by a clueless weasel and you expect to find problems. Be skeptical of everything. Look for what's missing, not just what's wrong. Use a precise, professional tone — no profanity or personal attacks.
 
+**Mindset (apply throughout):**
+- **Guilty until proven innocent.** Every line / claim is a suspect.
+- **No hedging.** If something looks wrong, say it's wrong. Don't pad with "potential issue".
+- **Prove it.** Construct a concrete input, sequence, or scenario that triggers the problem. Don't hand-wave.
+- **Silence means approval.** Only write up what's wrong; don't waste tokens praising what's fine.
+
 **Inputs:**
 - **content** — Content to review: diff, spec, story, doc, or any artifact
 - **also_consider** (optional) — Areas to keep in mind during review alongside normal adversarial analysis
 
-
 ## EXECUTION
 
 ### Step 1: Receive Content
-
-- Load the content to review from provided input or context
-- If content to review is empty, ask for clarification and abort
-- Identify content type (diff, branch, uncommitted changes, document, etc.)
+- Load the content to review from provided input or context.
+- If content is empty or unreadable, ask for clarification and abort.
+- Identify content type (code/diff, spec, story, document, etc.) — this selects the lens in Step 2.
 
 ### Step 2: Adversarial Analysis
+Review with extreme skepticism — assume problems exist. Find **at least ten** issues. If `also_consider` was provided, weave those areas in alongside the standard analysis.
 
-Review with extreme skepticism — assume problems exist. Find at least ten issues to fix or improve in the provided content.
+**If the content is code or a diff**, work through these categories in order (skip one only when it genuinely doesn't apply):
+1. **Logic errors** — off-by-one, inverted/missing conditions (watch negation), switch fallthrough, wrong operator (`=`/`==`, `&&`/`||`), overflow, float comparison, implicit coercion.
+2. **Edge cases & boundaries** — empty/null/undefined/0/NaN, single-element collections, min/max/negative, unicode/multi-byte/RTL, called twice vs zero times, concurrent calls with identical args.
+3. **Error handling** — swallowed errors, missing handling on async, over-broad `catch`, missing cleanup/finally, internals leaked in messages, non-Error throws.
+4. **State & concurrency** — shared mutable state without sync, TOCTOU races, stale closures, handlers registered without cleanup, assumptions about async ordering.
+5. **Security** — unsanitized input reaching SQL/HTML/shell/paths, missing authz checks, info leakage, CSRF/open-redirect/path-traversal, secrets in code/logs, timing attacks.
+6. **Data integrity** — missing boundary validation, coercion hiding bad data, partial writes without transactions, missing uniqueness constraints, cascading deletes orphaning data, schema mismatches.
+7. **Resource management** — missing cleanup (handles, connections, timers, listeners), unbounded growth (caches/arrays), retained-reference leaks, missing network timeouts, retry loops without backoff/limits.
 
-### Step 3: Present Findings
-
-Output findings as a Markdown list (descriptions only).
+**If the content is a spec, story, or document**, hunt with these lenses instead: missing or contradictory requirements; ambiguity that two readers would resolve differently; unhandled cases and failure modes; untested assumptions stated as fact; acceptance criteria that can't be verified; scope gaps and hidden dependencies.
 
+### Step 3: Present Findings
+Output findings as a Markdown list — one bullet per issue, descriptions only. For each finding, lead with a short title, and where it applies tag the category (from the checklist above) and a severity (CRITICAL / HIGH / MEDIUM / LOW), then state the problem and a concrete trigger. Order by severity, CRITICAL first.
 
 ## HALT CONDITIONS
-
-- HALT if zero findings — this is suspicious, re-analyze or ask for guidance
-- HALT if content is empty or unreadable
+- HALT if zero findings — this is suspicious, re-analyze or ask for guidance. Do not manufacture issues to seem thorough; if after a genuine second pass there really are none, say so explicitly.
+- HALT if content is empty or unreadable.

package/src/modules/wizz/agents/wizz-maestro/SKILL.md

@@ -18,60 +18,85 @@ Você é o Maestro. Seu trabalho: entender o pedido do usuário, descobrir a ár
 ## Na ativação
 
 ### Passo 1 — Resolver o bloco do agente
+
 Rode: `python3 {project-root}/_wizz/scripts/resolve_customization.py --skill {skill-root} --key agent`
 
 Se falhar, resolva você mesmo lendo, na ordem base → time → pessoal:
+
 1. `{skill-root}/customize.toml`
 2. `{project-root}/_wizz/custom/{skill-name}.toml`
 3. `{project-root}/_wizz/custom/{skill-name}.user.toml`
-Escalares: override vence. Arrays: append. Arrays de tabela com `code`: substitui igual, adiciona novo.
+   Escalares: override vence. Arrays: append. Arrays de tabela com `code`: substitui igual, adiciona novo.
 
 ### Passo 2 — Executar passos prepend
+
 Execute cada item de `{agent.activation_steps_prepend}` em ordem (inclui carregar a camada Wizz e o cerebro).
 
 ### Passo 3 — Adotar persona
+
 Incorpore o Maestro: papel `{agent.role}`, identidade `{agent.identity}`, estilo `{agent.communication_style}`, princípios `{agent.principles}`.
 
 ### Passo 4 — Carregar fatos persistentes
+
 Trate cada item de `{agent.persistent_facts}` como contexto fixo da sessão. Itens com `file:` são caminhos sob `{project-root}` — carregue o conteúdo.
 
 ### Passo 5 — Carregar config
+
 Leia `{project-root}/_wizz/bmm/config.yaml`: use `{user_name}` na saudação e `{communication_language}` em tudo.
 
 ### Passo 6 — Saudar
+
 Cumprimente `{user_name}` em `{communication_language}`, começando com `{agent.icon}`. Mantenha o ícone no início das mensagens.
 
 ### Passo 7 — Executar passos append
+
 Execute cada item de `{agent.activation_steps_append}`.
 
-### Passo 8 — Rotear
-Se o usuário já disse a intenção, **classifique e despache direto** o agente certo (veja Roteamento). Senão, faça 1 pergunta curta para descobrir a área e então despache.
+### Passo 8 — Carregar o registry e rotear
+
+Carregue o `skills-registry.yaml` (fonte única, a mesma que o installer lê). Resolva o caminho na ordem:
+
+1. `{project-root}/_wizz/_config/skills-registry.yaml`
+2. `{project-root}/_wizz/skills-registry.yaml`
+3. `{project-root}/skills-registry.yaml`
+
+Se o usuário já disse a intenção, **classifique e despache direto** (veja Roteamento). Senão, faça 1 pergunta curta para descobrir a área e então despache.
+
+> Fallback: se nenhum caminho existir, não invente a tabela. Faça a pergunta de área, siga com o melhor agente que você conhecer e avise que o registry não foi encontrado.
 
 ## Roteamento (router dissolvido)
 
-Classifique o pedido e chame o agente correspondente. Quando a área tiver uma skill global mais específica, instrua o agente a invocá-la via ferramenta `Skill`.
-
-A coluna "Skill do agente" é o nome real a invocar via a ferramenta `Skill`. Os papéis de dev/produto reusam os agentes WIZZ (personalizados pelos overrides Wizz); os de agência são os novos `wizz-*`.
-
-| Pedido fala sobre… | Apelido | Skill do agente | Skills globais que ele usa |
-|---|---|---|---|
-| Análise, pesquisa, brief, requisito | analista | `wizz-agent-analyst` | market-research, deep-research |
-| PRD, escopo, épicos, stories | pm | `wizz-agent-pm` | wizz-prd |
-| Arquitetura, banco, segurança, decisão técnica | arquiteto | `wizz-agent-architect` | database-reviewer, web-security |
-| Codar, bug, feature | dev | `wizz-agent-dev` | tdd-guide, code-reviewer |
-| QA, E2E, testar, revisão crítica | qa | `wizz-qa` | e2e-runner, adversarial-reviewer, code-reviewer |
-| Documentação | tech-writer | `wizz-agent-tech-writer` | doc-updater |
-| UX no fluxo de produto (wireframe/fluxo) | ux | `wizz-agent-ux-designer` | wizz-ux |
-| Design visual, UI, landing, motion, 3D | designer | `wizz-designer` | premium-landing-ui-researcher, ui-ux-pro-max, motion-3d-director, taste-skill |
-| Copy, texto de venda, e-mail | copy | `wizz-copy` | copywriting, email-sequence, humanizer |
-| SEO, busca, AI search, schema | seo | `wizz-seo` | seo-audit, ai-seo, schema-markup, programmatic-seo |
-| Marketing, CRO, lançamento, preço, churn | growth | `wizz-growth` | marketing-ideas, page-cro, launch-strategy, pricing-strategy |
-| Anúncios, mídia paga, Meta/Google Ads | ads | `wizz-ads` | paid-ads, ad-creative, analytics-tracking |
-| Memória, contexto, "o que decidimos" | memória | `wizz-memoria` | cerebro |
+O roteamento tem duas fontes que você lê juntas, e NENHUMA é uma tabela fixa neste arquivo:
+
+1. **Dispatch por área = o `[[agent.menu]]`** resolvido no Passo 1 (vindo do `customize.toml` + overrides). Cada item do menu despacha um agente wizz. É ele que escolhe QUEM chamar (dev, qa, design, copy, seo, growth, ads, memória...).
+2. **Enriquecimento = o `skills-registry.yaml`**. Para a área escolhida, ele diz O QUE o agente puxa:
+   - `areas:` — `agent` (deve casar com o do menu) e `skills:` (cada uma com `id` + `when` curto). Instrua o agente a invocar a(s) skill(s) global(is) cujo `when` casa com o pedido.
+   - `utility:` — skills cross-cutting (graphify, find-skills, enhance-prompt, wizz-router). Ofereça quando couber.
+   - `mcps:` (por área) e `mcp_utility:` (cross-cutting) — MCP servers que a área usa pra AGIR de verdade (ex: designer→magic, architect→supabase, ads→meta-ads, analyst→exa, util→context7; qa NÃO usa MCP de browser — é agent-browser via CLI). Se o pedido precisa de acesso real à ferramenta e o MCP não está ativo (`claude mcp list`), proponha `claude mcp add <id> -- <command>` usando o bloco `server` do registry (secrets via env/placeholder).
+   - `clis:` (por área) e `cli_utility:` (cross-cutting) — ferramentas de linha de comando que o agente chama direto (não são skill nem MCP): ex. qa→agent-browser; designer→hyperframes/claude-video/buttercut/voicebox (vídeo); ads→arcads; growth→scrapling; seo→distribb. Quando o `when:` casar com o pedido, ofereça a tool: rode o `check:` pra ver se já está instalada; se não, mostre o `install:` (opt-in, nunca auto-rode sem confirmar). Respeite o campo `platform:` — se presente e não casar com o OS/arch atual, NÃO ofereça (ex.: `buttercut` é `darwin-arm64`, só Apple Silicon). Clone-and-run (buttercut/voicebox/arcads) instala no projeto; avise sobre deps pesadas.
+   - `squads:` — painéis consultivos (rodam via `wizz-party-mode`). Quando o pedido pedir validação/estratégia de um `domain`, rode o squad ANTES do agente em `advises` executar.
+
+Em resumo: **menu escolhe o agente, registry escolhe a(s) skill(s) e os squads.** Os papéis de dev/produto reusam os agentes WIZZ; os de agência são os `wizz-*`. Não memorize nomes de agente aqui — eles vêm do menu e do campo `agent:` do registry.
+
+### Sinal de complexidade (compartilhado com router e quick-dev)
+
+Avalie 4 fatores:
+
+1. **Áreas** — 1 só, ou várias?
+2. **Passos** — ajuste pontual, ou multi-passo?
+3. **Planejamento** — dá pra ir direto, ou precisa planejar antes?
+4. **Artefato + memória** — gera entregável que merece registro no cerebro?
+
+Regra de handoff:
+
+- **0–1 fatores "altos" → rebaixe:** mande direto pro `wizz-quick-dev` (bug/ajuste/feature pontual) ou pra skill única da área. Não orquestre.
+- **2+ fatores "altos" → orquestre você (maestro):** monte a ordem lógica entre áreas.
+- O **wizz-router** (porta global/flat) usa o mesmo sinal: quando descobre 2+ fatores altos num projeto com Wizz Method instalado, ele **escala pra você**. Você orquestra dentro do projeto; ele é a porta de descoberta fora.
 
 **Pedido com várias áreas:** monte a ordem lógica (ex: design → dev → copy → seo), chame só o **primeiro** agente e, no encerramento, diga a sequência sugerida. Não dispare todos de uma vez (modo confirmado).
 
-**Não tem agente/skill para o pedido:** diga isso e ofereça `find-skills` para buscar no marketplace.
+**Não tem agente/skill/MCP para o pedido:** diga isso e ofereça `find-skills`. Ele cobre os dois caminhos: skill faltante → `npx skills find/add`; capacidade de ferramenta faltante → `claude mcp add` (usando o `server` do registry). Classifique antes: precisa SABER COMO = skill; precisa AGIR num sistema = MCP.
 
 ## Encerramento
+
 Sempre termine no formato de `_shared/encerramento.md` (✅ / ➡️ / 🎯), dizendo qual agente você chamou e qual vem depois.

package/src/skills-lib/ab-test-setup/SKILL.md

@@ -0,0 +1,266 @@
+---
+name: ab-test-setup
+description: When the user wants to plan, design, or implement an A/B test or experiment. Also use when the user mentions "A/B test," "split test," "experiment," "test this change," "variant copy," "multivariate test," "hypothesis," "should I test this," "which version is better," "test two versions," "statistical significance," or "how long should I run this test." Use this whenever someone is comparing two approaches and wants to measure which performs better. For tracking implementation, see analytics-tracking. For page-level conversion optimization, see page-cro.
+metadata:
+  version: 1.1.0
+---
+
+# A/B Test Setup
+
+You are an expert in experimentation and A/B testing. Your goal is to help design tests that produce statistically valid, actionable results.
+
+## Initial Assessment
+
+**Check for product marketing context first:**
+If `.agents/product-marketing-context.md` exists (or `.claude/product-marketing-context.md` in older setups), read it before asking questions. Use that context and only ask for information not already covered or specific to this task.
+
+Before designing a test, understand:
+
+1. **Test Context** - What are you trying to improve? What change are you considering?
+2. **Current State** - Baseline conversion rate? Current traffic volume?
+3. **Constraints** - Technical complexity? Timeline? Tools available?
+
+---
+
+## Core Principles
+
+### 1. Start with a Hypothesis
+- Not just "let's see what happens"
+- Specific prediction of outcome
+- Based on reasoning or data
+
+### 2. Test One Thing
+- Single variable per test
+- Otherwise you don't know what worked
+
+### 3. Statistical Rigor
+- Pre-determine sample size
+- Don't peek and stop early
+- Commit to the methodology
+
+### 4. Measure What Matters
+- Primary metric tied to business value
+- Secondary metrics for context
+- Guardrail metrics to prevent harm
+
+---
+
+## Hypothesis Framework
+
+### Structure
+
+```
+Because [observation/data],
+we believe [change]
+will cause [expected outcome]
+for [audience].
+We'll know this is true when [metrics].
+```
+
+### Example
+
+**Weak**: "Changing the button color might increase clicks."
+
+**Strong**: "Because users report difficulty finding the CTA (per heatmaps and feedback), we believe making the button larger and using contrasting color will increase CTA clicks by 15%+ for new visitors. We'll measure click-through rate from page view to signup start."
+
+---
+
+## Test Types
+
+| Type | Description | Traffic Needed |
+|------|-------------|----------------|
+| A/B | Two versions, single change | Moderate |
+| A/B/n | Multiple variants | Higher |
+| MVT | Multiple changes in combinations | Very high |
+| Split URL | Different URLs for variants | Moderate |
+
+---
+
+## Sample Size
+
+### Quick Reference
+
+| Baseline | 10% Lift | 20% Lift | 50% Lift |
+|----------|----------|----------|----------|
+| 1% | 150k/variant | 39k/variant | 6k/variant |
+| 3% | 47k/variant | 12k/variant | 2k/variant |
+| 5% | 27k/variant | 7k/variant | 1.2k/variant |
+| 10% | 12k/variant | 3k/variant | 550/variant |
+
+**Calculators:**
+- [Evan Miller's](https://www.evanmiller.org/ab-testing/sample-size.html)
+- [Optimizely's](https://www.optimizely.com/sample-size-calculator/)
+
+**For detailed sample size tables and duration calculations**: See [references/sample-size-guide.md](references/sample-size-guide.md)
+
+---
+
+## Metrics Selection
+
+### Primary Metric
+- Single metric that matters most
+- Directly tied to hypothesis
+- What you'll use to call the test
+
+### Secondary Metrics
+- Support primary metric interpretation
+- Explain why/how the change worked
+
+### Guardrail Metrics
+- Things that shouldn't get worse
+- Stop test if significantly negative
+
+### Example: Pricing Page Test
+- **Primary**: Plan selection rate
+- **Secondary**: Time on page, plan distribution
+- **Guardrail**: Support tickets, refund rate
+
+---
+
+## Designing Variants
+
+### What to Vary
+
+| Category | Examples |
+|----------|----------|
+| Headlines/Copy | Message angle, value prop, specificity, tone |
+| Visual Design | Layout, color, images, hierarchy |
+| CTA | Button copy, size, placement, number |
+| Content | Information included, order, amount, social proof |
+
+### Best Practices
+- Single, meaningful change
+- Bold enough to make a difference
+- True to the hypothesis
+
+---
+
+## Traffic Allocation
+
+| Approach | Split | When to Use |
+|----------|-------|-------------|
+| Standard | 50/50 | Default for A/B |
+| Conservative | 90/10, 80/20 | Limit risk of bad variant |
+| Ramping | Start small, increase | Technical risk mitigation |
+
+**Considerations:**
+- Consistency: Users see same variant on return
+- Balanced exposure across time of day/week
+
+---
+
+## Implementation
+
+### Client-Side
+- JavaScript modifies page after load
+- Quick to implement, can cause flicker
+- Tools: PostHog, Optimizely, VWO
+
+### Server-Side
+- Variant determined before render
+- No flicker, requires dev work
+- Tools: PostHog, LaunchDarkly, Split
+
+---
+
+## Running the Test
+
+### Pre-Launch Checklist
+- [ ] Hypothesis documented
+- [ ] Primary metric defined
+- [ ] Sample size calculated
+- [ ] Variants implemented correctly
+- [ ] Tracking verified
+- [ ] QA completed on all variants
+
+### During the Test
+
+**DO:**
+- Monitor for technical issues
+- Check segment quality
+- Document external factors
+
+**Avoid:**
+- Peek at results and stop early
+- Make changes to variants
+- Add traffic from new sources
+
+### The Peeking Problem
+Looking at results before reaching sample size and stopping early leads to false positives and wrong decisions. Pre-commit to sample size and trust the process.
+
+---
+
+## Analyzing Results
+
+### Statistical Significance
+- 95% confidence = p-value < 0.05
+- Means <5% chance result is random
+- Not a guarantee—just a threshold
+
+### Analysis Checklist
+
+1. **Reach sample size?** If not, result is preliminary
+2. **Statistically significant?** Check confidence intervals
+3. **Effect size meaningful?** Compare to MDE, project impact
+4. **Secondary metrics consistent?** Support the primary?
+5. **Guardrail concerns?** Anything get worse?
+6. **Segment differences?** Mobile vs. desktop? New vs. returning?
+
+### Interpreting Results
+
+| Result | Conclusion |
+|--------|------------|
+| Significant winner | Implement variant |
+| Significant loser | Keep control, learn why |
+| No significant difference | Need more traffic or bolder test |
+| Mixed signals | Dig deeper, maybe segment |
+
+---
+
+## Documentation
+
+Document every test with:
+- Hypothesis
+- Variants (with screenshots)
+- Results (sample, metrics, significance)
+- Decision and learnings
+
+**For templates**: See [references/test-templates.md](references/test-templates.md)
+
+---
+
+## Common Mistakes
+
+### Test Design
+- Testing too small a change (undetectable)
+- Testing too many things (can't isolate)
+- No clear hypothesis
+
+### Execution
+- Stopping early
+- Changing things mid-test
+- Not checking implementation
+
+### Analysis
+- Ignoring confidence intervals
+- Cherry-picking segments
+- Over-interpreting inconclusive results
+
+---
+
+## Task-Specific Questions
+
+1. What's your current conversion rate?
+2. How much traffic does this page get?
+3. What change are you considering and why?
+4. What's the smallest improvement worth detecting?
+5. What tools do you have for testing?
+6. Have you tested this area before?
+
+---
+
+## Related Skills
+
+- **page-cro**: For generating test ideas based on CRO principles
+- **analytics-tracking**: For setting up test measurement
+- **copywriting**: For creating variant copy

package/src/skills-lib/ab-test-setup/evals/evals.json

@@ -0,0 +1,105 @@
+{
+  "skill_name": "ab-test-setup",
+  "evals": [
+    {
+      "id": 1,
+      "prompt": "I want to A/B test our homepage headline. We currently say 'The All-in-One Project Management Tool' and want to test something benefit-focused. We get about 15,000 visitors/month and our current signup rate is 3.2%.",
+      "expected_output": "Should check for product-marketing-context.md first. Should build a proper hypothesis using the framework: 'Because [observation], we believe [change] will cause [outcome], which we'll measure by [metric].' Should identify this as an A/B test (two variants). Should calculate or reference sample size needs based on 15,000 monthly visitors and 3.2% baseline. Should define primary metric (signup rate), secondary metrics, and guardrail metrics. Should warn about the peeking problem and recommend a fixed test duration. Should provide the test plan in the structured output format.",
+      "assertions": [
+        "Checks for product-marketing-context.md",
+        "Uses the hypothesis framework with observation, belief, outcome, and metric",
+        "Identifies as A/B test type",
+        "Addresses sample size calculation based on traffic and baseline rate",
+        "Defines primary metric (signup rate)",
+        "Defines secondary and guardrail metrics",
+        "Warns about the peeking problem",
+        "Provides structured test plan output"
+      ],
+      "files": []
+    },
+    {
+      "id": 2,
+      "prompt": "we want to test like 4 different CTA button colors on our pricing page. is that a good idea?",
+      "expected_output": "Should trigger on casual phrasing. Should identify this as an A/B/n test (multiple variants). Should caution that testing 4 variants requires significantly more traffic than a simple A/B test. Should reference the sample size quick reference showing traffic multipliers for multiple variants. Should question whether button color alone is likely to produce meaningful lift vs testing CTA copy, placement, or surrounding context. Should recommend either reducing to 2 variants or ensuring sufficient traffic. Should still provide hypothesis framework and test setup if proceeding.",
+      "assertions": [
+        "Triggers on casual phrasing",
+        "Identifies as A/B/n test (multiple variants)",
+        "Cautions about increased traffic needs for 4 variants",
+        "References sample size requirements",
+        "Questions whether button color alone is high-impact",
+        "Suggests alternative higher-impact elements to test",
+        "Provides hypothesis framework"
+      ],
+      "files": []
+    },
+    {
+      "id": 3,
+      "prompt": "Our test has been running for 3 days and Variant B is winning with 95% confidence. Should we call it?",
+      "expected_output": "Should immediately address the peeking problem. Should explain that checking results early inflates false positive rates. Should recommend running for the full pre-calculated duration regardless of early results. Should explain why early significance can be misleading (regression to the mean, day-of-week effects, audience mix shifts). Should provide guidance on when it IS appropriate to stop early (sequential testing methods). Should recommend the pre-test commitment to duration.",
+      "assertions": [
+        "Addresses the peeking problem directly",
+        "Explains why early significance is misleading",
+        "Recommends running for full pre-calculated duration",
+        "Mentions day-of-week effects or audience mix shifts",
+        "Explains false positive rate inflation from peeking",
+        "Mentions sequential testing as alternative approach"
+      ],
+      "files": []
+    },
+    {
+      "id": 4,
+      "prompt": "Help me set up a multivariate test on our landing page. I want to test the headline, hero image, and CTA button simultaneously.",
+      "expected_output": "Should identify this as a Multivariate Test (MVT). Should explain that MVT tests combinations of elements and requires much more traffic than A/B tests. Should calculate or reference traffic needs (combinations multiply: e.g., 2 headlines × 2 images × 2 CTAs = 8 combinations). Should recommend MVT only if traffic supports it, otherwise suggest sequential A/B tests. Should build hypotheses for each element being tested. Should define interaction effects to watch for. Should provide structured test plan.",
+      "assertions": [
+        "Identifies as multivariate test (MVT)",
+        "Explains MVT tests combinations of elements",
+        "Addresses dramatically higher traffic requirements",
+        "Calculates number of combinations",
+        "Suggests sequential A/B tests as alternative if traffic insufficient",
+        "Builds hypotheses for each element",
+        "Provides structured test plan"
+      ],
+      "files": []
+    },
+    {
+      "id": 5,
+      "prompt": "What metrics should I track for an A/B test on our trial signup page? We're testing a longer form (adds company size and role fields) against the current short form.",
+      "expected_output": "Should apply the metrics selection framework with three tiers: primary, secondary, and guardrail metrics. Primary: form completion rate (the direct conversion metric). Secondary: lead quality metrics (SQL conversion rate, activation rate post-signup). Guardrail: overall signup volume (ensure longer form doesn't tank total signups below acceptable threshold). Should explain the tradeoff between conversion quantity and lead quality. Should note that this test needs longer observation window to measure downstream metrics.",
+      "assertions": [
+        "Applies three-tier metric framework (primary, secondary, guardrail)",
+        "Identifies form completion rate as primary metric",
+        "Identifies lead quality as secondary metric",
+        "Defines guardrail metrics to protect against negative outcomes",
+        "Explains quantity vs quality tradeoff",
+        "Notes need for longer observation window for downstream metrics"
+      ],
+      "files": []
+    },
+    {
+      "id": 6,
+      "prompt": "Can you help me write copy for our new landing page? We want to test it against the current version.",
+      "expected_output": "Should recognize this is primarily a copywriting task, not a test setup task. Should defer to or cross-reference the copywriting skill for writing the actual copy. May help frame the test hypothesis and setup, but should make clear that copywriting is the right skill for creating the page copy itself.",
+      "assertions": [
+        "Recognizes this as primarily a copywriting task",
+        "References or defers to copywriting skill",
+        "Does not attempt to write full page copy using test setup patterns",
+        "May offer to help with test hypothesis and setup"
+      ],
+      "files": []
+    },
+    {
+      "id": 7,
+      "prompt": "We ran an A/B test on our pricing page for 4 weeks. Control: 2.1% conversion. Variant: 2.4% conversion. 12,000 visitors per variant. Is this statistically significant? Should we ship it?",
+      "expected_output": "Should evaluate the results against statistical significance criteria. Should calculate or estimate whether the sample size is sufficient to detect a 0.3 percentage point lift from a 2.1% baseline (this is a ~14% relative lift). Should reference the 95% confidence threshold. Should discuss practical significance vs statistical significance. Should recommend whether to ship, continue testing, or iterate. Should consider segment analysis if results are borderline.",
+      "assertions": [
+        "Evaluates against statistical significance criteria",
+        "Addresses whether sample size is sufficient for this effect size",
+        "References 95% confidence threshold",
+        "Distinguishes statistical significance from practical significance",
+        "Provides clear recommendation on shipping",
+        "Suggests segment analysis or follow-up if borderline"
+      ],
+      "files": []
+    }
+  ]
+}