feat: add SEO/GEO evidence workbook

docs/marketing/research/2026-06-03-seo-geo-evidence-workbook.md

@@ -0,0 +1,91 @@
+# 2026-06-03 SEO/GEO Evidence Workbook Research Note
+
+## Summary
+
+The SEO/GEO evidence workbook turns QuickFork's registered Search/AI baseline prompts into a concrete audit queue. Each baseline row expands into one Google Search Console task and five AI-answer audit tasks across ChatGPT Search, Perplexity, Google AI Overview, Gemini, and Claude.
+
+This is a measurement slice, not SEO/GEO validation.
+
+## Why This Slice Matters
+
+- QuickFork now has multiple published product, use-case, resource, tool, and example routes, but the active experiments still have `pending_evidence`.
+- Search Console exports and AI-answer audits are manual enough that evidence can drift unless the exact route, query, surface, required fields, and forbidden claims are fixed.
+- A supervisor-grade workbook lets growth reviewers audit the same experiment contract across GA4, Search Console, and answer-engine surfaces before choosing the next landing page or product slice.
+- This supports the current P0 priority: collect evidence before scaling more pages.
+
+## Target User
+
+- SEO/GEO supervisor reviewing whether QuickFork is indexable, extractable, and accurately summarized.
+- Growth operator collecting route-level Search Console and AI-answer evidence.
+- Product marketer deciding whether a page promise deserves more content, product, or paid-package investment.
+
+## Lifecycle Position
+
+- Discovery: query clusters and AI prompts are fixed by the Search/AI baseline contract.
+- Validation: workbook rows specify the evidence needed before route comparisons can be decided.
+- Activation: workbook decisions must be read alongside generation, artifact export, and launch materials map copy behavior.
+- Monetization: paid package decisions still require lead quality and willingness-to-pay evidence.
+
+## Growth Contract
+
+Primary metric:
+
+- Completion of Search Console and AI-answer audit rows for each active experiment that requires search or AI evidence.
+
+Supporting metrics:
+
+- `page_view`, `cta_clicked`, `generation_started`, `generation_completed`, and `generation_failed` over the same 14-day window.
+- `launch_materials_map_copied` for `2026_q2_launch_materials_map_intent_validation`.
+- Lead-quality fields for full launch package requests when bottom-funnel pages are involved.
+
+Guardrails:
+
+- Do not treat workbook rows, prompt coverage, or page publication as search demand.
+- Do not claim ranking wins, AI citations, revenue, customer count, conversion lift, Product Hunt results, guaranteed launches, or pricing validation.
+- Keep the workbook PII-free: no raw emails, raw repo content, source notes, private launch notes, tokens, secrets, raw artifact bodies, or customer proof.
+
+## Implemented Surface
+
+- `src/marketing/seo-geo-audit-workbook.ts`
+- `src/marketing/seo-geo-audit-workbook.test.ts`
+- `docs/marketing/data/seo-geo-audit-workbook.csv`
+- `docs/superpowers/plans/2026-06-03-seo-geo-evidence-workbook.md`
+
+## Workbook Contract
+
+Every Search/AI baseline row now produces six audit rows:
+
+- `google_search_console`: collect `window_start`, `window_end`, `query`, `page`, `country`, `device`, `impressions`, `clicks`, `ctr`, and `average_position`.
+- `chatgpt_search`, `perplexity`, `google_ai_overview`, `gemini`, and `claude`: collect `audit_date`, `mentioned`, `cited`, `source_url`, `answer_summary`, `accuracy_status`, `competitors_cited`, and `forbidden_claims_observed`.
+
+The workbook keeps each row tied to:
+
+- Baseline id.
+- Experiment id.
+- Control or variant role.
+- Canonical path.
+- Target user.
+- Query cluster.
+- Surface.
+- Expected terms.
+- Forbidden claims.
+- Pending status and next action.
+
+## Current Verification
+
+- RED test failed first because `src/marketing/seo-geo-audit-workbook.ts` did not exist.
+- GREEN focused verification passed after adding the module and CSV mirror: `npm test -- src/marketing/seo-geo-audit-workbook.test.ts` returned 1 file passed and 5 tests passed.
+- Related verification passed: `npm test -- src/marketing/seo-geo-audit-workbook.test.ts src/marketing/search-ai-baseline.test.ts src/marketing/growth-experiment-report.test.ts` returned 3 files passed and 21 tests passed.
+- Initial full `npm test` attempts with the default forks pool hit Vitest worker startup timeouts in this local environment.
+- Targeted reruns of the affected files passed, including `npm test -- src/App.test.tsx -t "keeps the generator studio|full launch package contact|landing architecture"` and `npm test -- src/server/marketing/launch-package-fit.test.ts`.
+- Full verification passed with threads pool: `npm test -- --testTimeout=30000 --pool=threads` returned 25 files passed and 168 tests passed.
+- Build verification passed: `npm run build` completed TypeScript build and Vite production build.
+- Diff hygiene passed: `git diff --check` returned no whitespace errors.
+
+## Decision
+
+Treat the workbook as P0 measurement infrastructure. It makes SEO/GEO evidence collection executable, but it does not prove indexed demand, answer-engine citation, activation quality, lead quality, or willingness to pay.
+
+## Next Validation Step
+
+Use the workbook to collect 14-day Search Console and AI-answer evidence for the active validation experiments, then fill `docs/marketing/data/growth-experiment-evidence.csv` and decide whether each page should be promoted, rewritten, consolidated, or paused.

docs/marketing/research/2026-06-03-sites-output-enablement.md

@@ -0,0 +1,45 @@
+# Sites Output Enablement
+
+Date: 2026-06-03
+
+## Hypothesis
+
+If QuickFork exposes Sites as part of the launch package, founders, open-source maintainers, and DevRel operators can turn the same repository evidence into a reviewable site/page spine instead of treating launch output as only README, deck, social, visual, and outreach files.
+
+## Lifecycle Stage
+
+- Stage: Activation to monetization.
+- Target user: AI/devtool builders, open-source maintainers, technical founders, and DevRel operators preparing a public launch.
+- Product surface: generated launch brief, studio output presets, and full launch package qualification form.
+- CTA: generate a similar launch package or request a reviewed full launch package.
+- Primary metric: `launch_artifact_copied` or `launch_artifact_downloaded` where `artifact_type=site`; `sales_contact_requested` with package scope including `sites`.
+- Guardrail metric: no raw README text, artifact body, email, token, secret, unsupported customer proof, ranking, revenue, or guaranteed-growth claim in browser analytics.
+
+## Change
+
+- Added `site` as a generated launch materials channel.
+- Added a source-backed "Launch site page spine" export artifact.
+- Added a Sites output preset for the studio.
+- Added Sites to full launch package qualification scope.
+- Kept Sites as a reviewable page spine, not automatic public publishing.
+
+## Evidence Observed
+
+- RED generation test failed first because the launch brief did not include the site channel.
+- RED App tests failed first because the studio and contact scope form did not expose Sites.
+- Focused GREEN verification passed after implementation:
+  - `npm test -- src/server/generation/generation.test.ts -t "source-backed free repo launch brief"`
+  - `npm test -- src/App.test.tsx -t "keeps the generator studio|full launch package contact"`
+  - `npm test -- src/server/marketing/launch-package-fit.test.ts src/server/marketing/lead-capture.test.ts -t "launch package fit|full launch package qualification"`
+- Related focused verification passed after the final scope-count alignment:
+  - `npm test -- src/server/generation/generation.test.ts -t "source-backed free repo launch brief"` returned 1 file passed and 1 selected test passed.
+  - `npm test -- src/App.test.tsx -t "keeps the generator studio|full launch package contact|landing architecture"` returned 1 file passed and 3 selected tests passed.
+  - `npm test -- src/server/marketing/launch-package-fit.test.ts src/server/marketing/lead-capture.test.ts -t "launch package fit|full launch package qualification"` returned 2 files passed and 2 selected tests passed.
+- Full verification with the default forks pool hit Vitest worker startup timeout in this local environment, but no Sites assertion failed after the scope-count fix.
+- Full verification passed with threads pool: `npm test -- --testTimeout=30000 --pool=threads` returned 25 files passed and 168 tests passed.
+- Build verification passed: `npm run build` completed TypeScript build and Vite production build.
+- Diff hygiene passed: `git diff --check` returned no whitespace errors.
+
+## Remaining Validation Gap
+
+No production users have copied, downloaded, or requested the Sites scope yet. Treat the output as a monetization hypothesis until artifact interactions, full-package requests, or interviews show that launch site/page support is more valuable than the existing README, social, deck, visual, and outreach outputs.

docs/plans/2026-06-01-quickfork-growth-lifecycle-prioritization.md

@@ -1646,3 +1646,113 @@ Decision:
 Next action:
 
 - Run full verification, merge, production smoke, then use `2026_q2_launch_materials_map_intent_validation` for 14 days of GA4, Search Console, and AI-answer evidence collection.
+
+## 2026-06-03 Sites Output Enablement Slice
+
+Hypothesis:
+
+- If QuickFork exposes Sites as part of the launch package, founders, open-source maintainers, and DevRel operators can turn repository evidence into a reviewable launch site/page spine instead of treating launch output as only README, deck, social, visual, and outreach material.
+
+Lifecycle stage:
+
+- Activation to monetization.
+
+Target user:
+
+- AI/devtool builders, open-source maintainers, technical founders, and DevRel operators preparing a public launch.
+
+Changed surface:
+
+- `src/server/generation/types.ts`
+- `src/server/generation/launch-brief.ts`
+- `src/server/generation/generation.test.ts`
+- `src/components/landing/HeroSection.tsx`
+- `src/components/studio/ProductStudio.tsx`
+- `src/components/marketing/LeadCaptureForm.tsx`
+- `src/core/pipeline.ts`
+- `src/App.test.tsx`
+- `src/server/marketing/launch-package-fit.test.ts`
+- `src/server/marketing/lead-capture.test.ts`
+- `docs/marketing/research/2026-06-03-sites-output-enablement.md`
+
+Metric:
+
+- `launch_artifact_copied` or `launch_artifact_downloaded` where `artifact_type=site`.
+- `sales_contact_requested` with package scope including `sites`.
+
+Guardrail:
+
+- Sites output is a reviewable page spine, not automatic publishing.
+- Do not send raw README text, raw artifact body, email, token, secret, unsupported customer proof, rankings, revenue, pricing, or guaranteed-growth claims to browser analytics.
+
+Evidence observed:
+
+- RED generation test failed first because the launch brief did not include the site channel.
+- RED App tests failed first because the studio and full launch package form did not expose Sites.
+- Focused generation verification passed: `npm test -- src/server/generation/generation.test.ts -t "source-backed free repo launch brief"` returned 1 file passed and 1 selected test passed.
+- Focused App verification passed: `npm test -- src/App.test.tsx -t "keeps the generator studio|full launch package contact|landing architecture"` returned 1 file passed and 3 selected tests passed.
+- Focused lead qualification verification passed: `npm test -- src/server/marketing/launch-package-fit.test.ts src/server/marketing/lead-capture.test.ts -t "launch package fit|full launch package qualification"` returned 2 files passed and 2 selected tests passed.
+- Full verification passed with threads pool: `npm test -- --testTimeout=30000 --pool=threads` returned 25 files passed and 168 tests passed.
+- Build verification passed: `npm run build` completed TypeScript build and Vite production build.
+- Diff hygiene passed: `git diff --check` returned no whitespace errors.
+
+Decision:
+
+- Treat Sites as a product-output and monetization hypothesis. It fills a landing-page material gap in the launch package, but production users have not yet copied, downloaded, or requested this scope.
+
+Next action:
+
+- Track site artifact copy/download and full launch package requests with `sites` scope before investing in full hosted site generation, CMS export, or public landing-page automation.
+
+## 2026-06-03 SEO/GEO Evidence Workbook Slice
+
+Hypothesis:
+
+- If QuickFork expands each registered Search/AI baseline row into a surface-specific audit workbook, SEO/GEO reviewers can collect Search Console and AI-answer evidence without drifting from the experiment contract.
+
+Lifecycle stage:
+
+- Validation, P0 measurement and evidence.
+
+Target user:
+
+- SEO/GEO supervisor auditing whether QuickFork is indexable, extractable, and accurately summarized.
+- Growth operator collecting Search Console and AI-answer evidence for active page-intent experiments.
+
+Changed surface:
+
+- `src/marketing/seo-geo-audit-workbook.ts`
+- `src/marketing/seo-geo-audit-workbook.test.ts`
+- `docs/marketing/data/seo-geo-audit-workbook.csv`
+- `docs/marketing/research/2026-06-03-seo-geo-evidence-workbook.md`
+- `docs/superpowers/plans/2026-06-03-seo-geo-evidence-workbook.md`
+
+Metric:
+
+- Completion of Search Console and AI-answer workbook rows for experiments that require search or AI evidence.
+- Supporting route metrics remain `page_view`, `cta_clicked`, `generation_started`, `generation_completed`, and `generation_failed` over the same 14-day window.
+
+Guardrail:
+
+- Do not treat a workbook row, prompt, route, or crawler asset as validated search demand or AI visibility.
+- Do not claim ranking wins, AI citations, customer traction, revenue, conversion lift, pricing validation, Product Hunt outcomes, or guaranteed launches.
+- Keep workbook rows free of raw emails, raw repo text, source notes, private launch notes, tokens, secrets, raw artifacts, or unsupported proof.
+
+Evidence observed:
+
+- RED test failed first because the SEO/GEO audit workbook module did not exist.
+- Focused verification passed after adding the module and CSV mirror: `npm test -- src/marketing/seo-geo-audit-workbook.test.ts` returned 1 file passed and 5 tests passed.
+- Related verification passed: `npm test -- src/marketing/seo-geo-audit-workbook.test.ts src/marketing/search-ai-baseline.test.ts src/marketing/growth-experiment-report.test.ts` returned 3 files passed and 21 tests passed.
+- Initial full `npm test` attempts with the default forks pool hit Vitest worker startup timeouts in this local environment.
+- Targeted reruns of the affected files passed, including `npm test -- src/App.test.tsx -t "keeps the generator studio|full launch package contact|landing architecture"` and `npm test -- src/server/marketing/launch-package-fit.test.ts`.
+- Full verification passed with threads pool: `npm test -- --testTimeout=30000 --pool=threads` returned 25 files passed and 168 tests passed.
+- Build verification passed: `npm run build` completed TypeScript build and Vite production build.
+- Diff hygiene passed: `git diff --check` returned no whitespace errors.
+
+Decision:
+
+- Treat this as P0 measurement infrastructure, not a validation result. It makes the current active experiments reviewable by a senior SEO/GEO operator before QuickFork scales more landing pages.
+
+Next action:
+
+- Run related/full verification, then use the workbook to collect 14-day Search Console and AI-answer evidence before deciding which active page-intent experiments should be promoted, rewritten, consolidated, or paused.