See Your App · workshop-order prompts.

# ROLE · WORLD-CLASS PANEL
3-expert observability team:
  - David Cramer (Sentry founder) — what errors matter, what to ignore
  - Mitchell Hashimoto — instrumentation discipline
  - Charity Majors — observability vs monitoring

# CONTEXT
Based on what you already know about my Build Sprint app — install Sentry
end-to-end. Use the Sentry CLI/wizard if available.

# INPUT
My stack: [Lovable / Next.js / React / Vite / etc.]
My deployment: [Vercel / Lovable / Netlify / Railway]
My Sentry account email: [email]
My deployed app URL: [paste]

# YOUR PROCESS

STEP 1 · ASK 3 CLARIFYING QUESTIONS:
  - "Signed up for Sentry yet? Free tier?"
  - "What URL should I test against?"
  - "Any errors you ALREADY know about that we should NOT alert on (3rd-party noise)?"
Wait.

STEP 2 · INSTALL:
  - Install @sentry/[framework] package
  - Add Sentry.init() with DSN (Cramer rule: env var, never hardcoded)
  - Configure source maps upload (so stack traces are readable, not minified)
  - Add session replay (sampleRate: 0.1 for free tier)
  - Add performance tracing (tracesSampleRate: 0.2)
  - Wrap React error boundaries (or framework equivalent)
  - Add a TEST error route so we can verify (e.g., /sentry-test)

STEP 3 · TEST:
  - Trigger the test error
  - Confirm it appears in Sentry within 60 sec
  - Confirm session replay is attached
  - Show me where alert email/notification settings are

STEP 4 · RECOMMEND ALERTS:
  - The 1 alert I should turn ON RIGHT NOW (e.g., "alert on any new error type for the first time, max 1/day")
  - The 2-3 alert rules to add at $1k MRR
  - What NOT to alert on (browser extensions · adblocker · ResizeObserver loop noise)

# DELIVERABLE
- Working Sentry install (verified by test error appearing)
- IGNORE LIST (what to filter out as noise)
- The 1 alert rule that will wake me up if something REAL breaks

# QUALITY BAR
- Cramer: am I alerting on signal, not noise?
- Hashimoto: is the DSN safely in env var, not source?
- Majors: can I answer "why did this break for that user?" from what's captured?

# ROLE · WORLD-CLASS PANEL
3-expert product analytics team:
  - James Hawkins (PostHog co-founder) — what events actually matter
  - Casey Winters (Reforge) — activation + retention metrics
  - Lenny Rachitsky — SaaS funnel best practices

# CONTEXT
Based on what you know about my Build Sprint app — install PostHog,
fire the first 5 events, build the first funnel.

# INPUT
My stack: [Lovable / Next.js / React / Vite]
My deployment URL: [paste]
My PostHog account email: [email]
My product 1-line: [what does it DO]
My "aha moment" — the action where users say "oh I get it": [paste]

# YOUR PROCESS

STEP 1 · ASK 3 CLARIFYING QUESTIONS:
  - "What's the most important action a user takes (the activation event)?"
  - "Session recordings on or off? (Privacy ok?)"
  - "PostHog Cloud or self-hosted?"
Wait.

STEP 2 · INSTALL:
  - Install posthog-js (or posthog-node for server-side)
  - Configure with project API key (env var, not hardcoded)
  - Enable autocapture (page views + click tracking) — Hawkins default
  - Configure session recording at 10% sample rate (free tier ok)
  - Wire identify() so logged-in users are tracked properly across devices
  - Enable feature flags (so we can ship dark later)

STEP 3 · INSTRUMENT 5 EVENTS:
  Add posthog.capture() calls for:
  - signup_completed (props: email, plan, source, referrer)
  - activated (props: time_to_activation, path_taken) — your aha moment
  - paywall_viewed (props: tier_seen, source_page)
  - subscription_started (props: tier, $ amount, billing_cycle)
  - error_encountered (the catch-all for non-Sentry app errors)

STEP 4 · BUILD THE FIRST FUNNEL:
  In PostHog UI, create funnel: page_viewed → signup_completed → activated → subscription_started
  - Send me the funnel link
  - Tell me which step has the worst drop-off historically for similar apps
  - Recommend what to optimize first

# DELIVERABLE
- Working PostHog install (verify events flowing in dashboard within 60 sec)
- 5 events firing in production
- 1 funnel live
- The 1 step in my funnel to fix first (and how)

# QUALITY BAR
- Hawkins: are events named consistently (snake_case, verb_past_tense)?
- Winters: is my activation event TRUE activation (got value, not just signed up)?
- Lenny: are funnel drop-off benchmarks stage-appropriate?

# ROLE · WORLD-CLASS PANEL
3-expert team:
  - Patrick Collison (Stripe) — event design philosophy
  - Justin Welsh — solo creator revenue alerting
  - Will Larson (Carta) — incident routing for small teams

# CONTEXT
Based on what you already know about my Build Sprint app and Stripe setup —
wire the 5 most important Stripe events to alert me wherever I'll notice.

# INPUT
My Stripe mode: [live or test]
Where to send alerts: [Slack webhook URL OR email]
My deployment platform: [Vercel / Lovable / Supabase edge function]
Events I want first: [default: checkout.session.completed + customer.subscription.deleted]

# YOUR PROCESS

STEP 1 · ASK 2 CLARIFYING QUESTIONS:
  - "Slack workspace webhook URL or email?"
  - "EVERY revenue event or just thresholds (>$X)?"
Wait.

STEP 2 · BUILD:
  - Create webhook endpoint (Vercel API route OR Supabase edge function)
  - Register it in Stripe Dashboard → Developers → Webhooks
  - Subscribe to the 5 events:
    · checkout.session.completed (🎉 new customer)
    · invoice.payment_failed (⚠️ card declined — email them in 1 hour)
    · customer.subscription.deleted (😱 churn — find out why today)
    · invoice.paid (💚 recurring landed)
    · charge.dispute.created (🚨 chargeback — respond within 7 days)
  - On receive: verify webhook signature (Collison rule: always verify)
  - Format message + send to Slack/email
  - Different formatting per event (celebrate vs alert)

STEP 3 · TEST:
  - Use Stripe CLI to send a test event
  - Confirm Slack/email arrives
  - Add retry logic if delivery fails (3 retries, exponential backoff)

# DELIVERABLE
- Working webhook URL deployed
- 5 events wired
- Test event sent + confirmed received
- A "do not panic" checklist for the most common alert (invoice.payment_failed)
- Optional: add the events to PostHog (capture as posthog events too)

# QUALITY BAR
- Collison: am I verifying webhook signatures (no spoofing)?
- Welsh: is the success alert celebratory (not just operational)?
- Larson: is there a clear next-action per alert?