feat: validate podcast cost estimation accuracy, document per-token costs, and fix subscription/plan enforcement

Issue #543 — Validate Estimated Cost Accuracy (UI vs Backend)

Backend:
- cost_estimator.py uses pricing catalog (APIProviderPricing) as single source of truth
- All 7 cost components: analysis, research (search+LLM), script, TTS, voice clone, avatar, video
- initialize_default_pricing() runs on every app startup for auto-sync

Frontend cost estimation fixes:
- Added missing analysisCost, scriptCost, voiceCloneCost to PodcastEstimate type
- toPodcastEstimate() now extracts all 7 backend fields (was dropping 3)
- headerCostEst maps analysisCost->Analyze, scriptCost->Write, voiceCloneCost->Produce
- EstimateCard shows 5 chips: Analysis, Research, Script, Voice(TTS+clone), Visuals(avatar+video)
- Chip sum now equals backend total for all configurations

Subscription & plan fixes:
- Removed Stripe re-verification from checkSubscription() (downgrade regression fix #539)
- Added verifyCheckoutRef pattern for reliable mount-time checkout polling
- One-time Stripe sync effect with pending_subscription_change flag for Customer Portal returns
- Free plan limits: stability_calls 3->10, audio_calls 5->10 (supports 2 podcasts)
- Image enforcement uses actual provider (GPT_PROVIDER), not hardcoded Stability
- Billing/pricing pages bypass onboarding check in ProtectedRoute
- Gradient buttons + loading spinner on plan chip in UserBadge
- Added metadata-based Stripe lookup fallback (Issue #538)

Documentation:
- TESTING_GUIDE.md: comprehensive testing instructions for non-technical testers
  - Free plan limits, usage tracking, cost estimation formulas
  - 10 test cases for UI verification
  - Troubleshooting guide
  - Quick-reference cost formulas with all default rates

Cleanup: removed legacy ToBeMigrated directory (70+ files, ~22K LOC)
GSC Brainstorm: service, hook, modal, and UI components for blog topic brainstorming

backend/services/subscription/pricing_service.py

@@ -564,11 +564,11 @@ class PricingService:
                 "serper_calls_limit": 10,
                 "metaphor_calls_limit": 0,  # DISABLED: Metaphor not in Free tier
                 "firecrawl_calls_limit": 0,  # DISABLED: Firecrawl not in Free tier
-                "stability_calls_limit": 3,  # 3 images - enough to try the product
+                "stability_calls_limit": 10,  # 10 images - enough for 2 podcasts (5 images each)
                 "exa_calls_limit": 10,  # 10 research queries - enough to try the product
                 "video_calls_limit": 2,  # 2 video renders - try podcast video on Free
                 "image_edit_calls_limit": 5,  # 5 image edits - enough to try the product
-                "audio_calls_limit": 5,  # 5 audio clips - enough to try the product
+                "audio_calls_limit": 10,  # 10 audio clips - enough for 2 podcasts (5 clips each)
                 "wavespeed_calls_limit": 0,  # 0 = unlimited for Free; video controlled via video_calls_limit
                 "gemini_tokens_limit": 50000,
                 "openai_tokens_limit": 0,  # DISABLED