Fix the entire codebase 🇫🇮

[alexechoi]

Fix the entire codebase 🇫🇮

High-level PR Summary

This PR implements a caching layer for entity research reports and redesigns the authentication page with a modern UI. The research API now checks a cache collection before querying entities, attempting multiple cache ID variations to locate cached reports. The dashboard redirects to /reports/{reportId} instead of /entity/{id} when cached data is found. The auth page receives a complete visual overhaul with a dark theme, gradient effects, stats display, and improved form design while maintaining the same functionality.

⏱️ Estimated Review Time: 15-30 minutes

💡 Review Order Suggestion

Order	File Path
1	junction-app/app/auth/page.tsx
2	junction-app/app/api/research/route.ts
3	junction-app/app/dashboard/page.tsx

⚠️ Inconsistent Changes Detected

File Path	Warning
junction-app/app/auth/page.tsx	A complete UI redesign of the authentication page seems unrelated to fixing caching/report lookup logic described in the other files

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
aegis	Ready	Preview	Comment	Nov 15, 2025 9:43pm

[recurseml]

Review by RecurseML

🔍 Review performed on 949a33d..0d9ce39

Severity	Location	Issue	Delete
	junction-app/app/api/research/route.ts:109	Runtime TypeError on non-strings

✅ Files analyzed, no issues (2)

• junction-app/app/auth/page.tsx
• junction-app/app/dashboard/page.tsx

[recurseml]

Runtime TypeError when non-string values are processed. The code filters possibleCacheIds from matchedEntity properties (cacheId, cache_id, id, name, etc.) using .filter(Boolean), which only removes falsy values but does NOT guarantee the remaining values are strings. Since matchedEntity is typed as Record<string, any> and populated from untyped Firestore data, these fields could be numbers, objects, arrays, or other non-string types. The type annotation (value: string) is merely a type assertion and does not perform runtime validation. When .toLowerCase() is called on a non-string value, it will throw: TypeError: value.toLowerCase is not a function. This will crash the API endpoint when triggered by Firestore data containing non-string cache ID fields. To fix, convert values to strings before calling toLowerCase(): .map((value) => String(value).toLowerCase())

---

React with 👍 to tell me that this comment was useful, or 👎 if not (and I'll stop posting more comments like this in the future)