fix(gateway): correct Codestral embedding dimension

[marius-kilocode]

Kilo indexing uses /api/gateway/embedding-models as the source of truth for local vector-store dimensions. After #3196 stopped forwarding dimensions to upstream providers, mistralai/codestral-embed-2505 returns its native 1536-dimensional vectors.

The catalog still advertised Codestral as 256 dimensions, so the client created a 256-dimensional Qdrant collection and then upserted 1536-dimensional vectors. Qdrant rejects that with Wrong input: Vector dimension error: expected dim: 256, got 1536, which surfaced in the extension as a generic Bad Request.

This updates the Codestral catalog entry to 1536 so the vector store is created with the dimension the gateway actually returns.

[kilo-code-bot]

Code Review Summary

Status: No Issues Found | Recommendation: Merge

Changes Since Last Review

The new commit reverted buildDownstreamResponse entirely (resolving both previously flagged issues — the .clone() stream leak and the as cast violations), and instead fixes the root cause: the Codestral catalog dimension was wrong (256 → 1536). This is a cleaner fix.

• ✅ RESOLVED — WARNING: .clone() stream leak (embedding-request.ts:49) — buildDownstreamResponse removed
• ✅ RESOLVED — SUGGESTION: as casts (embedding-request.ts:56) — buildDownstreamResponse removed

Files Reviewed (5 files)

• apps/web/src/lib/ai-gateway/embeddings/embedding-request.ts — no issues
• apps/web/src/lib/ai-gateway/embeddings/embedding-request.test.ts — no issues
• apps/web/src/lib/ai-gateway/embeddings/kilo-embedding-models.ts — no issues
• apps/web/src/app/api/gateway/embedding-models/route.test.ts — no issues
• apps/web/src/app/api/openrouter/embeddings/route.ts — no issues

---

_{Reviewed by claude-sonnet-4.6 · 245,185 tokens}