refactor(server): drop stale prompt text, duplicated routes and dead options

- lib/system-prompts.ts: delete the extended prompt kept only to reach
  the 4096-token cache threshold of two old Claude models; describe the
  current canvas-first UI in the App Features section
- app/api/chat/route.ts: remove the server-side example cache (the
  browser answers examples itself), the ENABLE_HISTORY_XML_REPLACE
  option and its helper, the unread messageMetadata fields; the
  DEBUG_LLM_PAYLOAD blocks become one describeMessages summary
- app/api/log-feedback and log-save share scoreLatestTrace in
  lib/langfuse.ts
- app/api/admin/test-model and validate-model call a shared
  testModelConnection() instead of one route faking a request to the
  other and passing trust through the body
- lib/dynamo-quota-manager.ts reads the failing item from the
  conditional update instead of a second GetItem
- lib/ai-providers.ts: geminiThinkingOptions helper for the google and
  vertexai cases, drop the unused fourth resolveBaseURL parameter
- edge-functions: always stream, shrink messageSchema, drop model
  aliases nothing sends
- delete app/api/chat/xml_guide.md, parseOutputTokenLimit, unused type
  exports and export keywords on file-internal symbols
- env.example: drop variables nothing reads, document the quota,
  validation and analytics variables, mention the apiKeyEnv array form
  (also in docs/*/ai-providers.md)

lib/langfuse.ts

@@ -1,10 +1,59 @@
 import { LangfuseClient } from "@langfuse/client"
 import { observe, updateActiveTrace } from "@langfuse/tracing"
 import * as api from "@opentelemetry/api"
+import { randomUUID } from "crypto"
 
 // Singleton LangfuseClient instance for direct API calls
 let langfuseClient: LangfuseClient | null = null
 
+type IngestionBatch = Parameters<
+    LangfuseClient["api"]["ingestion"]["batch"]
+>[0]["batch"]
+
+// Attach a score to the session's most recent chat trace. When the session
+// has no trace yet, a fallbackTrace creates a standalone trace to carry the
+// score; without one the score is skipped. Returns whether a score was written.
+export async function scoreLatestTrace(
+    langfuse: LangfuseClient,
+    opts: {
+        sessionId: string
+        name: string
+        value: number
+        comment: string
+        fallbackTrace?: {
+            name: string
+            userId: string
+            input: unknown
+            metadata: Record<string, unknown>
+        }
+    },
+): Promise<boolean> {
+    const { sessionId, name, value, comment, fallbackTrace } = opts
+    const traces = await langfuse.api.trace.list({ sessionId, limit: 1 })
+    let traceId = traces.data?.[0]?.id
+    const timestamp = new Date().toISOString()
+    const batch: IngestionBatch = []
+
+    if (!traceId) {
+        if (!fallbackTrace) return false
+        traceId = randomUUID()
+        batch.push({
+            type: "trace-create",
+            id: randomUUID(),
+            timestamp,
+            body: { id: traceId, sessionId, ...fallbackTrace, timestamp },
+        })
+    }
+    batch.push({
+        type: "score-create",
+        id: randomUUID(),
+        timestamp,
+        body: { id: randomUUID(), traceId, name, value, comment },
+    })
+    await langfuse.api.ingestion.batch({ batch })
+    return true
+}
+
 export function getLangfuseClient(): LangfuseClient | null {
     if (!process.env.LANGFUSE_PUBLIC_KEY || !process.env.LANGFUSE_SECRET_KEY) {
         return null
@@ -22,7 +71,7 @@ export function getLangfuseClient(): LangfuseClient | null {
 }
 
 // Check if Langfuse is configured (both keys required)
-export function isLangfuseEnabled(): boolean {
+function isLangfuseEnabled(): boolean {
     return !!(
         process.env.LANGFUSE_PUBLIC_KEY && process.env.LANGFUSE_SECRET_KEY
     )