Add MCP server and enhance AI content tools

Implement self-hosted MCP server for admin actions, update AI generation tools in admin dashboard, and refine mobile app stores and components.

apps/admin/src/lib/ai/analyzers/deletions.ts

@@ -0,0 +1,101 @@
+// =============================================================================
+// Deletion Analysis Functions
+// Suggest removals for low-quality or duplicate questions
+// =============================================================================
+
+import { getOpenAI, getModel, getTemperature } from '../client';
+import type { DeletionAnalysis } from '../types';
+
+/**
+ * Analyze questions and suggest deletions (duplicates, unsafe, etc.)
+ */
+export async function analyzeQuestionDeletions(
+    questions: { id: string; text: string; partner_text?: string | null; intensity?: number }[],
+    isExplicit: boolean = false
+): Promise<DeletionAnalysis[]> {
+    const openai = getOpenAI();
+
+    const simplifiedQuestions = questions.map(q => ({
+        id: q.id,
+        text: q.text,
+        partner_text: q.partner_text,
+        intensity: q.intensity,
+    }));
+
+    const toneInstruction = isExplicit
+        ? 'CONTENT: EXPLICIT pack. Adult content is allowed but must be consensual, safe, and clear. Avoid coercion, minors, illegal acts, or unsafe suggestions.'
+        : 'CONTENT: NON-EXPLICIT pack. Do NOT include explicit sexual acts, crude language, or NSFW content. If the core action is sexual, it should be removed, not rewritten.';
+
+    const prompt = `<task>
+Identify questions that should be DELETED instead of edited.
+Only include questions that are better removed than fixed.
+</task>
+
+<content_type>
+${toneInstruction}
+</content_type>
+
+<deletion_categories>
+- duplicate: same core action as another question (near-duplicate)
+- redundant: minor variation with no added value
+- off-tone: violates explicit vs non-explicit rules
+- unsafe: coercion, minors, illegal acts, or consent violations
+- too-vague: not actionable or not a clear proposal
+- broken: grammar/structure too broken to repair without inventing content
+- off-topic: not relevant to a couples intimacy app
+</deletion_categories>
+
+<rules>
+1. Prefer edits over deletions when possible.
+2. Duplicates: delete the weaker/less specific version and keep the clearer one.
+3. If duplicate, set duplicate_of_id to the kept question's id.
+4. Do NOT delete for minor wording issues or intensity mismatches.
+</rules>
+
+<questions_to_analyze>
+${JSON.stringify(simplifiedQuestions)}
+</questions_to_analyze>
+
+<output_format>
+{
+  "deletions": [
+    {
+      "id": string,
+      "category": "duplicate"|"redundant"|"off-tone"|"unsafe"|"too-vague"|"broken"|"off-topic",
+      "reason": string,
+      "duplicate_of_id": string|null
+    }
+  ]
+}
+
+Only return questions that should be deleted.
+</output_format>`;
+
+    const systemMessage = isExplicit
+        ? 'You are a strict content quality reviewer for an adult couples app. Remove only what should be deleted. Always respond with valid JSON only.'
+        : 'You are a strict content quality reviewer for a couples app. Remove only what should be deleted. Always respond with valid JSON only.';
+
+    const response = await openai.chat.completions.create({
+        model: getModel('fix'),
+        messages: [
+            {
+                role: 'system',
+                content: systemMessage,
+            },
+            { role: 'user', content: prompt },
+        ],
+        response_format: { type: 'json_object' },
+        temperature: getTemperature('fix', 0.2),
+    });
+
+    const content = response.choices[0].message.content;
+    if (!content) throw new Error('No content generated');
+
+    let jsonContent = content.trim();
+    if (jsonContent.startsWith('```')) {
+        jsonContent = jsonContent.replace(/^```(?:json)?\s*\n?/, '').replace(/\n?```\s*$/, '');
+    }
+
+    const parsed = JSON.parse(jsonContent);
+    return parsed.deletions || [];
+}