🏗️ Architecture: How should Vitto orchestrator be enhanced to stop producing ...

Date: 2026-02-07
Time: 14:13:13 UTC
Session: gspace1-0
Agent: angular
Confidence: 92%
Status: ✅ Documented

📋 Decision Context

Question:

How should Vitto orchestrator be enhanced to stop producing fake completions?

Task: Create comprehensive Vitto enhancement handoff for new coding agent

🎯 Decision Made

Chosen:

4-phase approach: type-aware strategy engine, output quality gates, stale threshold fix, standalone module + learning pipeline. Separate session, not alongside current work.

Options Considered:

• No alternatives documented

💭 Reasoning

Vitto's core problem is strategy selection blindness — it ignores handoff.type and routes everything to direct AI execution. The fix requires architectural changes to the 2100-line orchestrator service (strategy engine rewrite, quality validation layer, rejection capability). This is too large and risky to do alongside active inter-agent work. Dedicated session with clear phases ensures each change can be tested independently.

📁 Files Involved

• apis/backend-ai-api/src/app/vitto/vitto-orchestrator.service.ts
• apis/backend-ai-api/src/app/vitto/vitto.module.ts
• apis/backend-ai-api/src/app/vitto/vitto-standalone.module.ts
• apis/backend-ai-api/src/app/infrastructure/services/handoff.service.ts
• apis/backend-api/src/app/ai-agents/services/tenant-ai.service.ts
• docs/english/Handoffs/HANDOFF-VITTO-ORCHESTRATOR-ENHANCEMENT.md

🔗 Related

• Decision ID: dev_1770473593881_lqblt0e
• Session: 0
• Workspace: gspace1

Auto-generated by MCP Smart Consolidation