Architecture: How should knowledge flow between tiers (T3 private → space ...
Documentation for Sessions
🏗️ Architecture: How should knowledge flow between tiers (T3 private → space ...
Date: 2026-02-06
Time: 00:56:44 UTC
Session: editor2-93.3
Agent: backend
Confidence: 90%
Status: ✅ Documented
📋 Decision Context
Question:
How should knowledge flow between tiers (T3 private → space team → golden → core rule)?
Task: Phase 4A: Knowledge promotion pipeline
🎯 Decision Made
Chosen:
SpacePatternAggregatorService with 4-stage promotion pipeline: private (user) → team (5+ users) → platform/golden (3+ spaces, 90%+ success, T2 opt-in) → core rule (manual curation). Each promotion stage respects T2 SpaceAIPreferences controls.
Options Considered:
- No alternatives documented
💭 Reasoning
Leverages existing LearnedPattern.visibility field (private/team/platform) that was defined in Session 51.2 but never queried. The 5-user threshold for team promotion and 3-space/90% success for golden promotion balances knowledge quality with discovery speed. T2 controls (allowPatternPromotion, shareAnonymousPatternsToGlobal) respect the B2B2C data ownership model. PII anonymization before golden promotion ensures privacy.
📁 Files Involved
apis/backend-ai-api/src/app/agents/services/space-pattern-aggregator.service.tsapis/backend-ai-api/src/app/agents/schemas/space-ai-preferences.schema.ts
🔗 Related
- Decision ID:
dev_1770339404286_5r2m16n - Session: 93
- Workspace: editor2
Auto-generated by MCP Smart Consolidation