Architecture: How should session documentation be synced to production Mon...

Documentation for Sessions

🏗️ Architecture: How should session documentation be synced to production Mon...

Date: 2026-01-22
Time: 19:22:11 UTC
Session: editor2-58.12
Agent: backend
Confidence: 95%
Status: ✅ Documented

📋 Decision Context

Question:

How should session documentation be synced to production MongoDB for searchability?

Task: Fix session docs sync authentication and document the pattern

🎯 Decision Made

Chosen:

Via backend-docs-api with SYSTEM_TOKEN auth

Options Considered:

Direct MongoDB writes from scripts
Via backend-docs-api with JWT auth
Via backend-docs-api with SYSTEM_TOKEN auth
Manual upload via admin panel

💭 Reasoning

The JwtAuthGuard in @libs-api/auth-core has built-in support for SYSTEM_TOKEN authentication (lines 63-75). When a token matches the SYSTEM_TOKEN env var, it bypasses JWT verification and authenticates as a system user. This is the documented pattern for internal service-to-service communication. The sync script uses getSystemToken() from env-loader.js to read the actual token from .env.local, ensuring proper authentication without needing to generate/manage JWTs for scripts.

📁 Files Involved

tools/scripts/database/sync-session-docs-to-mongodb.js
libs/libs-api/auth-core/src/lib/guards/jwt-auth.guard.ts
tools/scripts/database/env-loader.js

🔗 Related

Decision ID: dev_1769109730984_c9xx6w5
Session: 58
Workspace: editor2

Auto-generated by MCP Smart Consolidation