Multinex Docs
Guide

Teams & Namespaces

MemQ organizes memory into namespaces. Personal memory stays private. Team memory is shared. Community knowledge lives in _commons.

Namespace tiers

Your VISA tier determines which namespaces you can access.

FREE_commons only

Read and write to the shared public commons.

BASE_commons + 1 private namespace

Personal memory with one private namespace.

VERIFIED_commons + multiple private namespaces

Multiple private namespaces for different projects.

TEAM_commons + org: namespaces

Shared team memory with org-level namespaces.

Namespace conventions

Naming patterns
# Personal
user:alex                    # Daily continuity
user:alex:research           # Research project
user:alex:code               # Coding workspace

# Team
org:infra                    # Core team memory
org:infra:launch-q2          # Project-specific
org:infra:incident-2026-04   # Time-scoped

# Shared
_commons                     # Public knowledge pool

Individual with multiple agents

Recommended tier: BASE (one namespace) or VERIFIED (multiple namespaces).

  • Put personal durable memory in user:handle or user:handle:project
  • Point every local agent at the same namespace when they should share memory
  • Give exploratory agents separate project namespaces when you want isolation
  • All agents can read _commons regardless of namespace

Recommended tools

  • journal_record, journal_search, journal_distill — for continuity across sessions
  • mnemosyne_context — for token-efficient context packing
  • brain_reinforce — for post-task learning
  • commons_resonance — for tapping shared knowledge

Team with multiple agents

Recommended tier: TEAM.

  • Use one stable org: namespace per team, then branch by project
  • Keep durable team context in org:team
  • Use project sub-namespaces for higher-churn work
  • Let all agents read _commons in addition to the team namespace

Recommended split

  • Coordinator agent — writes summaries and decisions to org:team
  • Specialist agents — write transient work product to org:team:project
  • Review/audit agents — read both the team namespace and _commons

_commons — shared knowledge pool

_commons is the governed public memory plane. All tiers can read it. Contributions go through governed flows — regular users don't write to _commons directly.

Default content

Nexus Ranger automatically seeds _commons with validated feeds:

  • threat_intel — CISA advisories, NVD, CrowdStrike, Cisco Talos
  • breaking_news — OpenAI, Hugging Face, Google AI, GitHub, Cloudflare
  • industry_trade — SEC EDGAR, ecosystem shifts
  • resolved_issue — migration notes, changelog fixes, workarounds

Accessing _commons

Use commons_resonance to search the shared pool. Your AI can surface community-wide knowledge — threat intelligence, industry news, resolved issues — alongside your private memories.