Configure Massu AI for team collaboration with knowledge sharing and conflict detection

Team Setup

When multiple developers use AI assistants on the same codebase, Massu AI's team features help coordinate work, share knowledge, and prevent conflicts.

Enabling Team Features

In your massu.config.yaml:

yaml

team:
  enabled: true
  sync_backend: local
  expertise_weights:
    session: 20
    observation: 10
  privacy:
    share_file_paths: true
    share_code_snippets: false
    share_observations: true

How Team Knowledge Works

Expertise Tracking

When a developer works on a module (edits files, fixes bugs, makes decisions), Massu AI tracks their expertise:

Session count: How many sessions touched the module
Observation count: How many meaningful observations were made
Expertise score: Logarithmic formula: min(100, log2(sessions+1) 20 + log2(observations+1) 10)

The logarithmic scale means early experience counts more -- going from 0 to 5 sessions is more impactful than going from 50 to 55.

Conflict Detection

Massu AI detects when two developers edit the same file within a configurable time window:

> Use massu_team_conflicts --days 7

This helps prevent merge conflicts and duplicated effort.

Privacy Controls

Team features respect privacy settings:

Setting	Default	What It Controls
`share_file_paths`	true	File paths in shared observations
`share_code_snippets`	false	Actual code content
`share_observations`	true	Observation summaries

With default settings, team members can see which files were worked on and what was decided, but not the actual code changes.

Sync Backends

Local (Default)

yaml

team:
  sync_backend: local

Each developer has their own local database. Team data is shared through git (the .claude/session-state/ directory).

Shared

yaml

team:
  sync_backend: shared

Team data is synced through a shared database location (e.g., a network drive or shared directory).

Getting Started

Enable team features in your config
Have each team member run a few sessions
Use massu_team_expertise to see the expertise map
Use massu_team_conflicts daily to catch overlapping work
Use massu_team_search to find who knows about specific modules

Tips

Start with share_code_snippets: false for privacy
Expertise scores are most useful after 2-3 weeks of data collection
Run massu_team_conflicts at standup meetings to coordinate work
High expertise in a module means that developer is the go-to person for questions about it