Team Analytics MCP
Relevant source files
The following files were used as context for generating this wiki page:
Team Analytics MCP is a specialized Model Context Protocol (MCP) server designed for proactive engineering health monitoring, tool effectiveness tracking, and training recommendations. It integrates multiple data sources, including Git history and shared team memory, to provide automated scoring and actionable insights for engineering teams.
The system focuses on support and coaching rather than punishment, utilizing a multi-dimensional scoring algorithm to track code quality, knowledge sharing, velocity, and collaboration. It provides automated alerts and mentoring suggestions to improve team performance and individual professional growth.
Sources: mcp-servers/team-analytics/README.md:1-12, mcp-servers/team-analytics/README.md:144-149
System Architecture
The Team Analytics MCP operates as a Node.js-based server that processes raw engineering data into high-level health scores and recommendations. It consumes data from team-memory.json for knowledge sharing and local Git repositories for technical metrics.
The architecture illustrates how raw data from Git and memory files flows through specific analyzers into a central scoring and alert engine before being surfaced to MCP clients.
Sources: mcp-servers/team-analytics/README.md:23-28, mcp-servers/TEST_RESULTS.md:13-20, mcp-servers/README-SETUP.md:7-25
Scoring Mechanisms
The server calculates an overall health score (0-100) based on weighted categories.
| Category | Weight | Key Metrics |
|---|---|---|
| Code Quality | 35% | Bug fix ratio, commit size, reverted commits, message clarity |
| Knowledge Sharing | 25% | Org-wide/Project contributions, confidence level, references by others |
| Velocity | 25% | Commits/Lines per week, velocity trends |
| Collaboration | 15% | (Future) PR review patterns, feedback loops |
Sources: mcp-servers/team-analytics/README.md:14-21, mcp-servers/team-analytics/README.md:112-127
Alert Thresholds
The Scoring Engine categorizes results into three primary alert levels:
- Critical (Score < 50): Indicates immediate intervention is needed due to low code quality or high bug rates (> 25%).
- Warning (Score < 65): Suggests close monitoring for velocity drops (> 20%) or low knowledge sharing.
- Positive (Score >= 90): Acknowledges high collaboration and significant improvement trends (> 15%).
Sources: mcp-servers/team-analytics/README.md:129-142
MCP Tools
The server exposes five primary tools to AI agents and developers for retrieving insights and recommendations.
| Tool Name | Purpose | Key Parameters |
|---|---|---|
engineer_health | Detailed health report for a specific individual | engineer_name, timeframe |
team_insights | Organization-wide average health and at-risk list | timeframe |
daily_summary | Health check summary for standups and focus areas | timeframe |
training_recommendations | Suggested training topics and attendee lists | urgency |
find_mentors | Suggested mentor-mentee pairings based on skills | mentee |
Sources: mcp-servers/team-analytics/README.md:30-81, mcp-servers/README-SETUP.md:74-80
Configuration and Setup
The server requires a config.json file to map engineers to their respective emails and define paths to data repositories.
{
"memoryPath": "../../.shared/team-memory.json",
"repoPath": "../../../LegalAI_System",
"engineers": [
{
"name": "Brett",
"email": "brett@example.com"
}
]
}
Sources: mcp-servers/team-analytics/README.md:83-95
Installation Workflow
To integrate Team Analytics into an MCP client, the following environment variables and command arguments are typically configured in .mcp.json:
Conclusion
Team Analytics MCP serves as a critical component of the engineering platform by turning passive development data into active coaching opportunities. By monitoring metrics across code quality and knowledge sharing, it provides a data-driven foundation for 1-on-1 meetings, sprint reviews, and organizational training planning while maintaining a focus on engineer support and privacy.
Sources: mcp-servers/team-analytics/README.md:97-110, mcp-servers/team-analytics/README.md:144-149