signal insight

GitHub Copilot usage metrics API now reports on CLI skills, custom agents, MCP se…

GitHub expanded its Copilot usage metrics API to cover agentic CLI customizations — skills, custom agents, Model Context Protocol (MCP) servers, slash commands, and plugins — across per-user and aggregate 1-day and 28-day reports. GitHub-provided item names are shown while customer-defined names are grouped generically to protect privacy.

Published Sep 17, 2026 Updated Sep 19, 2026 1 sources

GitHubGitHub Copilot CLIenterprise controlsfeature updatelow impact

enterprise-controlsagentsai-coding-toolsfeature update

Impact: low
Confidence: 80%
Change type: feature update
First seen: Sep 17, 2026
Last updated: Sep 19, 2026
Audience: enterprise administratorsplatform engineersDevEx teams
Status: Ready

Summary

What changed

The Copilot usage metrics API added agentic activity metrics for CLI customizations (skills, custom agents, MCP servers, slash commands, plugins); for MCP servers, interaction_count increments only on connect/reconnect attempts, and customer-defined names are grouped under generic labels while recognized GitHub-provided items are named.

Why it matters

Organizations are accumulating home-grown agent skills, MCP servers, and plugins with little visibility into what developers actually use. Surfacing that in the metrics API lets platform teams find traction and enablement gaps and invest in the automations that pay off, turning a sprawling internal agent-tooling estate into something measurable and governable.

Evidence excerpt

GitHub Copilot expands existing CLI report coverage with agentic activity metrics for skills, custom agents, Model Context Protocol (MCP) servers, slash commands, and plugins.

Sources

github.blog