GitHub Copilot Usage Metrics API Now Names Custom Skills, Agents, and MCP Servers
GitHub updated the Copilot usage metrics API on October 22, 2026, so enterprise and organization reports now show the configured names of customer-defined skills, custom agents, Model Context Protocol (MCP) servers, slash commands, and plugins instead of grouping them under generic 'other' or 'custom' labels. The change builds on the agentic CLI customization metrics GitHub released in September.
The named entries appear across per-user 1-day and 28-day reports, aggregated 1-day reports, and day_totals in aggregated 28-day reports, in the totals_by_skill, totals_by_custom_agent, totals_by_mcp, totals_by_slash_cmd, and totals_by_plugin arrays. Each breakdown array now carries up to 25 entries per record, up from five, surfacing more of the long tail of customizations, with remaining activity consolidated into a trailing 'other' entry.
Name reporting covers Copilot CLI interactive sessions and prompt mode, the GitHub Copilot app, Agent Client Protocol integrations such as JetBrains, and the Copilot SDK. Prompt mode, the app, and ACP integrations require Copilot CLI 1.0.97 or later, and the Copilot SDK requires v1.0.21 or later - names are not reported from an IDE's built-in chat telemetry, so outdated clients will undercount.
Names are reported exactly as configured, trimmed to 128 characters, and identifiers that look like file system paths stay grouped under 'other' or 'custom' for security. GitHub advises reviewing names before sharing reports outside an organization, since internal codenames become visible in telemetry exports.
Access is limited to enterprise owners and billing managers, organization owners, and holders of custom roles with the 'View Copilot Metrics' permission, and the organization's Copilot usage metrics policy must be enabled for data to flow.
For engineering leadership and platform teams, the practical gain is observability: organizations investing in bespoke domain skills or private MCP servers can now measure actual developer adoption by name, benchmark post-rollout utilization, and spot underused internal tooling. The trade-offs are governance-related - names become telemetry data - and coverage gaps where developers run older Copilot clients.