GitHub Copilot adds telemetry for skills, custom agents, MCP and plugins to the usage metrics API
GitHub has expanded the usage metrics API for Copilot CLI with new fields tracking the use of skills, custom agents, MCP servers, slash commands and plugins – administrators can now see which agentic tools developers actually use.
GitHub has expanded the usage metrics API for Copilot CLI with telemetry focused on agentic customizations – skills, custom agents, MCP servers, slash commands and plugins. The new fields totals_by_skill, totals_by_custom_agent, totals_by_mcp, totals_by_slash_cmd and totals_by_plugin list up to five of the most-used items in each category along with an interaction count (interaction_count), which, depending on the category, counts executions of skills, slash commands and plugin skills, starts of custom agents, or attempts to connect to an MCP server (for MCP, both successful and unsuccessful connection or reconnection attempts are counted, rather than individual tool calls). The additional fields distinct_skill_use_count, distinct_custom_agent_use_count, distinct_mcp_use_count, distinct_slash_cmd_use_count and distinct_plugin_use_count count the number of distinct items used – in aggregate reports, each item is counted once across the entire organization, not once per user.
The data is available in both per-user and aggregate reports at the enterprise and organization levels, specifically in 1-day reports (both per-user and aggregate), in 28-day per-user reports and in the day_totals records of aggregate 28-day reports. Access requires the View Copilot Metrics permission and the Copilot usage metrics policy to be enabled.
According to GitHub, these additions allow enterprise and organization administrators to determine which customizations in Copilot CLI are gaining popularity, identify gaps in adoption and focus investments on automations that developers actually find useful.
Why it matters
For Copilot Enterprise/Organization administrators, this provides a concrete tool for assessing which agentic CLI extensions (skills, custom agents, MCP integrations, plugins) are actually gaining traction within the team, and using that information to decide on further investments in training or deployment. For individual developers, it means that their use of these extensions is now explicitly measured and visible to administrators with the appropriate permission.
Two audiences, two different impacts
What this means
For individuals
The use of agentic extensions in Copilot CLI is now tracked in per-user reports, so an individual developer’s activity with these tools is visible to administrators with the appropriate permission.
For a business
GitHub Enterprise/Organization administrators gain insight through the usage metrics API into which agentic customizations (skills, custom agents, MCP servers, slash commands, plugins) developers actually use in Copilot CLI, which, according to GitHub, helps target investments and identify gaps in tool adoption.
ProcessesCheck the original
Event sources
clearly official source · 1 publisher, 0 independent. We count feeds from the same owner only once.