Coding agents can now access eve Agent Runs via Vercel MCP and CLI

Coding agents can now directly access eve Agent Runs through Vercel MCP and the CLI to view execution details and debug. With this update, agents can browse run history, inspect execution details, and fetch full Trace data including reasoning, tool calls, and token usage.

The newly added Vercel MCP tools are as follows.

• list_agent_run_projects: Search projects within a team that have Agent Runs activity
• list_agent_runs: View a list of recent runs for a project (status, model, trigger, usage, etc.)
• get_agent_run: Inspect metadata, lifecycle events, usage, and sub-agent data
• get_agent_run_trace: Retrieve full trace data including turns, messages, reasoning, tool calls, and inputs/outputs

The Vercel CLI also provides similar functionality through the vercel agent-runs command. All CLI subcommands support machine-readable --json output, and trace data is rendered in Markdown format when passed through a pipe. This allows even agents that cannot use MCP to directly call the CLI to debug their own execution process.