Claude Code gets plugin evaluation with six graders and a baseline run

Anthropic has added a plugin evaluation command to Claude Code, its agentic coding tool. `claude plugin eval` runs a plugin against realistic prompts, grades what Claude produces, and compares that with a run where the plugin is not loaded. The workflow includes six grader types, a no-plugin baseline, and a CI gate for skills. It is meant to answer three questions plugin developers previously could not measure: whether the skill triggers, whether it changes the output, and whether that change is an improvement. Details of the grader types and CI integration were not included in the initial announcement.