Claude Code adds a tool to compare results with and without plugins
A post introduces “claude plugin eval” for testing plugin performance. Separately, an Opus 5.5 user reports using 43% of their weekly allowance after almost a week of hours-long daily use.
TLDR
A post announcing Claude Code’s “claude plugin eval” says it runs a plugin against test cases, scores the results, then reruns each case without the plugin for comparison. The author suggests checking whether plugins still earn their keep after Opus 5.5’s release. Separately, one Opus 5.5 user praises its efficiency, reporting that heavy use for hours a day over almost a week consumed 43% of their weekly usage allowance.
Combined views
—
2 Sources, first seen ago
