[Bug] Degradation in Claude model output quality and reasoning capability

Resolved 💬 3 comments Opened Mar 9, 2026 by Vivallo04 Closed Mar 13, 2026

Bug Description
There is a huge regression in Claude's output and intelligence....

Environment Info

  • Platform: darwin
  • Terminal: ghostty
  • Version: 2.1.71
  • Feedback ID: fd184f1c-03fb-41a9-92df-9715d61b28e5

Errors

[{"error":"Error: Plugin rust-analyzer-lsp has an invalid manifest file at /Users/vivallo/.claude/plugins/cache/claude-plugins-official/rust-analyzer-lsp/1.0.0/.claude-plugin/plugin.json.\n\nValidation errors: Unrecognized keys: \"source\", \"category\", \"strict\"\n    at gyq (/$bunfs/root/src/entrypoints/cli.js:2152:15375)\n    at async kR7 (/$bunfs/root/src/entrypoints/cli.js:2156:780)\n    at async zf$ (/$bunfs/root/src/entrypoints/cli.js:2156:9570)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-09T12:28:40.670Z"},{"error":"Error\n    at call (/$bunfs/root/src/entrypoints/cli.js:1461:5976)\n    at processTicksAndRejections (unknown:7:39)","timestamp":"2026-03-09T14:57:19.820Z"}]

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗