HomePeopleCompaniesAI ModelsOpen SourceAgentsResearchApps
AllFrameworksCoding agentsProtocolsProducts
AgentsCoding agentsAnthropic 9 Jul 2026 aifirst-review

Claude Code 2.1.205 review: fabricated-approval prevention and full /doctor

The brief

2.1.205 adds an auto-mode rule blocking session transcript tampering, turns /doctor into a full diagnostic, fixes structured output on invalid schemas, a Windows worktree deletion bug via NTFS junctions, and.

Read the original GitHub

Sources