Agents gain per-agent model selection for mixed-model pipelines; hooks add systemMessage for warnings; hidden files now appear in search and @-mentions.
Changes
5
Total Changes (vs Average)
0.4×
Version Type
Patch
Release Focus
Mixed
Day Since Prior Release
1
Release # This Month
#22 of 23
Bug Fix Ratio
20%
Post-Release Edits
—
Change Category
Ecosystem Impact
Windows featured in 4 of the last 10 releases, including the immediately prior v1.0.63 patch. This release returns to General-only changes, concentrating on agent infrastructure and the hooks extension system.
Key Themes
2Agent System Maturation|1Hooks Output Expansion|2Discovery & SDK Reliability
May Help If You've Noticed
Your agents all use the same model but you need different models for different agent roles in a pipeline
Your hook scripts can't surface warnings or context messages back to the user during execution
Hidden dotfiles or config files don't show up when searching or typing @ to reference files
An agent unexpectedly invokes another agent recursively in ways you didn't intend
Multi-turn conversations via the SDK aren't correctly attributing which messages came from the user
Changes in Claude Code v1.0.64
Added1 change
•Added hidden files to file search and @-mention suggestions
Fixed1 change
•SDK: Fixed user input tracking across multi-turn conversations
Other3 changes
•Agents: Added model customization support - you can now specify which model an agent should use
•Agents: Fixed unintended access to the recursive agent tool
•Hooks: Added systemMessage field to hook JSON output for displaying warnings and context