Claude Code · Source-level changelog

Version 0.2.72

Package @anthropic-ai/claude-code Diff v0.2.70 → v0.2.72
9
Features & Changes
3
Bug Fixes
0
In Development
0
Env Vars / Flags

Changes

9 items

# Claude Code v0.2.72 Changelog

Parallel Agent Execution #

  • New command: /parallel-agents [on|off|count] - Toggle parallel agent execution or set the number of agents to run concurrently
  • parallel-agents on - Enable with default 3 agents
  • parallel-agents off - Disable parallel execution
  • parallel-agents 5 - Enable with 5 parallel agents
  • parallel-agents or parallel-agents status - Show current status
  • Note: Using parallel agents will increase token usage by approximately (count + 1)x due to the synthesis agent

Enhanced Progress Indicators #

  • Offline indicator: Progress spinner now shows "offline" status when internet connection is lost
  • Network connectivity monitoring: Real-time detection of internet connectivity with visual feedback
  • Mode-specific spinners: Different visual indicators for different operations:
  • for tool use
  • for responding/thinking
  • for requesting

Improved File Size Display #

  • New human-readable file size formatting (e.g., "1.5MB" instead of "1572864 bytes")
  • Automatic unit conversion: bytes → KB → MB → GB

Streamlined Command Output Display #

  • Simplified display of command execution results
  • Cleaner separation of stdout and stderr output
  • More concise "No output" message when commands produce no output

Better Progress Tracking #

  • Token counting now updates smoothly during generation
  • Elapsed time counter with 10ms precision
  • Improved animation timing for character counting

Enhanced Settings Display #

  • Local settings now show: On this machine in <path>
  • Project settings now show: Checked in at .claude/settings.json
  • Clearer distinction between local and project-wide settings

Parallel Agent Architecture #

When enabled, the Task tool can now:

  1. Spawn multiple agents to work on different aspects of a problem simultaneously
  2. Collect results from all agents
  3. Use a synthesis agent to combine findings into a cohesive response
  4. Display progress for each agent separately in the UI

Example output with parallel agents:

3 total tool uses across 3 agents

Agent 1:
  [tool use progress]

Agent 2:
  [tool use progress]

Synthesis (combining results):
  [synthesis progress]

Stream Event Handling #

  • New stream_request_start event type for better request lifecycle tracking
  • Improved mode detection for different content block types
  • More granular control over UI updates during streaming

Removed Legacy Features #

  • Removed persistent shell functionality and related classes
  • Removed unused spinner configurations and animation styles
  • Cleaned up deprecated utility functions

Bug Fixes

3 items
  • Fixed duplicate import statements
  • Improved error handling for offline scenarios
  • Better cleanup of temporary resources