Version 1.0.71

What: Customizable status line at the bottom of your terminal showing contextual information Why: Provides at-a-glance information about your current session, model, and working directory How to use:

# Configure in ~/.claude/settings.json
{
  "statusLine": {
    "type": "command",
    "command": "cat | jq -r '\"\\(.model.display_name) in \\(.workspace.current_dir)\"'",
    "padding": 1
  }
}

Details:

• Receives JSON input with session_id, model info, workspace paths, and more
• Updates every 300ms (debounced) to show current context
• Can display any command output, limited to first line
• Supports ANSI color codes for custom styling

What: New specialized agent that converts your shell PS1 prompt to Claude Code statusLine Why: Makes it easy to replicate your familiar shell prompt in Claude Code without manual configuration How to use:

# Ask Claude to set up your statusLine
"Use the statusline-setup agent to convert my PS1 to a statusLine"

# Or use it directly
/task statusline-setup "Convert my shell prompt"

Details:

• Automatically reads ~/.zshrc, ~/.bashrc, ~/.bash_profile, or ~/.profile
• Converts PS1 escape sequences (\\u → $(whoami), \\w → $(pwd), etc.)
• Preserves your ANSI color codes
• Updates settings.json automatically
• Removes trailing $ or > characters from output

What: New setting to disable all hooks globally Why: Provides an escape hatch when hooks are causing issues or impacting performance How to use:

// In ~/.claude/settings.json
{
  "disableAllHooks": true
}

Details:

• Disables all hook executions including statusLine updates
• Useful for debugging or performance optimization
• Logs a message when hooks are skipped

What: Detection and tracking of simultaneous query executions Why: Prevents race conditions and undefined behavior from rapid command execution Details:

• Automatically detects when multiple queries run simultaneously
• Logs telemetry event tengu_concurrent_onquery_detected
• Helps identify problematic usage patterns

What changed: Complete rewrite of memory saving from AI-based merging to direct file append Impact: Memory operations are now 10-100x faster and much more reliable

What changed: Switched from full module imports to specific function imports Impact: Reduced bundle size and improved startup performance

What changed: New memory and context management thresholds Impact: Better handling of large contexts and improved memory efficiency Details:

• Updated context size limits to 13000/20000 tokens
• Improved memory management for long-running sessions

• Issue: StatusLine feature always returned false, making it completely unusable
• Cause: Logic error in feature detection function that ignored configuration
• Resolution: Properly checks for statusLine configuration and enables feature
• Affected users: Anyone trying to use statusLine in v1.0.70

• Issue: Memory additions would sometimes fail with "No tool use found" errors
• Cause: Complex AI-based approach that required perfect tool responses
• Resolution: Simplified to direct file append operations
• Affected users: Users frequently saving memories

• Issue: Multiple simultaneous queries could cause undefined behavior
• Cause: No protection against concurrent execution
• Resolution: Added detection and tracking of concurrent queries
• Affected users: Users triggering rapid commands or using automation

• Issue: Importing entire modules when only specific functions were needed
• Resolution: Changed to named imports for crypto, stream, and process modules
• Affected users: All users benefit from smaller bundle size

• Removed deprecated memory management functions (AuB, lhB, l$8)
• Cleaned up unused variables (uw2, ZG8, GG8, FG8)
• Reorganized GitHub Actions workflow variables for consistency

• Switched to named imports for better tree-shaking
• Updated memory thresholds for improved performance
• Added context management feature flag for future API support