Source Intelligence

DisclaimerUnofficial, and not affiliated with Anthropic. Nearly all of this is read straight out of what ships: npm bundles, captured prompts, published docs. Anthropic's own notes go in verbatim, marked as theirs. The rest is my reading, and every entry carries the strings behind it. If one looks wrong, vote it down and say why.

All releases Home olderv0.2.62 v0.2.65newer

Claude Code v0.2.64

5 entries read diff v0.2.62 → v0.2.64 Markdown
Find
Pick an entry · j / k steps through
5 entries

Changesopen

New Features#

Tool Call Counter

A new visual indicator that displays the number of tool calls being made during interactions. This can be enabled through the showToolCallCount configuration option, helping users track Claude's tool usage in real-time.

ULTRACLAUDE.md Support (Experimental)

Introduced experimental support for a new user preferences file called ULTRACLAUDE.md. This file:

  • Lives in the .claude directory alongside other configuration files
  • Provides critical user preferences that override default behavior
  • Includes a warning when the file exceeds 1000 characters
  • Can be edited using the /memory command

Example warning display:

⚠️ ULTRACLAUDE.md exceeds 1000 chars (1234 chars) • /memory to edit

Enhanced Permission Debug Info

New permission debugging display that shows detailed information about permission decisions:

  • Rule-based decisions with their sources (CLI argument, local settings, or project settings)
  • Mode information (Default, Accept Edits, or Bypass Permissions)
  • Subcommand results with visual checkmarks (✓) or crosses (✗)
  • Hierarchical display of permission decision reasons

Example output:

Permission debug info:
✓ read-file
  ⎿ allow rule from CLI argument
✗ write-file
  ⎿ Default mode

Improvements#

Enhanced Instruction System

  • Updated the instruction preamble from "Codebase-specific instructions" to more authoritative language
  • New critical user preferences reminder system that seamlessly injects user preferences into conversations
  • Preferences are automatically appended to user messages without disrupting the conversation flow

️ Stream Processing Enhancement

The stream event handler now includes an optional callback for tool use detection, allowing the UI to respond immediately when Claude begins using tools.

Technical Changes#

New Utility Functions

  • Added array utility functions for improved performance:
  • includes() implementation with optimized search
  • uniqBy() for removing duplicates with custom comparator
  • Enhanced array searching with NaN handling

Import Reorganization

  • Migrated from direct stream imports to named imports (PassThrough)
  • Updated process and OS imports to use modern Node.js module syntax
  • Better organization of platform-specific imports (EOL, platform, homedir)

Related

Other releases about the same thing. Found by shared names or similar wording; neither means one caused the other.

Bug Fixes#

  • Fixed potential issues with instruction content injection by properly handling edge cases in message arrays
  • Improved error handling for EACCES permission errors with better telemetry

Related

Other releases about the same thing. Found by shared names or similar wording; neither means one caused the other.

Configuration#

New configuration option added to the animation settings:

{
  useHaiku: false,
  haikuInterval: 5,
  thinking: "default",
  responding: "default", 
  toolUse: "default",
  normal: "default",
  charAnimation: "none",
  showToolCallCount: false  // New option
}
Verbatim
No official entry

Not individually listed upstream

v0.2.64 does not have its own entry in Anthropic’s official changelog, because prerelease builds are often folded into a neighbouring release. Browse the full changelog for the closest official notes. Everything else on this page came out of the bundle instead, which is why the two lists don't match.

System prompt

The system prompt was not captured for this release, so this page cannot say whether it moved.