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 of v2.1.74 Home All releases olderv2.1.73 v2.1.75newer
Claude Code v2.1.74

The changelog for version 2.1.74 is complete. Key highlights include:

  • 6 new features: Context usage analysis system, custom auto-memory directories, AI-generated session titles, /stuck diagnostic skill, session-end hooks timeout config, and assistant-mode auto-backgrounding
  • 12 improvements: Flexible agent model IDs, improved auto-mode description, smarter memory selection, bidirectional text rendering, better command parsing, git submodule auto-init, and more
  • 8 bug fixes: OAuth error normalization, abort controller leaks, argument parsing edge cases, case-insensitive model matching, team cleanup, and system message preservation
  • 3 in-development features: Context collapse (Hawthorn), plugin dependency resolution (Orchid Trellis), and per-message budget persistence

See this entry in the whole of v2.1.74 →