최고의 tui 기반 에이전틱 코딩 도구인 claude code 가 지난 일주일간 업데이트한 내역


Version 2.1.41:
     • Added guard against launching Claude Code inside another Claude Code session
     • Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
     • Fixed a crash when MCP tools return image content during streaming
     • Fixed /resume session previews showing raw XML tags instead of readable command names
     • Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
     • Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
     • Fixed hook blocking errors (exit code 2) not showing stderr to the user
     • Added speed attribute to OTel events and trace spans for fast mode visibility
     • Added claude auth login, claude auth status, and claude auth logout CLI subcommands
     • Added Windows ARM64 (win32-arm64) native binary support
     • Improved /rename to auto-generate session name from conversation context when called without arguments
     • Improved narrow terminal layout for prompt footer
     • Fixed file resolution failing for @-mentions with anchor fragments (e.g., @README.md#installation)
     • Fixed FileReadTool blocking the process on FIFOs, /dev/stdin, and large files
     • Fixed background task notifications not being delivered in streaming Agent SDK mode
     • Fixed cursor jumping to end on each keystroke in classifier rule input
     • Fixed markdown link display text being dropped for raw URL
     • Fixed auto-compact failure error notifications being shown to users
     • Fixed permission wait time being included in subagent elapsed time display
     • Fixed proactive ticks firing while in plan mode
     • Fixed clear stale permission rules when settings change on disk
     • Fixed hook blocking errors showing stderr content in UI

     Version 2.1.42:
     • Improved startup performance by deferring Zod schema construction
     • Improved prompt cache hit rates by moving date out of system prompt
     • Added one-time Opus 4.6 effort callout for eligible users
     • Fixed /resume showing interrupt messages as session titles
     • Fixed image dimension limit errors to suggest /compact

     Version 2.1.43:
     • Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
     • Fixed spurious warnings for non-agent markdown files in .claude/agents/ directory
     • Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock

     Version 2.1.44:
     • Fixed ENAMETOOLONG errors for deeply-nested directory paths
     • Fixed auth refresh errors

     Version 2.1.45:
     • Added support for Claude Sonnet 4.6
     • Added support for reading enabledPlugins and extraKnownMarketplaces from --add-dir directories
     • Added spinnerTipsOverride setting to customize spinner tips — configure tips with an array of custom tip strings, and optionally set excludeDefault:
     true to show only your custom tips instead of the built-in ones
     • Added SDKRateLimitInfo and SDKRateLimitEvent types to the SDK, enabling consumers to receive rate limit status updates including utilization, reset
     times, and overage information
     • Fixed Agent Teams teammates failing on Bedrock, Vertex, and Foundry by propagating API provider environment variables to tmux-spawned processes
     (anthropics/claude-code#23561)
     • Fixed sandbox "operation not permitted" errors when writing temporary files on macOS by using the correct per-user temp directory
     (anthropics/claude-code#21654)
     • Fixed Task tool (backgrounded agents) crashing with a ReferenceError on completion (anthropics/claude-code#22087)
     • Fixed autocomplete suggestions not being accepted on Enter when images are pasted in the input
     • Fixed skills invoked by subagents incorrectly appearing in main session context after compaction
     • Fixed excessive .claude.json.backup files accumulating on every startup
     • Fixed plugin-provided commands, agents, and hooks not being available immediately after installation without requiring a restart
     • Improved startup performance by removing eager loading of session history for stats caching
     • Improved memory usage for shell commands that produce large output — RSS no longer grows unboundedly with command output size
     • Improved collapsed read/search groups to show the current file or search pattern being processed beneath the summary line while active
     • [VSCode] Improved permission destination choice (project/user/session) to persist across sessions

     Version 2.1.46:
     • Fixed orphaned CC processes after terminal disconnect on macOS
     • Added support for using claude.ai MCP connectors in Claude Code

     Version 2.1.47:
     • Fixed FileWriteTool line counting to preserve intentional trailing blank lines instead of stripping them with trimEnd().
     • Fixed Windows terminal rendering bugs caused by os.EOL (\r\n) in display code — line counts now show correct values instead of always showing 1 on
     Windows.
     • Improved VS Code plan preview: auto-updates as Claude iterates, enables commenting only when the plan is ready for review, and keeps the preview open
      when rejecting so Claude can revise.
     • Fixed a bug where bold and colored text in markdown output could shift to the wrong characters on Windows due to \r\n line endings.
     • Fixed compaction failing when conversation contains many PDF documents by stripping document blocks alongside images before sending to the compaction
      API (anthropics/claude-code#26188)
     • Improved memory usage in long-running sessions by releasing API stream buffers, agent context, and skill state after use
     • Improved startup performance by deferring SessionStart hook execution, reducing time-to-interactive by ~500ms.
     • Fixed an issue where bash tool output was silently discarded on Windows when using MSYS2 or Cygwin shells.
     • Improved performance of @ file mentions - file suggestions now appear faster by pre-warming the index on startup and using session-based caching with
      background refresh.
     • Improved memory usage by trimming agent task message history after tasks complete
     • Improved memory usage during long agent sessions by eliminating O(n²) message accumulation in progress updates
     • Fixed the bash permission classifier to validate that returned match descriptions correspond to actual input rules, preventing hallucinated
     descriptions from incorrectly granting permissions
     • Fixed user-defined agents only loading one file on NFS/FUSE filesystems that report zero inodes (anthropics/claude-code#26044)
     • Fixed plugin agent skills silently failing to load when referenced by bare name instead of fully-qualified plugin name (anthropics/claude-code#25834)
     • Search patterns in collapsed tool results are now displayed in quotes for clarity
     • Windows: Fixed CWD tracking temp files never being cleaned up, causing them to accumulate indefinitely (anthropics/claude-code#17600)
     • Use ctrl+f to kill all background agents instead of double-pressing ESC. Background agents now continue running when you press ESC to cancel the main
      thread, giving you more control over agent lifecycle.
     • Fixed API 400 errors ("thinking blocks cannot be modified") that occurred in sessions with concurrent agents, caused by interleaved streaming content
      blocks preventing proper message merging.
     • Simplified teammate navigation to use only Shift+Down (with wrapping) instead of both Shift+Up and Shift+Down.
     • Fixed an issue where a single file write/edit error would abort all other parallel file write/edit operations. Independent file mutations now
     complete even when a sibling fails.
     • Added last_assistant_message field to Stop and SubagentStop hook inputs, providing the final assistant response text so hooks can access it without
     parsing transcript files.
     • Fixed custom session titles set via /rename being lost after resuming a conversation (anthropics/claude-code#23610)
     • Fixed collapsed read/search hint text overflowing on narrow terminals by truncating from the start.
     • Fixed an issue where bash commands with backslash-newline continuation lines (e.g., long commands split across multiple lines with \) would produce
     spurious empty arguments, potentially breaking command execution.
     • Fixed built-in slash commands (/help, /model, /compact, etc.) being hidden from the autocomplete dropdown when many user skills are installed
     (anthropics/claude-code#22020)
     • Fixed MCP servers not appearing in the MCP Management Dialog after deferred loading
     • Fixed session name persisting in status bar after /clear command (anthropics/claude-code#26082)
     • Fixed crash when a skill's name or description in SKILL.md frontmatter is a bare number (e.g., name: 3000) — the value is now properly coerced to a
     string (anthropics/claude-code#25837)
     • Fixed /resume silently dropping sessions when the first message exceeds 16KB or uses array-format content (anthropics/claude-code#25721)
     • Added chat:newline keybinding action for configurable multi-line input (anthropics/claude-code#26075)
     • Added added_dirs to the statusline JSON workspace section, exposing directories added via /add-dir to external scripts (anthropics/claude-code#26096)
     • Fixed claude doctor misclassifying mise and asdf-managed installations as native installs (anthropics/claude-code#26033)
     • Fixed zsh heredoc failing with "read-only file system" error in sandboxed commands (anthropics/claude-code#25990)
     • Fixed agent progress indicator showing inflated tool use count (anthropics/claude-code#26023)
     • Fixed image pasting not working on WSL2 systems where Windows copies images as BMP format (anthropics/claude-code#25935)
     • Fixed background agent results returning raw transcript data instead of the agent's final answer (anthropics/claude-code#26012)
     • Fixed Warp terminal incorrectly prompting for Shift+Enter setup when it supports it natively (anthropics/claude-code#25957)
     • Fixed CJK wide characters causing misaligned timestamps and layout elements in the TUI (anthropics/claude-code#26084)
     • Fixed custom agent model field in .claude/agents/*.md being ignored when spawning team teammates (anthropics/claude-code#26064)
     • Fixed plan mode being lost after context compaction, causing the model to switch from planning to implementation mode (anthropics/claude-code#26061)
     • Fixed alwaysThinkingEnabled: true in settings.json not enabling thinking mode on Bedrock and Vertex providers (anthropics/claude-code#26074)
     • Fixed tool_decision OTel telemetry event not being emitted in headless/SDK mode (anthropics/claude-code#26059)
     • Fixed session name being lost after context compaction — renamed sessions now preserve their custom title through compaction
     (anthropics/claude-code#26121)
     • Increased initial session count in resume picker from 10 to 50 for faster session discovery (anthropics/claude-code#26123)
     • Windows: fixed worktree session matching when drive letter casing differs (anthropics/claude-code#26123)
     • Fixed /resume <session-id> failing to find sessions whose first message exceeds 16KB (anthropics/claude-code#25920)
     • Fixed "Always allow" on multiline bash commands creating invalid permission patterns that corrupt settings (anthropics/claude-code#25909)
     • Fixed React crash (error #31) when a skill's argument-hint in SKILL.md frontmatter uses YAML sequence syntax (e.g., [topic: foo | bar]) — the value
     is now properly coerced to a string (anthropics/claude-code#25826)
     • Fixed crash when using /fork on sessions that used web search — null entries in search results from transcript deserialization are now handled
     gracefully (anthropics/claude-code#25811)
     • Fixed read-only git commands triggering FSEvents file watcher loops on macOS by adding --no-optional-locks flag (anthropics/claude-code#25750)
     • Fixed custom agents and skills not being discovered when running from a git worktree — project-level .claude/agents/ and .claude/skills/ from the
     main repository are now included (anthropics/claude-code#25816)
     • Fixed non-interactive subcommands like claude doctor and claude plugin validate being blocked inside nested Claude sessions
     (anthropics/claude-code#25803)
     • Windows: Fixed the same CLAUDE.md file being loaded twice when drive letter casing differs between paths (anthropics/claude-code#25756)
     • Fixed inline code spans in markdown being incorrectly parsed as bash commands (anthropics/claude-code#25792)
     • Fixed teammate spinners not respecting custom spinnerVerbs from settings (anthropics/claude-code#25748)
     • Fixed shell commands permanently failing after a command deletes its own working directory (anthropics/claude-code#26136)
     • Fixed hooks (PreToolUse, PostToolUse) silently failing to execute on Windows by using Git Bash instead of cmd.exe (anthropics/claude-code#25981)
     • Fixed LSP findReferences and other location-based operations returning results from gitignored files (e.g., node_modules/, venv/)
     (anthropics/claude-code#26051)
     • Moved config backup files from home directory root to ~/.claude/backups/ to reduce home directory clutter (anthropics/claude-code#26130)
     • Fixed sessions with large first prompts (>16KB) disappearing from the /resume list (anthropics/claude-code#26140)
     • Fixed shell functions with double-underscore prefixes (e.g., __git_ps1) not being preserved across shell sessions (anthropics/claude-code#25824)
     • Fixed spinner showing "0 tokens" counter before any tokens have been received (anthropics/claude-code#26105)
     • VSCode: Fixed conversation messages appearing dimmed while the AskUserQuestion dialog is open (anthropics/claude-code#26078)
     • Fixed background tasks failing in git worktrees due to remote URL resolution reading from worktree-specific gitdir instead of the main repository
     config (anthropics/claude-code#26065)
     • Fixed Right Alt key leaving visible [25~ escape sequence residue in the input field on Windows/Git Bash terminals (anthropics/claude-code#25943)
     • The /rename command now updates the terminal tab title by default (anthropics/claude-code#25789)
     • Fixed Edit tool silently corrupting Unicode curly quotes (\u201c\u201d \u2018\u2019) by replacing them with straight quotes when making edits
     (anthropics/claude-code#26141)
     • Fixed OSC 8 hyperlinks only being clickable on the first line when link text wraps across multiple terminal lines.

     Version 2.1.49:
     • Added --worktree (-w) flag to start Claude in an isolated git worktree
     • Subagents support isolation: "worktree" for working in a temporary git worktree
     • Added Ctrl+F keybinding to kill background agents (two-press confirmation)
     • Agent definitions support background: true to always run as a background task
     • Plugins can ship settings.json for default configuration
     • Fixed file-not-found errors to suggest corrected paths when the model drops the repo folder
     • Fixed Ctrl+C and ESC being silently ignored when background agents are running and the main thread is idle. Pressing twice within 3 seconds now kills
      all background agents.
     • Fixed prompt suggestion cache regression that reduced cache hit rates.
     • Fixed plugin enable and plugin disable to auto-detect the correct scope when --scope is not specified, instead of always defaulting to user scope
     • Simple mode (CLAUDE_CODE_SIMPLE) now includes the file edit tool in addition to the Bash tool, allowing direct file editing in simple mode.
     • Permission suggestions are now populated when safety checks trigger an ask response, enabling SDK consumers to display permission options
     • Sonnet 4.5 with 1M context is being removed from the Max plan in favor of our frontier Sonnet 4.6 model, which now has 1M context. Please switch in
     /model.
     • Fixed verbose mode not updating thinking block display when toggled via /config — memo comparators now correctly detect verbose changes
     • Fixed unbounded WASM memory growth during long sessions by periodically resetting the tree-sitter parser
     • Fixed potential rendering issues caused by stale yoga layout references
     • Improved performance in non-interactive mode (-p) by skipping unnecessary API calls during startup
     • Improved performance by caching authentication failures for HTTP and SSE MCP servers, avoiding repeated connection attempts to servers requiring auth
     • Fixed unbounded memory growth during long-running sessions caused by Yoga WASM linear memory never shrinking
     • SDK model info now includes supportsEffort, supportedEffortLevels, and supportsAdaptiveThinking fields so consumers can discover model capabilities.
     • Added ConfigChange hook event that fires when configuration files change during a session, enabling enterprise security auditing and optional
     blocking of settings changes.
     • Improved startup performance by caching MCP auth failures to avoid redundant connection attempts
     • Improved startup performance by reducing HTTP calls for analytics token counting
     • Improved startup performance by batching MCP tool token counting into a single API call
     • Fixed disableAllHooks setting to respect managed settings hierarchy — non-managed settings can no longer disable managed hooks set by policy (#26637)
     • Fixed --resume session picker showing raw XML tags for sessions that start with commands like /clear. Now correctly falls through to the session ID
     fallback.
     • Improved permission prompts for path safety and working directory blocks to show the reason for the restriction instead of a bare prompt with no
     context



1년동안 업데이트 할 만한 분량을 7일동안 한 느낌을 받았음.
이게 ai 프론티어 업체(앤트로픽) 의 개발속도임. 
AI 가 AI 를 개선하고 수정하는 loop 즉, 자기 개선 work flow 를 구축한것으로 보임.
메인 개발자의 개입이 있겠지만, 아마 코딩은 1도 안했을꺼고 중요한 부분 리뷰만 해주었을거라고 짐작.
-----
이걸 어떻게  사람이 따라감?