spacehub

surshar / claude-code

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

Built by Anthropic · github.com/anthropics

claude-code / CHANGELOG.md
GitHub Actionschore: Update CHANGELOG.md 4523c00 · on Feb 17, 2026 History
1,730 lines · 91.9 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.45
4
5- Added support for Claude Sonnet 4.6
6- Added support for reading `enabledPlugins` and `extraKnownMarketplaces` from `--add-dir` directories
7- 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
8- Added `SDKRateLimitInfo` and `SDKRateLimitEvent` types to the SDK, enabling consumers to receive rate limit status updates including utilization, reset times, and overage information
9- Fixed Agent Teams teammates failing on Bedrock, Vertex, and Foundry by propagating API provider environment variables to tmux-spawned processes (anthropics/claude-code#23561)
10- Fixed sandbox "operation not permitted" errors when writing temporary files on macOS by using the correct per-user temp directory (anthropics/claude-code#21654)
11- Fixed Task tool (backgrounded agents) crashing with a `ReferenceError` on completion (anthropics/claude-code#22087)
12- Fixed autocomplete suggestions not being accepted on Enter when images are pasted in the input
13- Fixed skills invoked by subagents incorrectly appearing in main session context after compaction
14- Fixed excessive `.claude.json.backup` files accumulating on every startup
15- Fixed plugin-provided commands, agents, and hooks not being available immediately after installation without requiring a restart
16- Improved startup performance by removing eager loading of session history for stats caching
17- Improved memory usage for shell commands that produce large output — RSS no longer grows unboundedly with command output size
18- Improved collapsed read/search groups to show the current file or search pattern being processed beneath the summary line while active
19- [VSCode] Improved permission destination choice (project/user/session) to persist across sessions
20
21## 2.1.44
22
23- Fixed ENAMETOOLONG errors for deeply-nested directory paths
24- Fixed auth refresh errors
25
26## 2.1.43
27
28- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
29- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
30- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
31
32## 2.1.42
33
34- Improved startup performance by deferring Zod schema construction
35- Improved prompt cache hit rates by moving date out of system prompt
36- Added one-time Opus 4.6 effort callout for eligible users
37- Fixed /resume showing interrupt messages as session titles
38- Fixed image dimension limit errors to suggest /compact
39
40## 2.1.41
41
42- Added guard against launching Claude Code inside another Claude Code session
43- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
44- Fixed a crash when MCP tools return image content during streaming
45- Fixed /resume session previews showing raw XML tags instead of readable command names
46- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
47- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
48- Fixed hook blocking errors (exit code 2) not showing stderr to the user
49- Added `speed` attribute to OTel events and trace spans for fast mode visibility
50- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
51- Added Windows ARM64 (win32-arm64) native binary support
52- Improved `/rename` to auto-generate session name from conversation context when called without arguments
53- Improved narrow terminal layout for prompt footer
54- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
55- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
56- Fixed background task notifications not being delivered in streaming Agent SDK mode
57- Fixed cursor jumping to end on each keystroke in classifier rule input
58- Fixed markdown link display text being dropped for raw URL
59- Fixed auto-compact failure error notifications being shown to users
60- Fixed permission wait time being included in subagent elapsed time display
61- Fixed proactive ticks firing while in plan mode
62- Fixed clear stale permission rules when settings change on disk
63- Fixed hook blocking errors showing stderr content in UI
64
65## 2.1.39
66
67- Improved terminal rendering performance
68- Fixed fatal errors being swallowed instead of displayed
69- Fixed process hanging after session close
70- Fixed character loss at terminal screen boundary
71- Fixed blank lines in verbose transcript view
72
73## 2.1.38
74
75- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
76- Fixed Tab key queueing slash commands instead of autocompleting
77- Fixed bash permission matching for commands using environment variable wrappers
78- Fixed text between tool uses disappearing when not using streaming
79- Fixed duplicate sessions when resuming in VS Code extension
80- Improved heredoc delimiter parsing to prevent command smuggling
81- Blocked writes to `.claude/skills` directory in sandbox mode
82
83## 2.1.37
84
85- Fixed an issue where /fast was not immediately available after enabling /extra-usage
86
87## 2.1.36
88
89- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
90
91## 2.1.34
92
93- Fixed a crash when agent teams setting changed between renders
94- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
95
96## 2.1.33
97
98- Fixed agent teammate sessions in tmux to send and receive messages
99- Fixed warnings about agent teams not being available on your current plan
100- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
101- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
102- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
103- Added plugin name to skill descriptions and `/skills` menu for better discoverability
104- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
105- Fixed an API error that could occur when aborting mid-stream, where whitespace text combined with a thinking block would bypass normalization and produce an invalid request
106- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
107- Fixed an issue where proxy settings configured via `settings.json` environment variables were not applied to WebFetch and other HTTP requests on the Node.js build
108- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
109- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
110- Errors from invalid managed settings are now surfaced
111- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
112- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
113- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
114
115## 2.1.32
116
117- Claude Opus 4.6 is now available!
118- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
119- Claude now automatically records and recalls memories as it works
120- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
121- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
122- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
123- Updated --resume to re-use --agent value specified in previous conversation by default.
124- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
125- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
126- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
127- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
128- VSCode: Added spinner when loading past conversations list
129
130## 2.1.31
131
132- Added session resume hint on exit, showing how to continue your conversation later
133- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
134- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
135- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
136- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
137- Fixed `temperatureOverride` being silently ignored in the streaming API path, causing all streaming requests to use the default temperature (1) regardless of the configured override
138- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
139- Improved system prompts to more clearly guide the model toward using dedicated tools (Read, Edit, Glob, Grep) instead of bash equivalents (`cat`, `sed`, `grep`, `find`), reducing unnecessary bash command usage
140- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
141- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
142- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
143
144## 2.1.30
145
146- Added `pages` parameter to the Read tool for PDFs, allowing specific page ranges to be read (e.g., `pages: "1-5"`). Large PDFs (>10 pages) now return a lightweight reference when `@` mentioned instead of being inlined into context.
147- Added pre-configured OAuth client credentials for MCP servers that don't support Dynamic Client Registration (e.g., Slack). Use `--client-id` and `--client-secret` with `claude mcp add`.
148- Added `/debug` for Claude to help troubleshoot the current session
149- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
150- Added token count, tool uses, and duration metrics to Task tool results
151- Added reduced motion mode to the config
152- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
153- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
154- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
155- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
156- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
157- Fixed permission dialogs stealing focus while actively typing
158- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
159- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
160- Improved memory usage for `--resume` (68% reduction for users with many sessions) by replacing the session index with lightweight stat-based loading and progressive enrichment
161- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
162- Changed `/model` to execute immediately instead of being queued
163- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
164- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
165
166## 2.1.29
167
168- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
169
170## 2.1.27
171
172- Added tool call failures and denials to debug logs
173- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
174- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
175- Sessions are now automatically linked to PRs when created via `gh pr create`
176- Fixed /context command not displaying colored output
177- Fixed status bar duplicating background task indicator when PR status was shown
178- Windows: Fixed bash command execution failing for users with `.bashrc` files
179- Windows: Fixed console windows flashing when spawning child processes
180- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
181
182## 2.1.25
183
184- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
185
186## 2.1.23
187
188- Added customizable spinner verbs setting (`spinnerVerbs`)
189- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
190- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
191- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
192- Fixed pending async hooks not being cancelled when headless streaming sessions ended
193- Fixed tab completion not updating the input field when accepting a suggestion
194- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
195- Improved terminal rendering performance with optimized screen data layout
196- Changed Bash commands to show timeout duration alongside elapsed time
197- Changed merged pull requests to show a purple status indicator in the prompt footer
198- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
199
200## 2.1.22
201
202- Fixed structured outputs for non-interactive (-p) mode
203
204## 2.1.21
205
206- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
207- Fixed shell completion cache files being truncated on exit
208- Fixed API errors when resuming sessions that were interrupted during tool execution
209- Fixed auto-compact triggering too early on models with large output token limits
210- Fixed task IDs potentially being reused after deletion
211- Fixed file search not working in VS Code extension on Windows
212- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
213- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
214- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
215- [VSCode] Fixed message action buttons having incorrect background colors
216
217## 2.1.20
218
219- Added arrow key history navigation in vim normal mode when cursor cannot move further
220- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
221- Added PR review status indicator to the prompt footer, showing the current branch's PR state (approved, changes requested, pending, or draft) as a colored dot with a clickable link
222- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
223- Added ability to delete tasks via the `TaskUpdate` tool
224- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
225- Fixed agents sometimes ignoring user messages sent while actively working on a task
226- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
227- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
228- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
229- Fixed draft prompt being lost when pressing UP arrow to navigate command history
230- Fixed ghost text flickering when typing slash commands mid-input
231- Fixed marketplace source removal not properly deleting settings
232- Fixed duplicate output in some commands like `/context`
233- Fixed task list sometimes showing outside the main conversation view
234- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
235- Fixed crashes when cancelling tool use
236- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
237- Improved thinking status text with a subtle shimmer animation
238- Improved task list to dynamically adjust visible items based on terminal height
239- Improved fork conversation hint to show how to resume the original session
240- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
241- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
242- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
243- Changed the `/copy` command to be available to all users
244- Changed background agents to prompt for tool permissions before launching
245- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
246- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
247
248## 2.1.19
249
250- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
251- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
252- Fixed crashes on processors without AVX instruction support
253- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
254- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
255- Fixed resuming sessions by custom title not working when run from a different directory
256- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
257- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
258- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
259- Fixed file write preview omitting empty lines
260- Changed skills without additional permissions or hooks to be allowed without requiring approval
261- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
262- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
263- [VSCode] Enabled session forking and rewind functionality for all users
264
265## 2.1.18
266
267- Added customizable keyboard shortcuts. Configure keybindings per context, create chord sequences, and personalize your workflow. Run `/keybindings` to get started. Learn more at https://code.claude.com/docs/en/keybindings
268
269## 2.1.17
270
271- Fixed crashes on processors without AVX instruction support
272
273## 2.1.16
274
275- Added new task management system, including new capabilities like dependency tracking
276- [VSCode] Added native plugin management support
277- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
278- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
279- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
280- Fixed session titles on the resume screen not respecting the user's language setting
281- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
282
283## 2.1.15
284
285- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
286- Improved UI rendering performance with React Compiler
287- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
288- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
289
290## 2.1.14
291
292- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
293- Added search to installed plugins list - type to filter by name or description
294- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
295- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
296- Fixed memory issues that could cause crashes when running parallel subagents
297- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
298- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
299- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
300- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
301- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
302- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
303- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
304- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
305- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
306- Improved backspace to delete pasted text as a single token instead of one character at a time
307- [VSCode] Added `/usage` command to display current plan usage
308
309## 2.1.12
310
311- Fixed message rendering bug
312
313## 2.1.11
314
315- Fixed excessive MCP connection requests for HTTP/SSE transports
316
317## 2.1.10
318
319- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
320- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
321- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
322- Improved startup to capture keystrokes typed before the REPL is fully ready
323- Improved file suggestions to show as removable attachments instead of inserting text when accepted
324- [VSCode] Added install count display to plugin listings
325- [VSCode] Added trust warning when installing plugins
326
327## 2.1.9
328
329- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
330- Added `plansDirectory` setting to customize where plan files are stored
331- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
332- Added session URL attribution to commits and PRs created from web sessions
333- Added support for `PreToolUse` hooks to return `additionalContext` to the model
334- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
335- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
336- Fixed MCP server reconnection hanging when cached connection promise never resolves
337- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
338
339## 2.1.7
340
341- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
342- Added ability to provide feedback when accepting permission prompts
343- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
344- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
345- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
346- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
347- Fixed context window blocking limit being calculated using the full context window instead of the effective context window (which reserves space for max output tokens)
348- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
349- Fixed terminal title animation jitter by using fixed-width braille characters
350- Fixed plugins with git submodules not being fully initialized when installed
351- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
352- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
353- Enabled MCP tool search auto mode by default for all users. When MCP tool descriptions exceed 10% of the context window, they are automatically deferred and discovered via the MCPSearch tool instead of being loaded upfront. This reduces context usage for users with many MCP tools configured. Users can disable this by adding `MCPSearch` to `disallowedTools` in their settings.
354- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
355- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
356
357## 2.1.6
358
359- Added search functionality to `/config` command for quickly filtering settings
360- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
361- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
362- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
363- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
364- Added an error display when the editor fails during Ctrl+G
365- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
366- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
367- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
368- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
369- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
370- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
371- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
372- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
373- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
374- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
375- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
376- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
377- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
378- Improved @ autocomplete with icons for different suggestion types and single-line formatting
379- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
380- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
381- Changed terminal title to "Claude Code" on startup for better window identification
382- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
383- [VSCode] Fixed usage indicator not updating after manual compact
384
385## 2.1.5
386
387- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
388
389## 2.1.4
390
391- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
392- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
393
394## 2.1.3
395
396- Merged slash commands and skills, simplifying the mental model with no change in behavior
397- Added release channel (`stable` or `latest`) toggle to `/config`
398- Added detection and warnings for unreachable permission rules, with warnings in `/doctor` and after saving rules that include the source of each rule and actionable fix guidance
399- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
400- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
401- Fixed mismatch between background task count in status bar and items shown in tasks dialog
402- Fixed sub-agents using the wrong model during conversation compaction
403- Fixed web search in sub-agents using incorrect model
404- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
405- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
406- Improved slash command suggestion readability by truncating long descriptions to 2 lines
407- Changed tool hook execution timeout from 60 seconds to 10 minutes
408- [VSCode] Added clickable destination selector for permission requests, allowing you to choose where settings are saved (this project, all projects, shared with team, or session only)
409
410## 2.1.2
411
412- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
413- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
414- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
415- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
416- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
417- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
418- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
419- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
420- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
421- Fixed updates incorrectly claiming another installation is in progress
422- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
423- Fixed remote session URL and teleport being broken when using `/tasks` command
424- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
425- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
426- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
427- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
428- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
429- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
430- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
431- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
432- [SDK] Changed minimum zod peer dependency to ^4.0.0
433- [VSCode] Fixed usage display not updating after manual compact
434
435## 2.1.0
436
437- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
438- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
439- Added support for `agent` field in skills to specify agent type for execution
440- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
441- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
442- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
443- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
444- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
445- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
446- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
447- Fixed Esc key with queued prompts to only move them to input without canceling the running task
448- Reduced permission prompts for complex bash commands
449- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
450- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
451- Fixed plugin path resolution for file-based marketplace sources
452- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
453- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
454- Fixed Claude in Chrome support for WSL environments
455- Fixed Windows native installer silently failing when executable creation fails
456- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
457- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
458- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
459- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
460- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
461- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
462- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
463- Added hooks support for skill and slash command frontmatter
464- Added new Vim motions: `;` and `,` to repeat f/F/t/T motions, `y` operator for yank with `yy`/`Y`, `p`/`P` for paste, text objects (`iw`, `aw`, `iW`, `aW`, `i"`, `a"`, `i'`, `a'`, `i(`, `a(`, `i[`, `a[`, `i{`, `a{`), `>>` and `<<` for indent/dedent, and `J` to join lines
465- Added `/plan` command shortcut to enable plan mode directly from the prompt
466- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
467- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
468- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
469- Added support for `once: true` config for hooks
470- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
471- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
472- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
473- Added left/right arrow key navigation for cycling through tabs in dialogs
474- Added real-time thinking block display in Ctrl+O transcript mode
475- Added filepath to full output in background bash task details dialog
476- Added Skills as a separate category in the context visualization
477- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
478- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
479- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
480- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
481- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
482- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
483- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
484- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
485- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
486- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
487- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
488- Fixed commands with `$()` command substitution failing with parse errors
489- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
490- Fixed bash command prefix extraction to correctly identify subcommands after global options (e.g., `git -C /path log` now correctly matches `Bash(git log:*)` rules)
491- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
492- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
493- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
494- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
495- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
496- Fixed subagents sometimes not inheriting the parent's model by default
497- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
498- Fixed duplicate Bash commands appearing in permission request option labels
499- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
500- Fixed background task completion notifications to appear proactively with bullet point
501- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
502- Fixed cursor disappearing after dismissing permission dialogs
503- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
504- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
505- Fixed images being silently dropped when queueing messages while backgrounding a task
506- Fixed large pasted images failing with "Image was too large" error
507- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
508- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
509- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
510- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
511- Fixed spinner token counter to properly accumulate tokens from subagents during execution
512- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
513- Fixed race condition where LSP tool could return "no server available" during startup
514- Fixed feedback submission hanging indefinitely when network requests timeout
515- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
516- Fixed hook success message showing trailing colon when hook has no output
517- Multiple optimizations to improve startup performance
518- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
519- Improved performance when reading Jupyter notebooks with many cells
520- Improved reliability for piped input like `cat refactor.md | claude`
521- Improved reliability for AskQuestion tool
522- Improved sed in-place edit commands to render as file edits with diff preview
523- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
524- Improved compaction reliability
525- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
526- Improved skills to show progress while executing, displaying tool uses as they happen
527- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
528- Improved skill suggestions to prioritize recently and frequently used skills
529- Improved spinner feedback when waiting for the first response token
530- Improved token count display in spinner to include tokens from background agents
531- Improved incremental output for async agents to give the main thread more control and visibility
532- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
533- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
534- Improved macOS screenshot paste reliability with TIFF format support
535- Improved `/stats` output
536- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
537- Changed "Interrupted" message color from red to grey for a less alarming appearance
538- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
539- Removed underline styling from image reference links
540- [SDK] Changed minimum zod peer dependency to ^4.0.0
541- [VSCode] Added currently selected model name to the context menu
542- [VSCode] Added descriptive labels on auto-accept permission button (e.g., "Yes, allow npm for this project" instead of "Yes, and don't ask again")
543- [VSCode] Fixed paragraph breaks not rendering in markdown content
544- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
545- [Windows] Fixed issue with improper rendering
546
547## 2.0.76
548
549- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
550
551## 2.0.75
552
553- Minor bugfixes
554
555## 2.0.74
556
557- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
558- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
559- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
560- Added syntax highlighting info to theme picker
561- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
562- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
563- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
564- Fixed a potential crash when syntax highlighting isn't initialized correctly
565- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
566- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
567- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
568- [Windows] Fixed issue with improper rendering
569- [VSCode] Added gift tag pictogram for year-end promotion message
570
571## 2.0.73
572
573- Added clickable `[Image #N]` links that open attached images in the default viewer
574- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
575- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
576- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
577- Fixed slow input history cycling and race condition that could overwrite text after message submission
578- Improved `/theme` command to open theme picker directly
579- Improved theme picker UI
580- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
581- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
582
583## 2.0.72
584
585- Added Claude in Chrome (Beta) feature that works with the Chrome extension (https://claude.ai/chrome) to let you control your browser directly from Claude Code
586- Reduced terminal flickering
587- Added scannable QR code to mobile app tip for quick app downloads
588- Added loading indicator when resuming conversations for better feedback
589- Fixed `/context` command not respecting custom system prompts in non-interactive mode
590- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
591- Improved @ mention file suggestion speed (~3× faster in git repositories)
592- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
593- Improved settings validation errors to be more prominent
594- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
595
596## 2.0.71
597
598- Added /config toggle to enable/disable prompt suggestions
599- Added `/settings` as an alias for the `/config` command
600- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
601- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
602- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
603- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
604- New syntax highlighting engine for native build
605
606## 2.0.70
607
608- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
609- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
610- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
611- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
612- Fixed input being cleared when processing queued commands while the user was typing
613- Fixed prompt suggestions replacing typed input when pressing Tab
614- Fixed diff view not updating when terminal is resized
615- Improved memory usage by 3x for large conversations
616- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
617- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
618- Fix thinking mode toggle in /config not persisting correctly
619- Improve UI for file creation permission dialog
620
621## 2.0.69
622
623- Minor bugfixes
624
625## 2.0.68
626
627- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
628- Fixed a bug where disallowed MCP tools were visible to the model
629- Fixed an issue where steering messages could be lost while a subagent is working
630- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
631- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
632- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
633
634## 2.0.67
635
636- Thinking mode is now enabled by default for Opus 4.5
637- Thinking mode configuration has moved to /config
638- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
639- Show reason why autoupdater is disabled in `/doctor`
640- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
641- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
642- Fixed scroll position resetting after deleting a permission rule in `/permissions`
643- Fixed word deletion (opt+delete) and word navigation (opt+arrow) not working correctly with non-Latin text such as Cyrillic, Greek, Arabic, Hebrew, Thai, and Chinese
644- Fixed `claude install --force` not bypassing stale lock files
645- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
646- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
647
648## 2.0.65
649
650- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
651- Added context window information to status line input
652- Added `fileSuggestion` setting for custom `@` file search commands
653- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
654- Fixed prompt not being saved to history when aborting a query with Escape
655- Fixed Read tool image handling to identify format from bytes instead of file extension
656
657## 2.0.64
658
659- Made auto-compacting instant
660- Agents and bash commands can run asynchronously and send messages to wake up the main agent
661- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
662- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
663- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
664- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
665- Fixed auto-loading .env when using native installer
666- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
667- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
668- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
669- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
670- Bedrock: Improve efficiency of token counting
671- Bedrock: Add support for `aws login` AWS Management Console credentials
672- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
673
674## 2.0.62
675
676- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
677- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
678- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
679- Fixed slash command selection not working when multiple commands share the same name
680- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
681- Fixed running versions getting removed because lock file incorrectly going stale
682- Fixed IDE diff tab not closing when rejecting file changes
683
684## 2.0.61
685
686- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
687
688## 2.0.60
689
690- Added background agent support. Agents run in the background while you work
691- Added --disable-slash-commands CLI flag to disable all slash commands
692- Added model name to "Co-Authored-By" commit messages
693- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
694- Updated Fetch to skip summarization for pre-approved websites
695- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
696
697## 2.0.59
698
699- Added --agent CLI flag to override the agent setting for the current session
700- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
701- VS Code: Fixed .claude.json config file being read from incorrect location
702
703## 2.0.58
704
705- Pro users now have access to Opus 4.5 as part of their subscription!
706- Fixed timer duration showing "11m 60s" instead of "12m 0s"
707- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
708
709## 2.0.57
710
711- Added feedback input when rejecting plans, allowing users to tell Claude what to change
712- VSCode: Added streaming message support for real-time response display
713
714## 2.0.56
715
716- Added setting to enable/disable terminal progress bar (OSC 9;4)
717- VSCode Extension: Added support for VS Code's secondary sidebar (VS Code 1.97+), allowing Claude Code to be displayed in the right sidebar while keeping the file explorer on the left. Requires setting sidebar as Preferred Location in the config.
718
719## 2.0.55
720
721- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
722- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
723- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
724- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
725
726## 2.0.54
727
728- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
729- Fix issue with excessive iTerm notifications
730
731## 2.0.52
732
733- Fixed duplicate message display when starting Claude with a command line argument
734- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
735- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
736- Permit some uses of `$!` in bash commands
737
738## 2.0.51
739
740- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
741- Introducing Claude Code for Desktop: https://claude.com/download
742- To give you room to try out our new model, we've updated usage limits for Claude Code users. See the Claude Opus 4.5 blog for full details
743- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
744- Plan Mode now builds more precise plans and executes more thoroughly
745- Usage limit notifications now easier to understand
746- Switched `/usage` back to "% used"
747- Fixed handling of thinking errors
748- Fixed performance regression
749
750## 2.0.50
751
752- Fixed bug preventing calling MCP tools that have nested references in their input schemas
753- Silenced a noisy but harmless error during upgrades
754- Improved ultrathink text display
755- Improved clarity of 5-hour session limit warning message
756
757## 2.0.49
758
759- Added readline-style ctrl-y for pasting deleted text
760- Improved clarity of usage limit warning message
761- Fixed handling of subagent permissions
762
763## 2.0.47
764
765- Improved error messages and validation for `claude --teleport`
766- Improved error handling in `/usage`
767- Fixed race condition with history entry not getting logged at exit
768- Fixed Vertex AI configuration not being applied from `settings.json`
769
770## 2.0.46
771
772- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
773
774## 2.0.45
775
776- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
777- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
778- Send background tasks to Claude Code on the web by starting a message with `&`
779
780## 2.0.43
781
782- Added `permissionMode` field for custom agents
783- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
784- Added skills frontmatter field to declare skills to auto-load for subagents
785- Added the `SubagentStart` hook event
786- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
787- Fixed duplicate rendering of some messages in the UI
788- Fixed some visual flickers
789- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
790
791## 2.0.42
792
793- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
794
795## 2.0.41
796
797- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
798- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
799- Fixed incorrect labeling of user settings vs project settings in command descriptions
800- Fixed crash when plugin command hooks timeout during execution
801- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
802- Fixed broken security documentation links in trust dialogs and onboarding
803- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
804- ctrl-r history search landing on a slash command no longer cancels the search
805- SDK: Support custom timeouts for hooks
806- Allow more safe git commands to run without approval
807- Plugins: Added support for sharing and installing output styles
808- Teleporting a session from web will automatically set the upstream branch
809
810## 2.0.37
811
812- Fixed how idleness is computed for notifications
813- Hooks: Added matcher values for Notification hook events
814- Output Styles: Added `keep-coding-instructions` option to frontmatter
815
816## 2.0.36
817
818- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
819- Fixed queued messages being incorrectly executed as bash commands
820- Fixed input being lost when typing while a queued message is processed
821
822## 2.0.35
823
824- Improve fuzzy search results when searching commands
825- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
826- Added `CLAUDE_CODE_EXIT_AFTER_STOP_DELAY` environment variable to automatically exit SDK mode after a specified idle duration, useful for automated workflows and scripts
827- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
828- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
829
830## 2.0.34
831
832- VSCode Extension: Added setting to configure the initial permission mode for new conversations
833- Improved file path suggestion performance with native Rust-based fuzzy finder
834- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
835- Fixed memory crash when reading or writing large files (especially base64-encoded images)
836
837## 2.0.33
838
839- Native binary installs now launch quicker.
840- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
841- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
842
843## 2.0.32
844
845- Un-deprecate output styles based on community feedback
846- Added `companyAnnouncements` setting for displaying announcements on startup
847- Fixed hook progress messages not updating correctly during PostToolUse hook execution
848
849## 2.0.31
850
851- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
852- Vertex: add support for Web Search on supported models
853- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
854- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
855- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
856- Fixed plugin uninstall not removing plugins
857
858## 2.0.30
859
860- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
861- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
862- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
863- Added prompt-based stop hooks
864- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
865- Enabled SSE MCP servers on native build
866- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
867- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
868- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
869- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
870- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
871- Fixed bug that saved session permissions to local settings
872- Fixed MCP tools not being available to sub-agents
873- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
874- Fixed delay when navigating through typeahead suggestions with arrow keys
875- VSCode: Restored selection indicator in input footer showing current file or code selection status
876
877## 2.0.28
878
879- Plan mode: introduced new Plan subagent
880- Subagents: claude can now choose to resume subagents
881- Subagents: claude can dynamically choose the model used by its subagents
882- SDK: added --max-budget-usd flag
883- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
884- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
885- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
886- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
887- Various other bug fixes
888
889## 2.0.27
890
891- New UI for permission prompts
892- Added current branch filtering and search to session resume screen for easier navigation
893- Fixed directory @-mention causing "No assistant message found" error
894- VSCode Extension: Add config setting to include .gitignored files in file searches
895- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
896
897## 2.0.25
898
899- Removed legacy SDK entrypoint. Please migrate to @anthropic-ai/claude-agent-sdk for future SDK updates: https://platform.claude.com/docs/en/agent-sdk/migration-guide
900
901## 2.0.24
902
903- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
904- Claude Code Web: Support for Web -> CLI teleport
905- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
906- Bedrock: Display awsAuthRefresh output when auth is required
907
908## 2.0.22
909
910- Fixed content layout shift when scrolling through slash commands
911- IDE: Add toggle to enable/disable thinking.
912- Fix bug causing duplicate permission prompts with parallel tool calls
913- Add support for enterprise managed MCP allowlist and denylist
914
915## 2.0.21
916
917- Support MCP `structuredContent` field in tool responses
918- Added an interactive question tool
919- Claude will now ask you questions more often in plan mode
920- Added Haiku 4.5 as a model option for Pro users
921- Fixed an issue where queued commands don't have access to previous messages' output
922
923## 2.0.20
924
925- Added support for Claude Skills
926
927## 2.0.19
928
929- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
930- Fixed a bug where Haiku was unnecessarily called in print mode
931
932## 2.0.17
933
934- Added Haiku 4.5 to model selector!
935- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
936- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
937- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
938- OTEL: support HTTP_PROXY and HTTPS_PROXY
939- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
940
941## 2.0.15
942
943- Fixed bug with resuming where previously created files needed to be read again before writing
944- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
945
946## 2.0.14
947
948- Fix @-mentioning MCP servers to toggle them on/off
949- Improve permission checks for bash with inline env vars
950- Fix ultrathink + thinking toggle
951- Reduce unnecessary logins
952- Document --system-prompt
953- Several improvements to rendering
954- Plugins UI polish
955
956## 2.0.13
957
958- Fixed `/plugin` not working on native build
959
960## 2.0.12
961
962- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
963- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
964- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
965- `/plugin validate` command for validating plugin structure and configuration
966- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
967- Plugin documentation available at https://code.claude.com/docs/en/plugins
968- Comprehensive error messages and diagnostics via `/doctor` command
969- Avoid flickering in `/model` selector
970- Improvements to `/help`
971- Avoid mentioning hooks in `/resume` summaries
972- Changes to the "verbose" setting in `/config` now persist across sessions
973
974## 2.0.11
975
976- Reduced system prompt size by 1.4k tokens
977- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
978- Fixed Opus fallback rate limit errors appearing incorrectly
979- Fixed /add-dir command selecting wrong default tab
980
981## 2.0.10
982
983- Rewrote terminal renderer for buttery smooth UI
984- Enable/disable MCP servers by @mentioning, or in /mcp
985- Added tab completion for shell commands in bash mode
986- PreToolUse hooks can now modify tool inputs
987- Press Ctrl-G to edit your prompt in your system's configured text editor
988- Fixes for bash permission checks with environment variables in the command
989
990## 2.0.9
991
992- Fix regression where bash backgrounding stopped working
993
994## 2.0.8
995
996- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
997- IDE: Add drag-and-drop support for files and folders in chat
998- /context: Fix counting for thinking blocks
999- Improve message rendering for users with light themes on dark terminals
1000- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
1001
1002## 2.0.5
1003
1004- IDE: Fix IME unintended message submission with Enter and Tab
1005- IDE: Add "Open in Terminal" link in login screen
1006- Fix unhandled OAuth expiration 401 API errors
1007- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
1008
1009## 2.0.1
1010
1011- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
1012- Various bug fixes and presentation improvements
1013
1014## 2.0.0
1015
1016- New native VS Code extension
1017- Fresh coat of paint throughout the whole app
1018- /rewind a conversation to undo code changes
1019- /usage command to see plan limits
1020- Tab to toggle thinking (sticky across sessions)
1021- Ctrl-R to search history
1022- Unshipped claude config command
1023- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1024- SDK: The Claude Code SDK is now the Claude Agent SDK
1025- Add subagents dynamically with `--agents` flag
1026
1027## 1.0.126
1028
1029- Enable /context command for Bedrock and Vertex
1030- Add mTLS support for HTTP-based OpenTelemetry exporters
1031
1032## 1.0.124
1033
1034- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1035- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1036- No longer inform Claude of the list of allowed tools when permission is denied
1037- Fixed security vulnerability in Bash tool permission checks
1038- Improved VSCode extension performance for large files
1039
1040## 1.0.123
1041
1042- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1043- Fixed thinking mode triggering on negation phrases like "don't think"
1044- Fixed rendering performance degradation during token streaming
1045- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1046- Enhanced BashTool environment snapshot logging
1047- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1048- Migrated --debug logging to a file, to enable easy tailing & filtering
1049
1050## 1.0.120
1051
1052- Fix input lag during typing, especially noticeable with large prompts
1053- Improved VSCode extension command registry and sessions dialog user experience
1054- Enhanced sessions dialog responsiveness and visual feedback
1055- Fixed IDE compatibility issue by removing worktree support check
1056- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1057
1058## 1.0.119
1059
1060- Fix Windows issue where process visually freezes on entering interactive mode
1061- Support dynamic headers for MCP servers via headersHelper configuration
1062- Fix thinking mode not working in headless sessions
1063- Fix slash commands now properly update allowed tools instead of replacing them
1064
1065## 1.0.117
1066
1067- Add Ctrl-R history search to recall previous commands like bash/zsh
1068- Fix input lag while typing, especially on Windows
1069- Add sed command to auto-allowed commands in acceptEdits mode
1070- Fix Windows PATH comparison to be case-insensitive for drive letters
1071- Add permissions management hint to /add-dir output
1072
1073## 1.0.115
1074
1075- Improve thinking mode display with enhanced visual effects
1076- Type /t to temporarily disable thinking mode in your prompt
1077- Improve path validation for glob and grep tools
1078- Show condensed output for post-tool hooks to reduce visual clutter
1079- Fix visual feedback when loading state completes
1080- Improve UI consistency for permission request dialogs
1081
1082## 1.0.113
1083
1084- Deprecated piped input in interactive mode
1085- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1086
1087## 1.0.112
1088
1089- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1090- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1091- Hooks: Added systemMessage support for SessionEnd hooks
1092- Added `spinnerTipsEnabled` setting to disable spinner tips
1093- IDE: Various improvements and bug fixes
1094
1095## 1.0.111
1096
1097- /model now validates provided model names
1098- Fixed Bash tool crashes caused by malformed shell syntax parsing
1099
1100## 1.0.110
1101
1102- /terminal-setup command now supports WezTerm
1103- MCP: OAuth tokens now proactively refresh before expiration
1104- Fixed reliability issues with background Bash processes
1105
1106## 1.0.109
1107
1108- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1109
1110## 1.0.106
1111
1112- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1113
1114## 1.0.97
1115
1116- Settings: /doctor now validates permission rule syntax and suggests corrections
1117
1118## 1.0.94
1119
1120- Vertex: add support for global endpoints for supported models
1121- /memory command now allows direct editing of all imported memory files
1122- SDK: Add custom tools as callbacks
1123- Added /todos command to list current todo items
1124
1125## 1.0.93
1126
1127- Windows: Add alt + v shortcut for pasting images from clipboard
1128- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1129
1130## 1.0.90
1131
1132- Settings file changes take effect immediately - no restart required
1133
1134## 1.0.88
1135
1136- Fixed issue causing "OAuth authentication is currently not supported"
1137- Status line input now includes `exceeds_200k_tokens`
1138- Fixed incorrect usage tracking in /cost.
1139- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1140- Bedrock: Updated default Sonnet model to Sonnet 4
1141
1142## 1.0.86
1143
1144- Added /context to help users self-serve debug context issues
1145- SDK: Added UUID support for all SDK messages
1146- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1147
1148## 1.0.85
1149
1150- Status line input now includes session cost info
1151- Hooks: Introduced SessionEnd hook
1152
1153## 1.0.84
1154
1155- Fix tool_use/tool_result id mismatch error when network is unstable
1156- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1157- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1158- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1159
1160## 1.0.83
1161
1162- @-mention: Support files with spaces in path
1163- New shimmering spinner
1164
1165## 1.0.82
1166
1167- SDK: Add request cancellation support
1168- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1169- Settings: Validation prevents invalid fields in .claude/settings.json files
1170- MCP: Improve tool name consistency
1171- Bash: Fix crash when Claude tries to automatically read large files
1172
1173## 1.0.81
1174
1175- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1176- Agents: Fix custom agent loading when agent files are unparsable
1177
1178## 1.0.80
1179
1180- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1181
1182## 1.0.77
1183
1184- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1185- SDK: Add session support and permission denial tracking
1186- Fix token limit errors in conversation summarization
1187- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1188
1189## 1.0.73
1190
1191- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1192- MCP: Press Esc to cancel OAuth authentication flows
1193- Bash: Improved command validation and reduced false security warnings
1194- UI: Enhanced spinner animations and status line visual hierarchy
1195- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1196
1197## 1.0.72
1198
1199- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1200
1201## 1.0.71
1202
1203- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1204- Customizable status line: add your terminal prompt to Claude Code with /statusline
1205
1206## 1.0.70
1207
1208- Performance: Optimized message rendering for better performance with large contexts
1209- Windows: Fixed native file search, ripgrep, and subagent functionality
1210- Added support for @-mentions in slash command arguments
1211
1212## 1.0.69
1213
1214- Upgraded Opus to version 4.1
1215
1216## 1.0.68
1217
1218- Fix incorrect model names being used for certain commands like `/pr-comments`
1219- Windows: improve permissions checks for allow / deny tools and project trust. This may create a new project entry in `.claude.json` - manually merge the history field if desired.
1220- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1221- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1222- SDK: Added canUseTool callback support for tool confirmation
1223- Added `disableAllHooks` setting
1224- Improved file suggestions performance in large repos
1225
1226## 1.0.65
1227
1228- IDE: Fixed connection stability issues and error handling for diagnostics
1229- Windows: Fixed shell environment setup for users without .bashrc files
1230
1231## 1.0.64
1232
1233- Agents: Added model customization support - you can now specify which model an agent should use
1234- Agents: Fixed unintended access to the recursive agent tool
1235- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1236- SDK: Fixed user input tracking across multi-turn conversations
1237- Added hidden files to file search and @-mention suggestions
1238
1239## 1.0.63
1240
1241- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1242
1243## 1.0.62
1244
1245- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1246- Hooks: Added SessionStart hook for new session initialization
1247- /add-dir command now supports typeahead for directory paths
1248- Improved network connectivity check reliability
1249
1250## 1.0.61
1251
1252- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1253- Settings: Added `--settings` flag to load settings from a JSON file
1254- Settings: Fixed resolution of settings files paths that are symlinks
1255- OTEL: Fixed reporting of wrong organization after authentication changes
1256- Slash commands: Fixed permissions checking for allowed-tools with Bash
1257- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1258- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1259- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1260
1261## 1.0.60
1262
1263- You can now create custom subagents for specialized tasks! Run /agents to get started
1264
1265## 1.0.59
1266
1267- SDK: Added tool confirmation support with canUseTool callback
1268- SDK: Allow specifying env for spawned process
1269- Hooks: Exposed PermissionDecision to hooks (including "ask")
1270- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1271- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1272
1273## 1.0.58
1274
1275- Added support for reading PDFs
1276- MCP: Improved server health status display in 'claude mcp list'
1277- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1278
1279## 1.0.57
1280
1281- Added support for specifying a model in slash commands
1282- Improved permission messages to help Claude understand allowed tools
1283- Fix: Remove trailing newlines from bash output in terminal wrapping
1284
1285## 1.0.56
1286
1287- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1288- Fixes for WSL IDE detection
1289- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1290
1291## 1.0.55
1292
1293- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1294- Windows: fixed Ctrl+Z crash
1295- SDK: Added ability to capture error logging
1296- Add --system-prompt-file option to override system prompt in print mode
1297
1298## 1.0.54
1299
1300- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1301- Custom slash commands: Added argument-hint to frontmatter
1302- Windows: OAuth uses port 45454 and properly constructs browser URL
1303- Windows: mode switching now uses alt + m, and plan mode renders properly
1304- Shell: Switch to in-memory shell snapshot to fix file-related errors
1305
1306## 1.0.53
1307
1308- Updated @-mention file truncation from 100 lines to 2000 lines
1309- Add helper script settings for AWS token refresh: awsAuthRefresh (for foreground operations like aws sso login) and awsCredentialExport (for background operation with STS-like response).
1310
1311## 1.0.52
1312
1313- Added support for MCP server instructions
1314
1315## 1.0.51
1316
1317- Added support for native Windows (requires Git for Windows)
1318- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1319- Settings: /doctor can now help you identify and fix invalid setting files
1320- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1321- Increased auto-compact warning threshold from 60% to 80%
1322- Fixed an issue with handling user directories with spaces for shell snapshots
1323- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1324- Custom slash commands: Fixed user-level commands in subdirectories
1325- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1326
1327## 1.0.48
1328
1329- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1330- Added progress messages to Bash tool based on the last 5 lines of command output
1331- Added expanding variables support for MCP server configuration
1332- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1333- Improved IDE extension path handling when Claude Code runs in WSL
1334- Hooks: Added a PreCompact hook
1335- Vim mode: Added c, f/F, t/T
1336
1337## 1.0.45
1338
1339- Redesigned Search (Grep) tool with new tool input parameters and features
1340- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1341- Fixed config file corruption issue by enforcing atomic writes
1342- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1343- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1344- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1345
1346## 1.0.44
1347
1348- New /export command lets you quickly export a conversation for sharing
1349- MCP: resource_link tool results are now supported
1350- MCP: tool annotations and tool titles now display in /mcp view
1351- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1352
1353## 1.0.43
1354
1355- Fixed a bug where the theme selector was saving excessively
1356- Hooks: Added EPIPE system error handling
1357
1358## 1.0.42
1359
1360- Added tilde (`~`) expansion support to `/add-dir` command
1361
1362## 1.0.41
1363
1364- Hooks: Split Stop hook triggering into Stop and SubagentStop
1365- Hooks: Enabled optional timeout configuration for each command
1366- Hooks: Added "hook_event_name" to hook input
1367- Fixed a bug where MCP tools would display twice in tool list
1368- New tool parameters JSON for Bash tool in `tool_decision` event
1369
1370## 1.0.40
1371
1372- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1373
1374## 1.0.39
1375
1376- New Active Time metric in OpenTelemetry logging
1377
1378## 1.0.38
1379
1380- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1381
1382## 1.0.37
1383
1384- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1385
1386## 1.0.36
1387
1388- Web search now takes today's date into context
1389- Fixed a bug where stdio MCP servers were not terminating properly on exit
1390
1391## 1.0.35
1392
1393- Added support for MCP OAuth Authorization Server discovery
1394
1395## 1.0.34
1396
1397- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1398
1399## 1.0.33
1400
1401- Improved logging functionality with session ID support
1402- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1403- Improvements to plan mode
1404
1405## 1.0.32
1406
1407- Updated loopback config for litellm
1408- Added forceLoginMethod setting to bypass login selection screen
1409
1410## 1.0.31
1411
1412- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1413
1414## 1.0.30
1415
1416- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1417- Improved file path autocomplete with filename matching
1418- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1419- Enhanced jq regex support for complex filters with pipes and select
1420
1421## 1.0.29
1422
1423- Improved CJK character support in cursor navigation and rendering
1424
1425## 1.0.28
1426
1427- Slash commands: Fix selector display during history navigation
1428- Resizes images before upload to prevent API size limit errors
1429- Added XDG_CONFIG_HOME support to configuration directory
1430- Performance optimizations for memory usage
1431- New attributes (terminal.type, language) in OpenTelemetry logging
1432
1433## 1.0.27
1434
1435- Streamable HTTP MCP servers are now supported
1436- Remote MCP servers (SSE and HTTP) now support OAuth
1437- MCP resources can now be @-mentioned
1438- /resume slash command to switch conversations within Claude Code
1439
1440## 1.0.25
1441
1442- Slash commands: moved "project" and "user" prefixes to descriptions
1443- Slash commands: improved reliability for command discovery
1444- Improved support for Ghostty
1445- Improved web search reliability
1446
1447## 1.0.24
1448
1449- Improved /mcp output
1450- Fixed a bug where settings arrays got overwritten instead of merged
1451
1452## 1.0.23
1453
1454- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1455- Released Python SDK: pip install claude-code-sdk to get started
1456
1457## 1.0.22
1458
1459- SDK: Renamed `total_cost` to `total_cost_usd`
1460
1461## 1.0.21
1462
1463- Improved editing of files with tab-based indentation
1464- Fix for tool_use without matching tool_result errors
1465- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1466
1467## 1.0.18
1468
1469- Added --add-dir CLI argument for specifying additional working directories
1470- Added streaming input support without require -p flag
1471- Improved startup performance and session storage performance
1472- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1473- Added detailed MCP server tools display (/mcp)
1474- MCP authentication and permission improvements
1475- Added auto-reconnection for MCP SSE connections on disconnect
1476- Fixed issue where pasted content was lost when dialogs appeared
1477
1478## 1.0.17
1479
1480- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1481- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1482- MCP server list UI improvements
1483- Update Claude Code process title to display "claude" instead of "node"
1484
1485## 1.0.11
1486
1487- Claude Code can now also be used with a Claude Pro subscription
1488- Added /upgrade for smoother switching to Claude Max plans
1489- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1490- Improved shell configuration error handling
1491- Improved todo list handling during compaction
1492
1493## 1.0.10
1494
1495- Added markdown table support
1496- Improved streaming performance
1497
1498## 1.0.8
1499
1500- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1501- Increased default otel interval from 1s -> 5s
1502- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1503- Fixed a regression where search tools unnecessarily asked for permissions
1504- Added support for triggering thinking non-English languages
1505- Improved compacting UI
1506
1507## 1.0.7
1508
1509- Renamed /allowed-tools -> /permissions
1510- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1511- Deprecated claude config commands in favor of editing settings.json
1512- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1513- Improved error handling for /install-github-app
1514- Bugfixes, UI polish, and tool reliability improvements
1515
1516## 1.0.6
1517
1518- Improved edit reliability for tab-indented files
1519- Respect CLAUDE_CONFIG_DIR everywhere
1520- Reduced unnecessary tool permission prompts
1521- Added support for symlinks in @file typeahead
1522- Bugfixes, UI polish, and tool reliability improvements
1523
1524## 1.0.4
1525
1526- Fixed a bug where MCP tool errors weren't being parsed correctly
1527
1528## 1.0.1
1529
1530- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1531- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1532- Updated documentation links and OAuth process descriptions
1533
1534## 1.0.0
1535
1536- Claude Code is now generally available
1537- Introducing Sonnet 4 and Opus 4 models
1538
1539## 0.2.125
1540
1541- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1542- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1543
1544## 0.2.117
1545
1546- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1547- Introduced settings.cleanupPeriodDays
1548- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1549- Introduced --debug mode
1550
1551## 0.2.108
1552
1553- You can now send messages to Claude while it works to steer Claude in real-time
1554- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1555- Fixed a bug where thinking was not working in -p mode
1556- Fixed a regression in /cost reporting
1557- Deprecated MCP wizard interface in favor of other MCP commands
1558- Lots of other bugfixes and improvements
1559
1560## 0.2.107
1561
1562- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1563
1564## 0.2.106
1565
1566- MCP SSE server configs can now specify custom headers
1567- Fixed a bug where MCP permission prompt didn't always show correctly
1568
1569## 0.2.105
1570
1571- Claude can now search the web
1572- Moved system & account status to /status
1573- Added word movement keybindings for Vim
1574- Improved latency for startup, todo tool, and file edits
1575
1576## 0.2.102
1577
1578- Improved thinking triggering reliability
1579- Improved @mention reliability for images and folders
1580- You can now paste multiple large chunks into one prompt
1581
1582## 0.2.100
1583
1584- Fixed a crash caused by a stack overflow error
1585- Made db storage optional; missing db support disables --continue and --resume
1586
1587## 0.2.98
1588
1589- Fixed an issue where auto-compact was running twice
1590
1591## 0.2.96
1592
1593- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1594
1595## 0.2.93
1596
1597- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1598- Claude now has access to a Todo list that helps it stay on track and be more organized
1599
1600## 0.2.82
1601
1602- Added support for --disallowedTools
1603- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1604
1605## 0.2.75
1606
1607- Hit Enter to queue up additional messages while Claude is working
1608- Drag in or copy/paste image files directly into the prompt
1609- @-mention files to directly add them to context
1610- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1611- Improved performance for filename auto-complete
1612
1613## 0.2.74
1614
1615- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1616- Task tool can now perform writes and run bash commands
1617
1618## 0.2.72
1619
1620- Updated spinner to indicate tokens loaded and tool usage
1621
1622## 0.2.70
1623
1624- Network commands like curl are now available for Claude to use
1625- Claude can now run multiple web queries in parallel
1626- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1627
1628## 0.2.69
1629
1630- Fixed UI glitches with improved Select component behavior
1631- Enhanced terminal output display with better text truncation logic
1632
1633## 0.2.67
1634
1635- Shared project permission rules can be saved in .claude/settings.json
1636
1637## 0.2.66
1638
1639- Print mode (-p) now supports streaming output via --output-format=stream-json
1640- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1641
1642## 0.2.63
1643
1644- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1645
1646## 0.2.61
1647
1648- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1649- Enhanced image detection for more reliable clipboard paste functionality
1650- Fixed an issue where ESC key could crash the conversation history selector
1651
1652## 0.2.59
1653
1654- Copy+paste images directly into your prompt
1655- Improved progress indicators for bash and fetch tools
1656- Bugfixes for non-interactive mode (-p)
1657
1658## 0.2.54
1659
1660- Quickly add to Memory by starting your message with '#'
1661- Press ctrl+r to see full output for long tool results
1662- Added support for MCP SSE transport
1663
1664## 0.2.53
1665
1666- New web fetch tool lets Claude view URLs that you paste in
1667- Fixed a bug with JPEG detection
1668
1669## 0.2.50
1670
1671- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1672
1673## 0.2.49
1674
1675- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1676
1677## 0.2.47
1678
1679- Press Tab to auto-complete file and folder names
1680- Press Shift + Tab to toggle auto-accept for file edits
1681- Automatic conversation compaction for infinite conversation length (toggle with /config)
1682
1683## 0.2.44
1684
1685- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1686
1687## 0.2.41
1688
1689- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1690- MCP server startup no longer blocks the app from starting up
1691
1692## 0.2.37
1693
1694- New /release-notes command lets you view release notes at any time
1695- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1696
1697## 0.2.36
1698
1699- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1700- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1701
1702## 0.2.34
1703
1704- Vim bindings for text input - enable with /vim or /config
1705
1706## 0.2.32
1707
1708- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1709- Fix for some PersistentShell issues
1710
1711## 0.2.31
1712
1713- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1714- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1715
1716## 0.2.30
1717
1718- Added ANSI color theme for better terminal compatibility
1719- Fixed issue where slash command arguments weren't being sent properly
1720- (Mac-only) API keys are now stored in macOS Keychain
1721
1722## 0.2.26
1723
1724- New /approved-tools command for managing tool permissions
1725- Word-level diff display for improved code readability
1726- Fuzzy matching for slash commands
1727
1728## 0.2.21
1729
1730- Fuzzy matching for /commands