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 ef1e0ac · on Feb 6, 2026 History
1,637 lines · 86.5 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.33
4
5- Fixed agent teammate sessions in tmux to send and receive messages
6- Fixed warnings about agent teams not being available on your current plan
7- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
8- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
9- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
10- Added plugin name to skill descriptions and `/skills` menu for better discoverability
11- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
12- 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
13- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
14- 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
15- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
16- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
17- Errors from invalid managed settings are now surfaced
18- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
19- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
20- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
21
22## 2.1.32
23
24- Claude Opus 4.6 is now available!
25- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
26- Claude now automatically records and recalls memories as it works
27- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
28- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
29- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
30- Updated --resume to re-use --agent value specified in previous conversation by default.
31- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
32- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
33- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
34- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
35- VSCode: Added spinner when loading past conversations list
36
37## 2.1.31
38
39- Added session resume hint on exit, showing how to continue your conversation later
40- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
41- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
42- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
43- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
44- 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
45- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
46- 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
47- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
48- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
49- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
50
51## 2.1.30
52
53- 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.
54- 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`.
55- Added `/debug` for Claude to help troubleshoot the current session
56- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
57- Added token count, tool uses, and duration metrics to Task tool results
58- Added reduced motion mode to the config
59- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
60- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
61- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
62- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
63- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
64- Fixed permission dialogs stealing focus while actively typing
65- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
66- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
67- 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
68- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
69- Changed `/model` to execute immediately instead of being queued
70- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
71- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
72
73## 2.1.29
74
75- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
76
77## 2.1.27
78
79- Added tool call failures and denials to debug logs
80- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
81- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
82- Sessions are now automatically linked to PRs when created via `gh pr create`
83- Fixed /context command not displaying colored output
84- Fixed status bar duplicating background task indicator when PR status was shown
85- Windows: Fixed bash command execution failing for users with `.bashrc` files
86- Windows: Fixed console windows flashing when spawning child processes
87- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
88
89## 2.1.25
90
91- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
92
93## 2.1.23
94
95- Added customizable spinner verbs setting (`spinnerVerbs`)
96- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
97- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
98- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
99- Fixed pending async hooks not being cancelled when headless streaming sessions ended
100- Fixed tab completion not updating the input field when accepting a suggestion
101- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
102- Improved terminal rendering performance with optimized screen data layout
103- Changed Bash commands to show timeout duration alongside elapsed time
104- Changed merged pull requests to show a purple status indicator in the prompt footer
105- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
106
107## 2.1.22
108
109- Fixed structured outputs for non-interactive (-p) mode
110
111## 2.1.21
112
113- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
114- Fixed shell completion cache files being truncated on exit
115- Fixed API errors when resuming sessions that were interrupted during tool execution
116- Fixed auto-compact triggering too early on models with large output token limits
117- Fixed task IDs potentially being reused after deletion
118- Fixed file search not working in VS Code extension on Windows
119- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
120- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
121- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
122- [VSCode] Fixed message action buttons having incorrect background colors
123
124## 2.1.20
125
126- Added arrow key history navigation in vim normal mode when cursor cannot move further
127- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
128- 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
129- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
130- Added ability to delete tasks via the `TaskUpdate` tool
131- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
132- Fixed agents sometimes ignoring user messages sent while actively working on a task
133- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
134- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
135- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
136- Fixed draft prompt being lost when pressing UP arrow to navigate command history
137- Fixed ghost text flickering when typing slash commands mid-input
138- Fixed marketplace source removal not properly deleting settings
139- Fixed duplicate output in some commands like `/context`
140- Fixed task list sometimes showing outside the main conversation view
141- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
142- Fixed crashes when cancelling tool use
143- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
144- Improved thinking status text with a subtle shimmer animation
145- Improved task list to dynamically adjust visible items based on terminal height
146- Improved fork conversation hint to show how to resume the original session
147- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
148- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
149- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
150- Changed the `/copy` command to be available to all users
151- Changed background agents to prompt for tool permissions before launching
152- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
153- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
154
155## 2.1.19
156
157- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
158- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
159- Fixed crashes on processors without AVX instruction support
160- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
161- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
162- Fixed resuming sessions by custom title not working when run from a different directory
163- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
164- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
165- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
166- Fixed file write preview omitting empty lines
167- Changed skills without additional permissions or hooks to be allowed without requiring approval
168- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
169- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
170- [VSCode] Enabled session forking and rewind functionality for all users
171
172## 2.1.18
173
174- 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
175
176## 2.1.17
177
178- Fixed crashes on processors without AVX instruction support
179
180## 2.1.16
181
182- Added new task management system, including new capabilities like dependency tracking
183- [VSCode] Added native plugin management support
184- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
185- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
186- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
187- Fixed session titles on the resume screen not respecting the user's language setting
188- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
189
190## 2.1.15
191
192- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
193- Improved UI rendering performance with React Compiler
194- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
195- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
196
197## 2.1.14
198
199- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
200- Added search to installed plugins list - type to filter by name or description
201- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
202- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
203- Fixed memory issues that could cause crashes when running parallel subagents
204- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
205- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
206- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
207- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
208- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
209- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
210- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
211- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
212- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
213- Improved backspace to delete pasted text as a single token instead of one character at a time
214- [VSCode] Added `/usage` command to display current plan usage
215
216## 2.1.12
217
218- Fixed message rendering bug
219
220## 2.1.11
221
222- Fixed excessive MCP connection requests for HTTP/SSE transports
223
224## 2.1.10
225
226- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
227- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
228- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
229- Improved startup to capture keystrokes typed before the REPL is fully ready
230- Improved file suggestions to show as removable attachments instead of inserting text when accepted
231- [VSCode] Added install count display to plugin listings
232- [VSCode] Added trust warning when installing plugins
233
234## 2.1.9
235
236- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
237- Added `plansDirectory` setting to customize where plan files are stored
238- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
239- Added session URL attribution to commits and PRs created from web sessions
240- Added support for `PreToolUse` hooks to return `additionalContext` to the model
241- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
242- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
243- Fixed MCP server reconnection hanging when cached connection promise never resolves
244- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
245
246## 2.1.7
247
248- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
249- Added ability to provide feedback when accepting permission prompts
250- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
251- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
252- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
253- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
254- 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)
255- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
256- Fixed terminal title animation jitter by using fixed-width braille characters
257- Fixed plugins with git submodules not being fully initialized when installed
258- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
259- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
260- 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.
261- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
262- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
263
264## 2.1.6
265
266- Added search functionality to `/config` command for quickly filtering settings
267- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
268- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
269- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
270- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
271- Added an error display when the editor fails during Ctrl+G
272- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
273- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
274- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
275- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
276- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
277- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
278- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
279- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
280- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
281- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
282- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
283- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
284- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
285- Improved @ autocomplete with icons for different suggestion types and single-line formatting
286- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
287- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
288- Changed terminal title to "Claude Code" on startup for better window identification
289- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
290- [VSCode] Fixed usage indicator not updating after manual compact
291
292## 2.1.5
293
294- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
295
296## 2.1.4
297
298- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
299- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
300
301## 2.1.3
302
303- Merged slash commands and skills, simplifying the mental model with no change in behavior
304- Added release channel (`stable` or `latest`) toggle to `/config`
305- 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
306- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
307- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
308- Fixed mismatch between background task count in status bar and items shown in tasks dialog
309- Fixed sub-agents using the wrong model during conversation compaction
310- Fixed web search in sub-agents using incorrect model
311- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
312- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
313- Improved slash command suggestion readability by truncating long descriptions to 2 lines
314- Changed tool hook execution timeout from 60 seconds to 10 minutes
315- [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)
316
317## 2.1.2
318
319- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
320- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
321- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
322- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
323- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
324- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
325- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
326- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
327- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
328- Fixed updates incorrectly claiming another installation is in progress
329- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
330- Fixed remote session URL and teleport being broken when using `/tasks` command
331- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
332- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
333- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
334- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
335- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
336- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
337- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
338- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
339- [SDK] Changed minimum zod peer dependency to ^4.0.0
340- [VSCode] Fixed usage display not updating after manual compact
341
342## 2.1.0
343
344- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
345- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
346- Added support for `agent` field in skills to specify agent type for execution
347- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
348- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
349- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
350- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
351- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
352- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
353- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
354- Fixed Esc key with queued prompts to only move them to input without canceling the running task
355- Reduced permission prompts for complex bash commands
356- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
357- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
358- Fixed plugin path resolution for file-based marketplace sources
359- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
360- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
361- Fixed Claude in Chrome support for WSL environments
362- Fixed Windows native installer silently failing when executable creation fails
363- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
364- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
365- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
366- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
367- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
368- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
369- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
370- Added hooks support for skill and slash command frontmatter
371- 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
372- Added `/plan` command shortcut to enable plan mode directly from the prompt
373- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
374- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
375- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
376- Added support for `once: true` config for hooks
377- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
378- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
379- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
380- Added left/right arrow key navigation for cycling through tabs in dialogs
381- Added real-time thinking block display in Ctrl+O transcript mode
382- Added filepath to full output in background bash task details dialog
383- Added Skills as a separate category in the context visualization
384- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
385- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
386- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
387- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
388- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
389- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
390- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
391- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
392- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
393- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
394- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
395- Fixed commands with `$()` command substitution failing with parse errors
396- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
397- 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)
398- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
399- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
400- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
401- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
402- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
403- Fixed subagents sometimes not inheriting the parent's model by default
404- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
405- Fixed duplicate Bash commands appearing in permission request option labels
406- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
407- Fixed background task completion notifications to appear proactively with bullet point
408- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
409- Fixed cursor disappearing after dismissing permission dialogs
410- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
411- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
412- Fixed images being silently dropped when queueing messages while backgrounding a task
413- Fixed large pasted images failing with "Image was too large" error
414- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
415- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
416- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
417- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
418- Fixed spinner token counter to properly accumulate tokens from subagents during execution
419- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
420- Fixed race condition where LSP tool could return "no server available" during startup
421- Fixed feedback submission hanging indefinitely when network requests timeout
422- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
423- Fixed hook success message showing trailing colon when hook has no output
424- Multiple optimizations to improve startup performance
425- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
426- Improved performance when reading Jupyter notebooks with many cells
427- Improved reliability for piped input like `cat refactor.md | claude`
428- Improved reliability for AskQuestion tool
429- Improved sed in-place edit commands to render as file edits with diff preview
430- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
431- Improved compaction reliability
432- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
433- Improved skills to show progress while executing, displaying tool uses as they happen
434- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
435- Improved skill suggestions to prioritize recently and frequently used skills
436- Improved spinner feedback when waiting for the first response token
437- Improved token count display in spinner to include tokens from background agents
438- Improved incremental output for async agents to give the main thread more control and visibility
439- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
440- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
441- Improved macOS screenshot paste reliability with TIFF format support
442- Improved `/stats` output
443- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
444- Changed "Interrupted" message color from red to grey for a less alarming appearance
445- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
446- Removed underline styling from image reference links
447- [SDK] Changed minimum zod peer dependency to ^4.0.0
448- [VSCode] Added currently selected model name to the context menu
449- [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")
450- [VSCode] Fixed paragraph breaks not rendering in markdown content
451- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
452- [Windows] Fixed issue with improper rendering
453
454## 2.0.76
455
456- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
457
458## 2.0.75
459
460- Minor bugfixes
461
462## 2.0.74
463
464- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
465- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
466- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
467- Added syntax highlighting info to theme picker
468- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
469- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
470- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
471- Fixed a potential crash when syntax highlighting isn't initialized correctly
472- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
473- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
474- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
475- [Windows] Fixed issue with improper rendering
476- [VSCode] Added gift tag pictogram for year-end promotion message
477
478## 2.0.73
479
480- Added clickable `[Image #N]` links that open attached images in the default viewer
481- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
482- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
483- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
484- Fixed slow input history cycling and race condition that could overwrite text after message submission
485- Improved `/theme` command to open theme picker directly
486- Improved theme picker UI
487- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
488- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
489
490## 2.0.72
491
492- 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
493- Reduced terminal flickering
494- Added scannable QR code to mobile app tip for quick app downloads
495- Added loading indicator when resuming conversations for better feedback
496- Fixed `/context` command not respecting custom system prompts in non-interactive mode
497- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
498- Improved @ mention file suggestion speed (~3× faster in git repositories)
499- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
500- Improved settings validation errors to be more prominent
501- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
502
503## 2.0.71
504
505- Added /config toggle to enable/disable prompt suggestions
506- Added `/settings` as an alias for the `/config` command
507- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
508- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
509- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
510- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
511- New syntax highlighting engine for native build
512
513## 2.0.70
514
515- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
516- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
517- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
518- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
519- Fixed input being cleared when processing queued commands while the user was typing
520- Fixed prompt suggestions replacing typed input when pressing Tab
521- Fixed diff view not updating when terminal is resized
522- Improved memory usage by 3x for large conversations
523- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
524- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
525- Fix thinking mode toggle in /config not persisting correctly
526- Improve UI for file creation permission dialog
527
528## 2.0.69
529
530- Minor bugfixes
531
532## 2.0.68
533
534- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
535- Fixed a bug where disallowed MCP tools were visible to the model
536- Fixed an issue where steering messages could be lost while a subagent is working
537- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
538- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
539- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
540
541## 2.0.67
542
543- Thinking mode is now enabled by default for Opus 4.5
544- Thinking mode configuration has moved to /config
545- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
546- Show reason why autoupdater is disabled in `/doctor`
547- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
548- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
549- Fixed scroll position resetting after deleting a permission rule in `/permissions`
550- 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
551- Fixed `claude install --force` not bypassing stale lock files
552- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
553- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
554
555## 2.0.65
556
557- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
558- Added context window information to status line input
559- Added `fileSuggestion` setting for custom `@` file search commands
560- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
561- Fixed prompt not being saved to history when aborting a query with Escape
562- Fixed Read tool image handling to identify format from bytes instead of file extension
563
564## 2.0.64
565
566- Made auto-compacting instant
567- Agents and bash commands can run asynchronously and send messages to wake up the main agent
568- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
569- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
570- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
571- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
572- Fixed auto-loading .env when using native installer
573- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
574- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
575- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
576- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
577- Bedrock: Improve efficiency of token counting
578- Bedrock: Add support for `aws login` AWS Management Console credentials
579- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
580
581## 2.0.62
582
583- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
584- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
585- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
586- Fixed slash command selection not working when multiple commands share the same name
587- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
588- Fixed running versions getting removed because lock file incorrectly going stale
589- Fixed IDE diff tab not closing when rejecting file changes
590
591## 2.0.61
592
593- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
594
595## 2.0.60
596
597- Added background agent support. Agents run in the background while you work
598- Added --disable-slash-commands CLI flag to disable all slash commands
599- Added model name to "Co-Authored-By" commit messages
600- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
601- Updated Fetch to skip summarization for pre-approved websites
602- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
603
604## 2.0.59
605
606- Added --agent CLI flag to override the agent setting for the current session
607- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
608- VS Code: Fixed .claude.json config file being read from incorrect location
609
610## 2.0.58
611
612- Pro users now have access to Opus 4.5 as part of their subscription!
613- Fixed timer duration showing "11m 60s" instead of "12m 0s"
614- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
615
616## 2.0.57
617
618- Added feedback input when rejecting plans, allowing users to tell Claude what to change
619- VSCode: Added streaming message support for real-time response display
620
621## 2.0.56
622
623- Added setting to enable/disable terminal progress bar (OSC 9;4)
624- 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.
625
626## 2.0.55
627
628- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
629- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
630- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
631- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
632
633## 2.0.54
634
635- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
636- Fix issue with excessive iTerm notifications
637
638## 2.0.52
639
640- Fixed duplicate message display when starting Claude with a command line argument
641- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
642- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
643- Permit some uses of `$!` in bash commands
644
645## 2.0.51
646
647- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
648- Introducing Claude Code for Desktop: https://claude.com/download
649- 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
650- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
651- Plan Mode now builds more precise plans and executes more thoroughly
652- Usage limit notifications now easier to understand
653- Switched `/usage` back to "% used"
654- Fixed handling of thinking errors
655- Fixed performance regression
656
657## 2.0.50
658
659- Fixed bug preventing calling MCP tools that have nested references in their input schemas
660- Silenced a noisy but harmless error during upgrades
661- Improved ultrathink text display
662- Improved clarity of 5-hour session limit warning message
663
664## 2.0.49
665
666- Added readline-style ctrl-y for pasting deleted text
667- Improved clarity of usage limit warning message
668- Fixed handling of subagent permissions
669
670## 2.0.47
671
672- Improved error messages and validation for `claude --teleport`
673- Improved error handling in `/usage`
674- Fixed race condition with history entry not getting logged at exit
675- Fixed Vertex AI configuration not being applied from `settings.json`
676
677## 2.0.46
678
679- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
680
681## 2.0.45
682
683- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
684- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
685- Send background tasks to Claude Code on the web by starting a message with `&`
686
687## 2.0.43
688
689- Added `permissionMode` field for custom agents
690- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
691- Added skills frontmatter field to declare skills to auto-load for subagents
692- Added the `SubagentStart` hook event
693- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
694- Fixed duplicate rendering of some messages in the UI
695- Fixed some visual flickers
696- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
697
698## 2.0.42
699
700- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
701
702## 2.0.41
703
704- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
705- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
706- Fixed incorrect labeling of user settings vs project settings in command descriptions
707- Fixed crash when plugin command hooks timeout during execution
708- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
709- Fixed broken security documentation links in trust dialogs and onboarding
710- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
711- ctrl-r history search landing on a slash command no longer cancels the search
712- SDK: Support custom timeouts for hooks
713- Allow more safe git commands to run without approval
714- Plugins: Added support for sharing and installing output styles
715- Teleporting a session from web will automatically set the upstream branch
716
717## 2.0.37
718
719- Fixed how idleness is computed for notifications
720- Hooks: Added matcher values for Notification hook events
721- Output Styles: Added `keep-coding-instructions` option to frontmatter
722
723## 2.0.36
724
725- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
726- Fixed queued messages being incorrectly executed as bash commands
727- Fixed input being lost when typing while a queued message is processed
728
729## 2.0.35
730
731- Improve fuzzy search results when searching commands
732- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
733- 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
734- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
735- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
736
737## 2.0.34
738
739- VSCode Extension: Added setting to configure the initial permission mode for new conversations
740- Improved file path suggestion performance with native Rust-based fuzzy finder
741- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
742- Fixed memory crash when reading or writing large files (especially base64-encoded images)
743
744## 2.0.33
745
746- Native binary installs now launch quicker.
747- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
748- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
749
750## 2.0.32
751
752- Un-deprecate output styles based on community feedback
753- Added `companyAnnouncements` setting for displaying announcements on startup
754- Fixed hook progress messages not updating correctly during PostToolUse hook execution
755
756## 2.0.31
757
758- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
759- Vertex: add support for Web Search on supported models
760- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
761- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
762- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
763- Fixed plugin uninstall not removing plugins
764
765## 2.0.30
766
767- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
768- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
769- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
770- Added prompt-based stop hooks
771- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
772- Enabled SSE MCP servers on native build
773- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
774- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
775- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
776- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
777- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
778- Fixed bug that saved session permissions to local settings
779- Fixed MCP tools not being available to sub-agents
780- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
781- Fixed delay when navigating through typeahead suggestions with arrow keys
782- VSCode: Restored selection indicator in input footer showing current file or code selection status
783
784## 2.0.28
785
786- Plan mode: introduced new Plan subagent
787- Subagents: claude can now choose to resume subagents
788- Subagents: claude can dynamically choose the model used by its subagents
789- SDK: added --max-budget-usd flag
790- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
791- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
792- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
793- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
794- Various other bug fixes
795
796## 2.0.27
797
798- New UI for permission prompts
799- Added current branch filtering and search to session resume screen for easier navigation
800- Fixed directory @-mention causing "No assistant message found" error
801- VSCode Extension: Add config setting to include .gitignored files in file searches
802- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
803
804## 2.0.25
805
806- 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
807
808## 2.0.24
809
810- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
811- Claude Code Web: Support for Web -> CLI teleport
812- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
813- Bedrock: Display awsAuthRefresh output when auth is required
814
815## 2.0.22
816
817- Fixed content layout shift when scrolling through slash commands
818- IDE: Add toggle to enable/disable thinking.
819- Fix bug causing duplicate permission prompts with parallel tool calls
820- Add support for enterprise managed MCP allowlist and denylist
821
822## 2.0.21
823
824- Support MCP `structuredContent` field in tool responses
825- Added an interactive question tool
826- Claude will now ask you questions more often in plan mode
827- Added Haiku 4.5 as a model option for Pro users
828- Fixed an issue where queued commands don't have access to previous messages' output
829
830## 2.0.20
831
832- Added support for Claude Skills
833
834## 2.0.19
835
836- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
837- Fixed a bug where Haiku was unnecessarily called in print mode
838
839## 2.0.17
840
841- Added Haiku 4.5 to model selector!
842- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
843- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
844- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
845- OTEL: support HTTP_PROXY and HTTPS_PROXY
846- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
847
848## 2.0.15
849
850- Fixed bug with resuming where previously created files needed to be read again before writing
851- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
852
853## 2.0.14
854
855- Fix @-mentioning MCP servers to toggle them on/off
856- Improve permission checks for bash with inline env vars
857- Fix ultrathink + thinking toggle
858- Reduce unnecessary logins
859- Document --system-prompt
860- Several improvements to rendering
861- Plugins UI polish
862
863## 2.0.13
864
865- Fixed `/plugin` not working on native build
866
867## 2.0.12
868
869- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
870- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
871- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
872- `/plugin validate` command for validating plugin structure and configuration
873- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
874- Plugin documentation available at https://code.claude.com/docs/en/plugins
875- Comprehensive error messages and diagnostics via `/doctor` command
876- Avoid flickering in `/model` selector
877- Improvements to `/help`
878- Avoid mentioning hooks in `/resume` summaries
879- Changes to the "verbose" setting in `/config` now persist across sessions
880
881## 2.0.11
882
883- Reduced system prompt size by 1.4k tokens
884- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
885- Fixed Opus fallback rate limit errors appearing incorrectly
886- Fixed /add-dir command selecting wrong default tab
887
888## 2.0.10
889
890- Rewrote terminal renderer for buttery smooth UI
891- Enable/disable MCP servers by @mentioning, or in /mcp
892- Added tab completion for shell commands in bash mode
893- PreToolUse hooks can now modify tool inputs
894- Press Ctrl-G to edit your prompt in your system's configured text editor
895- Fixes for bash permission checks with environment variables in the command
896
897## 2.0.9
898
899- Fix regression where bash backgrounding stopped working
900
901## 2.0.8
902
903- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
904- IDE: Add drag-and-drop support for files and folders in chat
905- /context: Fix counting for thinking blocks
906- Improve message rendering for users with light themes on dark terminals
907- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
908
909## 2.0.5
910
911- IDE: Fix IME unintended message submission with Enter and Tab
912- IDE: Add "Open in Terminal" link in login screen
913- Fix unhandled OAuth expiration 401 API errors
914- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
915
916## 2.0.1
917
918- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
919- Various bug fixes and presentation improvements
920
921## 2.0.0
922
923- New native VS Code extension
924- Fresh coat of paint throughout the whole app
925- /rewind a conversation to undo code changes
926- /usage command to see plan limits
927- Tab to toggle thinking (sticky across sessions)
928- Ctrl-R to search history
929- Unshipped claude config command
930- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
931- SDK: The Claude Code SDK is now the Claude Agent SDK
932- Add subagents dynamically with `--agents` flag
933
934## 1.0.126
935
936- Enable /context command for Bedrock and Vertex
937- Add mTLS support for HTTP-based OpenTelemetry exporters
938
939## 1.0.124
940
941- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
942- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
943- No longer inform Claude of the list of allowed tools when permission is denied
944- Fixed security vulnerability in Bash tool permission checks
945- Improved VSCode extension performance for large files
946
947## 1.0.123
948
949- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
950- Fixed thinking mode triggering on negation phrases like "don't think"
951- Fixed rendering performance degradation during token streaming
952- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
953- Enhanced BashTool environment snapshot logging
954- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
955- Migrated --debug logging to a file, to enable easy tailing & filtering
956
957## 1.0.120
958
959- Fix input lag during typing, especially noticeable with large prompts
960- Improved VSCode extension command registry and sessions dialog user experience
961- Enhanced sessions dialog responsiveness and visual feedback
962- Fixed IDE compatibility issue by removing worktree support check
963- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
964
965## 1.0.119
966
967- Fix Windows issue where process visually freezes on entering interactive mode
968- Support dynamic headers for MCP servers via headersHelper configuration
969- Fix thinking mode not working in headless sessions
970- Fix slash commands now properly update allowed tools instead of replacing them
971
972## 1.0.117
973
974- Add Ctrl-R history search to recall previous commands like bash/zsh
975- Fix input lag while typing, especially on Windows
976- Add sed command to auto-allowed commands in acceptEdits mode
977- Fix Windows PATH comparison to be case-insensitive for drive letters
978- Add permissions management hint to /add-dir output
979
980## 1.0.115
981
982- Improve thinking mode display with enhanced visual effects
983- Type /t to temporarily disable thinking mode in your prompt
984- Improve path validation for glob and grep tools
985- Show condensed output for post-tool hooks to reduce visual clutter
986- Fix visual feedback when loading state completes
987- Improve UI consistency for permission request dialogs
988
989## 1.0.113
990
991- Deprecated piped input in interactive mode
992- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
993
994## 1.0.112
995
996- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
997- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
998- Hooks: Added systemMessage support for SessionEnd hooks
999- Added `spinnerTipsEnabled` setting to disable spinner tips
1000- IDE: Various improvements and bug fixes
1001
1002## 1.0.111
1003
1004- /model now validates provided model names
1005- Fixed Bash tool crashes caused by malformed shell syntax parsing
1006
1007## 1.0.110
1008
1009- /terminal-setup command now supports WezTerm
1010- MCP: OAuth tokens now proactively refresh before expiration
1011- Fixed reliability issues with background Bash processes
1012
1013## 1.0.109
1014
1015- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1016
1017## 1.0.106
1018
1019- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1020
1021## 1.0.97
1022
1023- Settings: /doctor now validates permission rule syntax and suggests corrections
1024
1025## 1.0.94
1026
1027- Vertex: add support for global endpoints for supported models
1028- /memory command now allows direct editing of all imported memory files
1029- SDK: Add custom tools as callbacks
1030- Added /todos command to list current todo items
1031
1032## 1.0.93
1033
1034- Windows: Add alt + v shortcut for pasting images from clipboard
1035- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1036
1037## 1.0.90
1038
1039- Settings file changes take effect immediately - no restart required
1040
1041## 1.0.88
1042
1043- Fixed issue causing "OAuth authentication is currently not supported"
1044- Status line input now includes `exceeds_200k_tokens`
1045- Fixed incorrect usage tracking in /cost.
1046- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1047- Bedrock: Updated default Sonnet model to Sonnet 4
1048
1049## 1.0.86
1050
1051- Added /context to help users self-serve debug context issues
1052- SDK: Added UUID support for all SDK messages
1053- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1054
1055## 1.0.85
1056
1057- Status line input now includes session cost info
1058- Hooks: Introduced SessionEnd hook
1059
1060## 1.0.84
1061
1062- Fix tool_use/tool_result id mismatch error when network is unstable
1063- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1064- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1065- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1066
1067## 1.0.83
1068
1069- @-mention: Support files with spaces in path
1070- New shimmering spinner
1071
1072## 1.0.82
1073
1074- SDK: Add request cancellation support
1075- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1076- Settings: Validation prevents invalid fields in .claude/settings.json files
1077- MCP: Improve tool name consistency
1078- Bash: Fix crash when Claude tries to automatically read large files
1079
1080## 1.0.81
1081
1082- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1083- Agents: Fix custom agent loading when agent files are unparsable
1084
1085## 1.0.80
1086
1087- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1088
1089## 1.0.77
1090
1091- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1092- SDK: Add session support and permission denial tracking
1093- Fix token limit errors in conversation summarization
1094- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1095
1096## 1.0.73
1097
1098- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1099- MCP: Press Esc to cancel OAuth authentication flows
1100- Bash: Improved command validation and reduced false security warnings
1101- UI: Enhanced spinner animations and status line visual hierarchy
1102- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1103
1104## 1.0.72
1105
1106- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1107
1108## 1.0.71
1109
1110- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1111- Customizable status line: add your terminal prompt to Claude Code with /statusline
1112
1113## 1.0.70
1114
1115- Performance: Optimized message rendering for better performance with large contexts
1116- Windows: Fixed native file search, ripgrep, and subagent functionality
1117- Added support for @-mentions in slash command arguments
1118
1119## 1.0.69
1120
1121- Upgraded Opus to version 4.1
1122
1123## 1.0.68
1124
1125- Fix incorrect model names being used for certain commands like `/pr-comments`
1126- 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.
1127- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1128- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1129- SDK: Added canUseTool callback support for tool confirmation
1130- Added `disableAllHooks` setting
1131- Improved file suggestions performance in large repos
1132
1133## 1.0.65
1134
1135- IDE: Fixed connection stability issues and error handling for diagnostics
1136- Windows: Fixed shell environment setup for users without .bashrc files
1137
1138## 1.0.64
1139
1140- Agents: Added model customization support - you can now specify which model an agent should use
1141- Agents: Fixed unintended access to the recursive agent tool
1142- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1143- SDK: Fixed user input tracking across multi-turn conversations
1144- Added hidden files to file search and @-mention suggestions
1145
1146## 1.0.63
1147
1148- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1149
1150## 1.0.62
1151
1152- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1153- Hooks: Added SessionStart hook for new session initialization
1154- /add-dir command now supports typeahead for directory paths
1155- Improved network connectivity check reliability
1156
1157## 1.0.61
1158
1159- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1160- Settings: Added `--settings` flag to load settings from a JSON file
1161- Settings: Fixed resolution of settings files paths that are symlinks
1162- OTEL: Fixed reporting of wrong organization after authentication changes
1163- Slash commands: Fixed permissions checking for allowed-tools with Bash
1164- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1165- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1166- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1167
1168## 1.0.60
1169
1170- You can now create custom subagents for specialized tasks! Run /agents to get started
1171
1172## 1.0.59
1173
1174- SDK: Added tool confirmation support with canUseTool callback
1175- SDK: Allow specifying env for spawned process
1176- Hooks: Exposed PermissionDecision to hooks (including "ask")
1177- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1178- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1179
1180## 1.0.58
1181
1182- Added support for reading PDFs
1183- MCP: Improved server health status display in 'claude mcp list'
1184- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1185
1186## 1.0.57
1187
1188- Added support for specifying a model in slash commands
1189- Improved permission messages to help Claude understand allowed tools
1190- Fix: Remove trailing newlines from bash output in terminal wrapping
1191
1192## 1.0.56
1193
1194- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1195- Fixes for WSL IDE detection
1196- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1197
1198## 1.0.55
1199
1200- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1201- Windows: fixed Ctrl+Z crash
1202- SDK: Added ability to capture error logging
1203- Add --system-prompt-file option to override system prompt in print mode
1204
1205## 1.0.54
1206
1207- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1208- Custom slash commands: Added argument-hint to frontmatter
1209- Windows: OAuth uses port 45454 and properly constructs browser URL
1210- Windows: mode switching now uses alt + m, and plan mode renders properly
1211- Shell: Switch to in-memory shell snapshot to fix file-related errors
1212
1213## 1.0.53
1214
1215- Updated @-mention file truncation from 100 lines to 2000 lines
1216- 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).
1217
1218## 1.0.52
1219
1220- Added support for MCP server instructions
1221
1222## 1.0.51
1223
1224- Added support for native Windows (requires Git for Windows)
1225- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1226- Settings: /doctor can now help you identify and fix invalid setting files
1227- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1228- Increased auto-compact warning threshold from 60% to 80%
1229- Fixed an issue with handling user directories with spaces for shell snapshots
1230- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1231- Custom slash commands: Fixed user-level commands in subdirectories
1232- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1233
1234## 1.0.48
1235
1236- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1237- Added progress messages to Bash tool based on the last 5 lines of command output
1238- Added expanding variables support for MCP server configuration
1239- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1240- Improved IDE extension path handling when Claude Code runs in WSL
1241- Hooks: Added a PreCompact hook
1242- Vim mode: Added c, f/F, t/T
1243
1244## 1.0.45
1245
1246- Redesigned Search (Grep) tool with new tool input parameters and features
1247- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1248- Fixed config file corruption issue by enforcing atomic writes
1249- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1250- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1251- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1252
1253## 1.0.44
1254
1255- New /export command lets you quickly export a conversation for sharing
1256- MCP: resource_link tool results are now supported
1257- MCP: tool annotations and tool titles now display in /mcp view
1258- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1259
1260## 1.0.43
1261
1262- Fixed a bug where the theme selector was saving excessively
1263- Hooks: Added EPIPE system error handling
1264
1265## 1.0.42
1266
1267- Added tilde (`~`) expansion support to `/add-dir` command
1268
1269## 1.0.41
1270
1271- Hooks: Split Stop hook triggering into Stop and SubagentStop
1272- Hooks: Enabled optional timeout configuration for each command
1273- Hooks: Added "hook_event_name" to hook input
1274- Fixed a bug where MCP tools would display twice in tool list
1275- New tool parameters JSON for Bash tool in `tool_decision` event
1276
1277## 1.0.40
1278
1279- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1280
1281## 1.0.39
1282
1283- New Active Time metric in OpenTelemetry logging
1284
1285## 1.0.38
1286
1287- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1288
1289## 1.0.37
1290
1291- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1292
1293## 1.0.36
1294
1295- Web search now takes today's date into context
1296- Fixed a bug where stdio MCP servers were not terminating properly on exit
1297
1298## 1.0.35
1299
1300- Added support for MCP OAuth Authorization Server discovery
1301
1302## 1.0.34
1303
1304- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1305
1306## 1.0.33
1307
1308- Improved logging functionality with session ID support
1309- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1310- Improvements to plan mode
1311
1312## 1.0.32
1313
1314- Updated loopback config for litellm
1315- Added forceLoginMethod setting to bypass login selection screen
1316
1317## 1.0.31
1318
1319- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1320
1321## 1.0.30
1322
1323- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1324- Improved file path autocomplete with filename matching
1325- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1326- Enhanced jq regex support for complex filters with pipes and select
1327
1328## 1.0.29
1329
1330- Improved CJK character support in cursor navigation and rendering
1331
1332## 1.0.28
1333
1334- Slash commands: Fix selector display during history navigation
1335- Resizes images before upload to prevent API size limit errors
1336- Added XDG_CONFIG_HOME support to configuration directory
1337- Performance optimizations for memory usage
1338- New attributes (terminal.type, language) in OpenTelemetry logging
1339
1340## 1.0.27
1341
1342- Streamable HTTP MCP servers are now supported
1343- Remote MCP servers (SSE and HTTP) now support OAuth
1344- MCP resources can now be @-mentioned
1345- /resume slash command to switch conversations within Claude Code
1346
1347## 1.0.25
1348
1349- Slash commands: moved "project" and "user" prefixes to descriptions
1350- Slash commands: improved reliability for command discovery
1351- Improved support for Ghostty
1352- Improved web search reliability
1353
1354## 1.0.24
1355
1356- Improved /mcp output
1357- Fixed a bug where settings arrays got overwritten instead of merged
1358
1359## 1.0.23
1360
1361- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1362- Released Python SDK: pip install claude-code-sdk to get started
1363
1364## 1.0.22
1365
1366- SDK: Renamed `total_cost` to `total_cost_usd`
1367
1368## 1.0.21
1369
1370- Improved editing of files with tab-based indentation
1371- Fix for tool_use without matching tool_result errors
1372- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1373
1374## 1.0.18
1375
1376- Added --add-dir CLI argument for specifying additional working directories
1377- Added streaming input support without require -p flag
1378- Improved startup performance and session storage performance
1379- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1380- Added detailed MCP server tools display (/mcp)
1381- MCP authentication and permission improvements
1382- Added auto-reconnection for MCP SSE connections on disconnect
1383- Fixed issue where pasted content was lost when dialogs appeared
1384
1385## 1.0.17
1386
1387- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1388- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1389- MCP server list UI improvements
1390- Update Claude Code process title to display "claude" instead of "node"
1391
1392## 1.0.11
1393
1394- Claude Code can now also be used with a Claude Pro subscription
1395- Added /upgrade for smoother switching to Claude Max plans
1396- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1397- Improved shell configuration error handling
1398- Improved todo list handling during compaction
1399
1400## 1.0.10
1401
1402- Added markdown table support
1403- Improved streaming performance
1404
1405## 1.0.8
1406
1407- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1408- Increased default otel interval from 1s -> 5s
1409- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1410- Fixed a regression where search tools unnecessarily asked for permissions
1411- Added support for triggering thinking non-English languages
1412- Improved compacting UI
1413
1414## 1.0.7
1415
1416- Renamed /allowed-tools -> /permissions
1417- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1418- Deprecated claude config commands in favor of editing settings.json
1419- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1420- Improved error handling for /install-github-app
1421- Bugfixes, UI polish, and tool reliability improvements
1422
1423## 1.0.6
1424
1425- Improved edit reliability for tab-indented files
1426- Respect CLAUDE_CONFIG_DIR everywhere
1427- Reduced unnecessary tool permission prompts
1428- Added support for symlinks in @file typeahead
1429- Bugfixes, UI polish, and tool reliability improvements
1430
1431## 1.0.4
1432
1433- Fixed a bug where MCP tool errors weren't being parsed correctly
1434
1435## 1.0.1
1436
1437- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1438- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1439- Updated documentation links and OAuth process descriptions
1440
1441## 1.0.0
1442
1443- Claude Code is now generally available
1444- Introducing Sonnet 4 and Opus 4 models
1445
1446## 0.2.125
1447
1448- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1449- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1450
1451## 0.2.117
1452
1453- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1454- Introduced settings.cleanupPeriodDays
1455- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1456- Introduced --debug mode
1457
1458## 0.2.108
1459
1460- You can now send messages to Claude while it works to steer Claude in real-time
1461- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1462- Fixed a bug where thinking was not working in -p mode
1463- Fixed a regression in /cost reporting
1464- Deprecated MCP wizard interface in favor of other MCP commands
1465- Lots of other bugfixes and improvements
1466
1467## 0.2.107
1468
1469- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1470
1471## 0.2.106
1472
1473- MCP SSE server configs can now specify custom headers
1474- Fixed a bug where MCP permission prompt didn't always show correctly
1475
1476## 0.2.105
1477
1478- Claude can now search the web
1479- Moved system & account status to /status
1480- Added word movement keybindings for Vim
1481- Improved latency for startup, todo tool, and file edits
1482
1483## 0.2.102
1484
1485- Improved thinking triggering reliability
1486- Improved @mention reliability for images and folders
1487- You can now paste multiple large chunks into one prompt
1488
1489## 0.2.100
1490
1491- Fixed a crash caused by a stack overflow error
1492- Made db storage optional; missing db support disables --continue and --resume
1493
1494## 0.2.98
1495
1496- Fixed an issue where auto-compact was running twice
1497
1498## 0.2.96
1499
1500- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1501
1502## 0.2.93
1503
1504- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1505- Claude now has access to a Todo list that helps it stay on track and be more organized
1506
1507## 0.2.82
1508
1509- Added support for --disallowedTools
1510- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1511
1512## 0.2.75
1513
1514- Hit Enter to queue up additional messages while Claude is working
1515- Drag in or copy/paste image files directly into the prompt
1516- @-mention files to directly add them to context
1517- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1518- Improved performance for filename auto-complete
1519
1520## 0.2.74
1521
1522- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1523- Task tool can now perform writes and run bash commands
1524
1525## 0.2.72
1526
1527- Updated spinner to indicate tokens loaded and tool usage
1528
1529## 0.2.70
1530
1531- Network commands like curl are now available for Claude to use
1532- Claude can now run multiple web queries in parallel
1533- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1534
1535## 0.2.69
1536
1537- Fixed UI glitches with improved Select component behavior
1538- Enhanced terminal output display with better text truncation logic
1539
1540## 0.2.67
1541
1542- Shared project permission rules can be saved in .claude/settings.json
1543
1544## 0.2.66
1545
1546- Print mode (-p) now supports streaming output via --output-format=stream-json
1547- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1548
1549## 0.2.63
1550
1551- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1552
1553## 0.2.61
1554
1555- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1556- Enhanced image detection for more reliable clipboard paste functionality
1557- Fixed an issue where ESC key could crash the conversation history selector
1558
1559## 0.2.59
1560
1561- Copy+paste images directly into your prompt
1562- Improved progress indicators for bash and fetch tools
1563- Bugfixes for non-interactive mode (-p)
1564
1565## 0.2.54
1566
1567- Quickly add to Memory by starting your message with '#'
1568- Press ctrl+r to see full output for long tool results
1569- Added support for MCP SSE transport
1570
1571## 0.2.53
1572
1573- New web fetch tool lets Claude view URLs that you paste in
1574- Fixed a bug with JPEG detection
1575
1576## 0.2.50
1577
1578- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1579
1580## 0.2.49
1581
1582- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1583
1584## 0.2.47
1585
1586- Press Tab to auto-complete file and folder names
1587- Press Shift + Tab to toggle auto-accept for file edits
1588- Automatic conversation compaction for infinite conversation length (toggle with /config)
1589
1590## 0.2.44
1591
1592- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1593
1594## 0.2.41
1595
1596- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1597- MCP server startup no longer blocks the app from starting up
1598
1599## 0.2.37
1600
1601- New /release-notes command lets you view release notes at any time
1602- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1603
1604## 0.2.36
1605
1606- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1607- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1608
1609## 0.2.34
1610
1611- Vim bindings for text input - enable with /vim or /config
1612
1613## 0.2.32
1614
1615- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1616- Fix for some PersistentShell issues
1617
1618## 0.2.31
1619
1620- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1621- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1622
1623## 0.2.30
1624
1625- Added ANSI color theme for better terminal compatibility
1626- Fixed issue where slash command arguments weren't being sent properly
1627- (Mac-only) API keys are now stored in macOS Keychain
1628
1629## 0.2.26
1630
1631- New /approved-tools command for managing tool permissions
1632- Word-level diff display for improved code readability
1633- Fuzzy matching for slash commands
1634
1635## 0.2.21
1636
1637- Fuzzy matching for /commands