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 8f0fe03 · on Feb 24, 2026 History
1,881 lines · 109.0 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.52
4
5- VS Code: Fixed extension crash on Windows ("command 'claude-vscode.editor.openLast' not found")
6
7## 2.1.51
8
9- Added `claude remote-control` subcommand for external builds, enabling local environment serving for all users.
10- Updated plugin marketplace default git timeout from 30s to 120s and added `CLAUDE_CODE_PLUGIN_GIT_TIMEOUT_MS` to configure.
11- Added support for custom npm registries and specific version pinning when installing plugins from npm sources
12- BashTool now skips login shell (`-l` flag) by default when a shell snapshot is available, improving command execution performance. Previously this required setting `CLAUDE_BASH_NO_LOGIN=true`.
13- Fixed a security issue where `statusLine` and `fileSuggestion` hook commands could execute without workspace trust acceptance in interactive mode.
14- Tool results larger than 50K characters are now persisted to disk (previously 100K). This reduces context window usage and improves conversation longevity.
15- Fixed a bug where duplicate `control_response` messages (e.g. from WebSocket reconnects) could cause API 400 errors by pushing duplicate assistant messages into the conversation.
16- Added `CLAUDE_CODE_ACCOUNT_UUID`, `CLAUDE_CODE_USER_EMAIL`, and `CLAUDE_CODE_ORGANIZATION_UUID` environment variables for SDK callers to provide account info synchronously, eliminating a race condition where early telemetry events lacked account metadata.
17- Fixed slash command autocomplete crashing when a plugin's SKILL.md description is a YAML array or other non-string type
18- The `/model` picker now shows human-readable labels (e.g., "Sonnet 4.5") instead of raw model IDs for pinned model versions, with an upgrade hint when a newer version is available.
19
20## 2.1.50
21
22- Added support for `startupTimeout` configuration for LSP servers
23- Added `WorktreeCreate` and `WorktreeRemove` hook events, enabling custom VCS setup and teardown when agent worktree isolation creates or removes worktrees.
24- Fixed a bug where resumed sessions could be invisible when the working directory involved symlinks, because the session storage path was resolved at different times during startup. Also fixed session data loss on SSH disconnect by flushing session data before hooks and analytics in the graceful shutdown sequence.
25- Linux: Fixed native modules not loading on systems with glibc older than 2.30 (e.g., RHEL 8)
26- Fixed memory leak in agent teams where completed teammate tasks were never garbage collected from session state
27- Fixed `CLAUDE_CODE_SIMPLE` to fully strip down skills, session memory, custom agents, and CLAUDE.md token counting
28- Fixed `/mcp reconnect` freezing the CLI when given a server name that doesn't exist
29- Fixed memory leak where completed task state objects were never removed from AppState
30- Added support for `isolation: worktree` in agent definitions, allowing agents to declaratively run in isolated git worktrees.
31- `CLAUDE_CODE_SIMPLE` mode now also disables MCP tools, attachments, hooks, and CLAUDE.md file loading for a fully minimal experience.
32- Fixed bug where MCP tools were not discovered when tool search is enabled and a prompt is passed in as a launch argument
33- Improved memory usage during long sessions by clearing internal caches after compaction
34- Added `claude agents` CLI command to list all configured agents
35- Improved memory usage during long sessions by clearing large tool results after they have been processed
36- Fixed a memory leak where LSP diagnostic data was never cleaned up after delivery, causing unbounded memory growth in long sessions
37- Fixed a memory leak where completed task output was not freed from memory, reducing memory usage in long sessions with many tasks
38- Improved startup performance for headless mode (`-p` flag) by deferring Yoga WASM and UI component imports
39- Fixed prompt suggestion cache regression that reduced cache hit rates
40- Fixed unbounded memory growth in long sessions by capping file history snapshots
41- Added `CLAUDE_CODE_DISABLE_1M_CONTEXT` environment variable to disable 1M context window support
42- Opus 4.6 (fast mode) now includes the full 1M context window
43- VSCode: Added `/extra-usage` command support in VS Code sessions
44- Fixed memory leak where TaskOutput retained recent lines after cleanup
45- Fixed memory leak in CircularBuffer where cleared items were retained in the backing array
46- Fixed memory leak in shell command execution where ChildProcess and AbortController references were retained after cleanup
47
48## 2.1.49
49
50- Improved MCP OAuth authentication with step-up auth support and discovery caching, reducing redundant network requests during server connections
51- Added `--worktree` (`-w`) flag to start Claude in an isolated git worktree
52- Subagents support `isolation: "worktree"` for working in a temporary git worktree
53- Added Ctrl+F keybinding to kill background agents (two-press confirmation)
54- Agent definitions support `background: true` to always run as a background task
55- Plugins can ship `settings.json` for default configuration
56- Fixed file-not-found errors to suggest corrected paths when the model drops the repo folder
57- Fixed Ctrl+C and ESC being silently ignored when background agents are running and the main thread is idle. Pressing twice within 3 seconds now kills all background agents.
58- Fixed prompt suggestion cache regression that reduced cache hit rates.
59- Fixed `plugin enable` and `plugin disable` to auto-detect the correct scope when `--scope` is not specified, instead of always defaulting to user scope
60- Simple mode (`CLAUDE_CODE_SIMPLE`) now includes the file edit tool in addition to the Bash tool, allowing direct file editing in simple mode.
61- Permission suggestions are now populated when safety checks trigger an ask response, enabling SDK consumers to display permission options
62- Sonnet 4.5 with 1M context is being removed from the Max plan in favor of our frontier Sonnet 4.6 model, which now has 1M context. Please switch in /model.
63- Fixed verbose mode not updating thinking block display when toggled via `/config` — memo comparators now correctly detect verbose changes
64- Fixed unbounded WASM memory growth during long sessions by periodically resetting the tree-sitter parser
65- Fixed potential rendering issues caused by stale yoga layout references
66- Improved performance in non-interactive mode (`-p`) by skipping unnecessary API calls during startup
67- Improved performance by caching authentication failures for HTTP and SSE MCP servers, avoiding repeated connection attempts to servers requiring auth
68- Fixed unbounded memory growth during long-running sessions caused by Yoga WASM linear memory never shrinking
69- SDK model info now includes `supportsEffort`, `supportedEffortLevels`, and `supportsAdaptiveThinking` fields so consumers can discover model capabilities.
70- Added `ConfigChange` hook event that fires when configuration files change during a session, enabling enterprise security auditing and optional blocking of settings changes.
71- Improved startup performance by caching MCP auth failures to avoid redundant connection attempts
72- Improved startup performance by reducing HTTP calls for analytics token counting
73- Improved startup performance by batching MCP tool token counting into a single API call
74- Fixed `disableAllHooks` setting to respect managed settings hierarchy — non-managed settings can no longer disable managed hooks set by policy (#26637)
75- Fixed `--resume` session picker showing raw XML tags for sessions that start with commands like `/clear`. Now correctly falls through to the session ID fallback.
76- Improved permission prompts for path safety and working directory blocks to show the reason for the restriction instead of a bare prompt with no context
77
78## 2.1.47
79
80- Fixed FileWriteTool line counting to preserve intentional trailing blank lines instead of stripping them with `trimEnd()`.
81- Fixed Windows terminal rendering bugs caused by `os.EOL` (`\r\n`) in display code — line counts now show correct values instead of always showing 1 on Windows.
82- Improved VS Code plan preview: auto-updates as Claude iterates, enables commenting only when the plan is ready for review, and keeps the preview open when rejecting so Claude can revise.
83- Fixed a bug where bold and colored text in markdown output could shift to the wrong characters on Windows due to `\r\n` line endings.
84- Fixed compaction failing when conversation contains many PDF documents by stripping document blocks alongside images before sending to the compaction API (anthropics/claude-code#26188)
85- Improved memory usage in long-running sessions by releasing API stream buffers, agent context, and skill state after use
86- Improved startup performance by deferring SessionStart hook execution, reducing time-to-interactive by ~500ms.
87- Fixed an issue where bash tool output was silently discarded on Windows when using MSYS2 or Cygwin shells.
88- Improved performance of `@` file mentions - file suggestions now appear faster by pre-warming the index on startup and using session-based caching with background refresh.
89- Improved memory usage by trimming agent task message history after tasks complete
90- Improved memory usage during long agent sessions by eliminating O(n²) message accumulation in progress updates
91- Fixed the bash permission classifier to validate that returned match descriptions correspond to actual input rules, preventing hallucinated descriptions from incorrectly granting permissions
92- Fixed user-defined agents only loading one file on NFS/FUSE filesystems that report zero inodes (anthropics/claude-code#26044)
93- Fixed plugin agent skills silently failing to load when referenced by bare name instead of fully-qualified plugin name (anthropics/claude-code#25834)
94- Search patterns in collapsed tool results are now displayed in quotes for clarity
95- Windows: Fixed CWD tracking temp files never being cleaned up, causing them to accumulate indefinitely (anthropics/claude-code#17600)
96- Use `ctrl+f` to kill all background agents instead of double-pressing ESC. Background agents now continue running when you press ESC to cancel the main thread, giving you more control over agent lifecycle.
97- Fixed API 400 errors ("thinking blocks cannot be modified") that occurred in sessions with concurrent agents, caused by interleaved streaming content blocks preventing proper message merging.
98- Simplified teammate navigation to use only Shift+Down (with wrapping) instead of both Shift+Up and Shift+Down.
99- Fixed an issue where a single file write/edit error would abort all other parallel file write/edit operations. Independent file mutations now complete even when a sibling fails.
100- Added `last_assistant_message` field to Stop and SubagentStop hook inputs, providing the final assistant response text so hooks can access it without parsing transcript files.
101- Fixed custom session titles set via `/rename` being lost after resuming a conversation (anthropics/claude-code#23610)
102- Fixed collapsed read/search hint text overflowing on narrow terminals by truncating from the start.
103- Fixed an issue where bash commands with backslash-newline continuation lines (e.g., long commands split across multiple lines with `\`) would produce spurious empty arguments, potentially breaking command execution.
104- Fixed built-in slash commands (`/help`, `/model`, `/compact`, etc.) being hidden from the autocomplete dropdown when many user skills are installed (anthropics/claude-code#22020)
105- Fixed MCP servers not appearing in the MCP Management Dialog after deferred loading
106- Fixed session name persisting in status bar after `/clear` command (anthropics/claude-code#26082)
107- Fixed crash when a skill's `name` or `description` in SKILL.md frontmatter is a bare number (e.g., `name: 3000`) — the value is now properly coerced to a string (anthropics/claude-code#25837)
108- Fixed /resume silently dropping sessions when the first message exceeds 16KB or uses array-format content (anthropics/claude-code#25721)
109- Added `chat:newline` keybinding action for configurable multi-line input (anthropics/claude-code#26075)
110- Added `added_dirs` to the statusline JSON `workspace` section, exposing directories added via `/add-dir` to external scripts (anthropics/claude-code#26096)
111- Fixed `claude doctor` misclassifying mise and asdf-managed installations as native installs (anthropics/claude-code#26033)
112- Fixed zsh heredoc failing with "read-only file system" error in sandboxed commands (anthropics/claude-code#25990)
113- Fixed agent progress indicator showing inflated tool use count (anthropics/claude-code#26023)
114- Fixed image pasting not working on WSL2 systems where Windows copies images as BMP format (anthropics/claude-code#25935)
115- Fixed background agent results returning raw transcript data instead of the agent's final answer (anthropics/claude-code#26012)
116- Fixed Warp terminal incorrectly prompting for Shift+Enter setup when it supports it natively (anthropics/claude-code#25957)
117- Fixed CJK wide characters causing misaligned timestamps and layout elements in the TUI (anthropics/claude-code#26084)
118- Fixed custom agent `model` field in `.claude/agents/*.md` being ignored when spawning team teammates (anthropics/claude-code#26064)
119- Fixed plan mode being lost after context compaction, causing the model to switch from planning to implementation mode (anthropics/claude-code#26061)
120- Fixed `alwaysThinkingEnabled: true` in settings.json not enabling thinking mode on Bedrock and Vertex providers (anthropics/claude-code#26074)
121- Fixed `tool_decision` OTel telemetry event not being emitted in headless/SDK mode (anthropics/claude-code#26059)
122- Fixed session name being lost after context compaction — renamed sessions now preserve their custom title through compaction (anthropics/claude-code#26121)
123- Increased initial session count in resume picker from 10 to 50 for faster session discovery (anthropics/claude-code#26123)
124- Windows: fixed worktree session matching when drive letter casing differs (anthropics/claude-code#26123)
125- Fixed `/resume <session-id>` failing to find sessions whose first message exceeds 16KB (anthropics/claude-code#25920)
126- Fixed "Always allow" on multiline bash commands creating invalid permission patterns that corrupt settings (anthropics/claude-code#25909)
127- Fixed React crash (error #31) when a skill's `argument-hint` in SKILL.md frontmatter uses YAML sequence syntax (e.g., `[topic: foo | bar]`) — the value is now properly coerced to a string (anthropics/claude-code#25826)
128- Fixed crash when using `/fork` on sessions that used web search — null entries in search results from transcript deserialization are now handled gracefully (anthropics/claude-code#25811)
129- Fixed read-only git commands triggering FSEvents file watcher loops on macOS by adding --no-optional-locks flag (anthropics/claude-code#25750)
130- Fixed custom agents and skills not being discovered when running from a git worktree — project-level `.claude/agents/` and `.claude/skills/` from the main repository are now included (anthropics/claude-code#25816)
131- Fixed non-interactive subcommands like `claude doctor` and `claude plugin validate` being blocked inside nested Claude sessions (anthropics/claude-code#25803)
132- Windows: Fixed the same CLAUDE.md file being loaded twice when drive letter casing differs between paths (anthropics/claude-code#25756)
133- Fixed inline code spans in markdown being incorrectly parsed as bash commands (anthropics/claude-code#25792)
134- Fixed teammate spinners not respecting custom spinnerVerbs from settings (anthropics/claude-code#25748)
135- Fixed shell commands permanently failing after a command deletes its own working directory (anthropics/claude-code#26136)
136- Fixed hooks (PreToolUse, PostToolUse) silently failing to execute on Windows by using Git Bash instead of cmd.exe (anthropics/claude-code#25981)
137- Fixed LSP `findReferences` and other location-based operations returning results from gitignored files (e.g., `node_modules/`, `venv/`) (anthropics/claude-code#26051)
138- Moved config backup files from home directory root to `~/.claude/backups/` to reduce home directory clutter (anthropics/claude-code#26130)
139- Fixed sessions with large first prompts (>16KB) disappearing from the /resume list (anthropics/claude-code#26140)
140- Fixed shell functions with double-underscore prefixes (e.g., `__git_ps1`) not being preserved across shell sessions (anthropics/claude-code#25824)
141- Fixed spinner showing "0 tokens" counter before any tokens have been received (anthropics/claude-code#26105)
142- VSCode: Fixed conversation messages appearing dimmed while the AskUserQuestion dialog is open (anthropics/claude-code#26078)
143- Fixed background tasks failing in git worktrees due to remote URL resolution reading from worktree-specific gitdir instead of the main repository config (anthropics/claude-code#26065)
144- Fixed Right Alt key leaving visible `[25~` escape sequence residue in the input field on Windows/Git Bash terminals (anthropics/claude-code#25943)
145- The `/rename` command now updates the terminal tab title by default (anthropics/claude-code#25789)
146- Fixed Edit tool silently corrupting Unicode curly quotes (\u201c\u201d \u2018\u2019) by replacing them with straight quotes when making edits (anthropics/claude-code#26141)
147- Fixed OSC 8 hyperlinks only being clickable on the first line when link text wraps across multiple terminal lines.
148
149## 2.1.46
150
151- Fixed orphaned CC processes after terminal disconnect on macOS
152- Added support for using claude.ai MCP connectors in Claude Code
153
154## 2.1.45
155
156- Added support for Claude Sonnet 4.6
157- Added support for reading `enabledPlugins` and `extraKnownMarketplaces` from `--add-dir` directories
158- Added `spinnerTipsOverride` setting to customize spinner tips — configure `tips` with an array of custom tip strings, and optionally set `excludeDefault: true` to show only your custom tips instead of the built-in ones
159- Added `SDKRateLimitInfo` and `SDKRateLimitEvent` types to the SDK, enabling consumers to receive rate limit status updates including utilization, reset times, and overage information
160- Fixed Agent Teams teammates failing on Bedrock, Vertex, and Foundry by propagating API provider environment variables to tmux-spawned processes (anthropics/claude-code#23561)
161- Fixed sandbox "operation not permitted" errors when writing temporary files on macOS by using the correct per-user temp directory (anthropics/claude-code#21654)
162- Fixed Task tool (backgrounded agents) crashing with a `ReferenceError` on completion (anthropics/claude-code#22087)
163- Fixed autocomplete suggestions not being accepted on Enter when images are pasted in the input
164- Fixed skills invoked by subagents incorrectly appearing in main session context after compaction
165- Fixed excessive `.claude.json.backup` files accumulating on every startup
166- Fixed plugin-provided commands, agents, and hooks not being available immediately after installation without requiring a restart
167- Improved startup performance by removing eager loading of session history for stats caching
168- Improved memory usage for shell commands that produce large output — RSS no longer grows unboundedly with command output size
169- Improved collapsed read/search groups to show the current file or search pattern being processed beneath the summary line while active
170- [VSCode] Improved permission destination choice (project/user/session) to persist across sessions
171
172## 2.1.44
173
174- Fixed ENAMETOOLONG errors for deeply-nested directory paths
175- Fixed auth refresh errors
176
177## 2.1.43
178
179- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
180- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
181- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
182
183## 2.1.42
184
185- Improved startup performance by deferring Zod schema construction
186- Improved prompt cache hit rates by moving date out of system prompt
187- Added one-time Opus 4.6 effort callout for eligible users
188- Fixed /resume showing interrupt messages as session titles
189- Fixed image dimension limit errors to suggest /compact
190
191## 2.1.41
192
193- Added guard against launching Claude Code inside another Claude Code session
194- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
195- Fixed a crash when MCP tools return image content during streaming
196- Fixed /resume session previews showing raw XML tags instead of readable command names
197- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
198- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
199- Fixed hook blocking errors (exit code 2) not showing stderr to the user
200- Added `speed` attribute to OTel events and trace spans for fast mode visibility
201- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
202- Added Windows ARM64 (win32-arm64) native binary support
203- Improved `/rename` to auto-generate session name from conversation context when called without arguments
204- Improved narrow terminal layout for prompt footer
205- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
206- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
207- Fixed background task notifications not being delivered in streaming Agent SDK mode
208- Fixed cursor jumping to end on each keystroke in classifier rule input
209- Fixed markdown link display text being dropped for raw URL
210- Fixed auto-compact failure error notifications being shown to users
211- Fixed permission wait time being included in subagent elapsed time display
212- Fixed proactive ticks firing while in plan mode
213- Fixed clear stale permission rules when settings change on disk
214- Fixed hook blocking errors showing stderr content in UI
215
216## 2.1.39
217
218- Improved terminal rendering performance
219- Fixed fatal errors being swallowed instead of displayed
220- Fixed process hanging after session close
221- Fixed character loss at terminal screen boundary
222- Fixed blank lines in verbose transcript view
223
224## 2.1.38
225
226- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
227- Fixed Tab key queueing slash commands instead of autocompleting
228- Fixed bash permission matching for commands using environment variable wrappers
229- Fixed text between tool uses disappearing when not using streaming
230- Fixed duplicate sessions when resuming in VS Code extension
231- Improved heredoc delimiter parsing to prevent command smuggling
232- Blocked writes to `.claude/skills` directory in sandbox mode
233
234## 2.1.37
235
236- Fixed an issue where /fast was not immediately available after enabling /extra-usage
237
238## 2.1.36
239
240- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
241
242## 2.1.34
243
244- Fixed a crash when agent teams setting changed between renders
245- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
246
247## 2.1.33
248
249- Fixed agent teammate sessions in tmux to send and receive messages
250- Fixed warnings about agent teams not being available on your current plan
251- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
252- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
253- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
254- Added plugin name to skill descriptions and `/skills` menu for better discoverability
255- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
256- 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
257- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
258- 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
259- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
260- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
261- Errors from invalid managed settings are now surfaced
262- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
263- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
264- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
265
266## 2.1.32
267
268- Claude Opus 4.6 is now available!
269- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
270- Claude now automatically records and recalls memories as it works
271- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
272- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
273- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
274- Updated --resume to re-use --agent value specified in previous conversation by default.
275- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
276- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
277- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
278- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
279- VSCode: Added spinner when loading past conversations list
280
281## 2.1.31
282
283- Added session resume hint on exit, showing how to continue your conversation later
284- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
285- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
286- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
287- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
288- 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
289- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
290- 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
291- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
292- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
293- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
294
295## 2.1.30
296
297- 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.
298- 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`.
299- Added `/debug` for Claude to help troubleshoot the current session
300- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
301- Added token count, tool uses, and duration metrics to Task tool results
302- Added reduced motion mode to the config
303- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
304- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
305- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
306- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
307- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
308- Fixed permission dialogs stealing focus while actively typing
309- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
310- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
311- 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
312- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
313- Changed `/model` to execute immediately instead of being queued
314- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
315- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
316
317## 2.1.29
318
319- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
320
321## 2.1.27
322
323- Added tool call failures and denials to debug logs
324- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
325- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
326- Sessions are now automatically linked to PRs when created via `gh pr create`
327- Fixed /context command not displaying colored output
328- Fixed status bar duplicating background task indicator when PR status was shown
329- Windows: Fixed bash command execution failing for users with `.bashrc` files
330- Windows: Fixed console windows flashing when spawning child processes
331- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
332
333## 2.1.25
334
335- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
336
337## 2.1.23
338
339- Added customizable spinner verbs setting (`spinnerVerbs`)
340- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
341- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
342- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
343- Fixed pending async hooks not being cancelled when headless streaming sessions ended
344- Fixed tab completion not updating the input field when accepting a suggestion
345- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
346- Improved terminal rendering performance with optimized screen data layout
347- Changed Bash commands to show timeout duration alongside elapsed time
348- Changed merged pull requests to show a purple status indicator in the prompt footer
349- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
350
351## 2.1.22
352
353- Fixed structured outputs for non-interactive (-p) mode
354
355## 2.1.21
356
357- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
358- Fixed shell completion cache files being truncated on exit
359- Fixed API errors when resuming sessions that were interrupted during tool execution
360- Fixed auto-compact triggering too early on models with large output token limits
361- Fixed task IDs potentially being reused after deletion
362- Fixed file search not working in VS Code extension on Windows
363- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
364- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
365- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
366- [VSCode] Fixed message action buttons having incorrect background colors
367
368## 2.1.20
369
370- Added arrow key history navigation in vim normal mode when cursor cannot move further
371- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
372- 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
373- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
374- Added ability to delete tasks via the `TaskUpdate` tool
375- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
376- Fixed agents sometimes ignoring user messages sent while actively working on a task
377- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
378- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
379- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
380- Fixed draft prompt being lost when pressing UP arrow to navigate command history
381- Fixed ghost text flickering when typing slash commands mid-input
382- Fixed marketplace source removal not properly deleting settings
383- Fixed duplicate output in some commands like `/context`
384- Fixed task list sometimes showing outside the main conversation view
385- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
386- Fixed crashes when cancelling tool use
387- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
388- Improved thinking status text with a subtle shimmer animation
389- Improved task list to dynamically adjust visible items based on terminal height
390- Improved fork conversation hint to show how to resume the original session
391- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
392- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
393- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
394- Changed the `/copy` command to be available to all users
395- Changed background agents to prompt for tool permissions before launching
396- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
397- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
398
399## 2.1.19
400
401- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
402- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
403- Fixed crashes on processors without AVX instruction support
404- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
405- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
406- Fixed resuming sessions by custom title not working when run from a different directory
407- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
408- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
409- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
410- Fixed file write preview omitting empty lines
411- Changed skills without additional permissions or hooks to be allowed without requiring approval
412- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
413- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
414- [VSCode] Enabled session forking and rewind functionality for all users
415
416## 2.1.18
417
418- 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
419
420## 2.1.17
421
422- Fixed crashes on processors without AVX instruction support
423
424## 2.1.16
425
426- Added new task management system, including new capabilities like dependency tracking
427- [VSCode] Added native plugin management support
428- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
429- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
430- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
431- Fixed session titles on the resume screen not respecting the user's language setting
432- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
433
434## 2.1.15
435
436- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
437- Improved UI rendering performance with React Compiler
438- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
439- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
440
441## 2.1.14
442
443- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
444- Added search to installed plugins list - type to filter by name or description
445- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
446- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
447- Fixed memory issues that could cause crashes when running parallel subagents
448- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
449- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
450- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
451- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
452- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
453- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
454- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
455- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
456- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
457- Improved backspace to delete pasted text as a single token instead of one character at a time
458- [VSCode] Added `/usage` command to display current plan usage
459
460## 2.1.12
461
462- Fixed message rendering bug
463
464## 2.1.11
465
466- Fixed excessive MCP connection requests for HTTP/SSE transports
467
468## 2.1.10
469
470- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
471- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
472- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
473- Improved startup to capture keystrokes typed before the REPL is fully ready
474- Improved file suggestions to show as removable attachments instead of inserting text when accepted
475- [VSCode] Added install count display to plugin listings
476- [VSCode] Added trust warning when installing plugins
477
478## 2.1.9
479
480- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
481- Added `plansDirectory` setting to customize where plan files are stored
482- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
483- Added session URL attribution to commits and PRs created from web sessions
484- Added support for `PreToolUse` hooks to return `additionalContext` to the model
485- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
486- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
487- Fixed MCP server reconnection hanging when cached connection promise never resolves
488- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
489
490## 2.1.7
491
492- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
493- Added ability to provide feedback when accepting permission prompts
494- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
495- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
496- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
497- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
498- 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)
499- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
500- Fixed terminal title animation jitter by using fixed-width braille characters
501- Fixed plugins with git submodules not being fully initialized when installed
502- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
503- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
504- 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.
505- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
506- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
507
508## 2.1.6
509
510- Added search functionality to `/config` command for quickly filtering settings
511- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
512- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
513- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
514- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
515- Added an error display when the editor fails during Ctrl+G
516- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
517- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
518- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
519- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
520- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
521- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
522- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
523- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
524- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
525- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
526- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
527- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
528- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
529- Improved @ autocomplete with icons for different suggestion types and single-line formatting
530- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
531- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
532- Changed terminal title to "Claude Code" on startup for better window identification
533- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
534- [VSCode] Fixed usage indicator not updating after manual compact
535
536## 2.1.5
537
538- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
539
540## 2.1.4
541
542- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
543- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
544
545## 2.1.3
546
547- Merged slash commands and skills, simplifying the mental model with no change in behavior
548- Added release channel (`stable` or `latest`) toggle to `/config`
549- 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
550- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
551- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
552- Fixed mismatch between background task count in status bar and items shown in tasks dialog
553- Fixed sub-agents using the wrong model during conversation compaction
554- Fixed web search in sub-agents using incorrect model
555- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
556- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
557- Improved slash command suggestion readability by truncating long descriptions to 2 lines
558- Changed tool hook execution timeout from 60 seconds to 10 minutes
559- [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)
560
561## 2.1.2
562
563- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
564- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
565- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
566- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
567- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
568- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
569- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
570- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
571- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
572- Fixed updates incorrectly claiming another installation is in progress
573- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
574- Fixed remote session URL and teleport being broken when using `/tasks` command
575- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
576- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
577- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
578- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
579- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
580- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
581- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
582- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
583- [SDK] Changed minimum zod peer dependency to ^4.0.0
584- [VSCode] Fixed usage display not updating after manual compact
585
586## 2.1.0
587
588- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
589- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
590- Added support for `agent` field in skills to specify agent type for execution
591- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
592- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
593- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
594- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
595- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
596- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
597- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
598- Fixed Esc key with queued prompts to only move them to input without canceling the running task
599- Reduced permission prompts for complex bash commands
600- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
601- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
602- Fixed plugin path resolution for file-based marketplace sources
603- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
604- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
605- Fixed Claude in Chrome support for WSL environments
606- Fixed Windows native installer silently failing when executable creation fails
607- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
608- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
609- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
610- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
611- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
612- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
613- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
614- Added hooks support for skill and slash command frontmatter
615- 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
616- Added `/plan` command shortcut to enable plan mode directly from the prompt
617- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
618- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
619- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
620- Added support for `once: true` config for hooks
621- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
622- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
623- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
624- Added left/right arrow key navigation for cycling through tabs in dialogs
625- Added real-time thinking block display in Ctrl+O transcript mode
626- Added filepath to full output in background bash task details dialog
627- Added Skills as a separate category in the context visualization
628- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
629- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
630- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
631- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
632- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
633- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
634- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
635- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
636- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
637- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
638- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
639- Fixed commands with `$()` command substitution failing with parse errors
640- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
641- 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)
642- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
643- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
644- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
645- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
646- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
647- Fixed subagents sometimes not inheriting the parent's model by default
648- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
649- Fixed duplicate Bash commands appearing in permission request option labels
650- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
651- Fixed background task completion notifications to appear proactively with bullet point
652- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
653- Fixed cursor disappearing after dismissing permission dialogs
654- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
655- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
656- Fixed images being silently dropped when queueing messages while backgrounding a task
657- Fixed large pasted images failing with "Image was too large" error
658- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
659- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
660- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
661- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
662- Fixed spinner token counter to properly accumulate tokens from subagents during execution
663- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
664- Fixed race condition where LSP tool could return "no server available" during startup
665- Fixed feedback submission hanging indefinitely when network requests timeout
666- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
667- Fixed hook success message showing trailing colon when hook has no output
668- Multiple optimizations to improve startup performance
669- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
670- Improved performance when reading Jupyter notebooks with many cells
671- Improved reliability for piped input like `cat refactor.md | claude`
672- Improved reliability for AskQuestion tool
673- Improved sed in-place edit commands to render as file edits with diff preview
674- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
675- Improved compaction reliability
676- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
677- Improved skills to show progress while executing, displaying tool uses as they happen
678- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
679- Improved skill suggestions to prioritize recently and frequently used skills
680- Improved spinner feedback when waiting for the first response token
681- Improved token count display in spinner to include tokens from background agents
682- Improved incremental output for async agents to give the main thread more control and visibility
683- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
684- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
685- Improved macOS screenshot paste reliability with TIFF format support
686- Improved `/stats` output
687- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
688- Changed "Interrupted" message color from red to grey for a less alarming appearance
689- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
690- Removed underline styling from image reference links
691- [SDK] Changed minimum zod peer dependency to ^4.0.0
692- [VSCode] Added currently selected model name to the context menu
693- [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")
694- [VSCode] Fixed paragraph breaks not rendering in markdown content
695- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
696- [Windows] Fixed issue with improper rendering
697
698## 2.0.76
699
700- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
701
702## 2.0.75
703
704- Minor bugfixes
705
706## 2.0.74
707
708- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
709- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
710- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
711- Added syntax highlighting info to theme picker
712- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
713- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
714- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
715- Fixed a potential crash when syntax highlighting isn't initialized correctly
716- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
717- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
718- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
719- [Windows] Fixed issue with improper rendering
720- [VSCode] Added gift tag pictogram for year-end promotion message
721
722## 2.0.73
723
724- Added clickable `[Image #N]` links that open attached images in the default viewer
725- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
726- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
727- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
728- Fixed slow input history cycling and race condition that could overwrite text after message submission
729- Improved `/theme` command to open theme picker directly
730- Improved theme picker UI
731- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
732- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
733
734## 2.0.72
735
736- 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
737- Reduced terminal flickering
738- Added scannable QR code to mobile app tip for quick app downloads
739- Added loading indicator when resuming conversations for better feedback
740- Fixed `/context` command not respecting custom system prompts in non-interactive mode
741- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
742- Improved @ mention file suggestion speed (~3× faster in git repositories)
743- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
744- Improved settings validation errors to be more prominent
745- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
746
747## 2.0.71
748
749- Added /config toggle to enable/disable prompt suggestions
750- Added `/settings` as an alias for the `/config` command
751- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
752- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
753- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
754- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
755- New syntax highlighting engine for native build
756
757## 2.0.70
758
759- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
760- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
761- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
762- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
763- Fixed input being cleared when processing queued commands while the user was typing
764- Fixed prompt suggestions replacing typed input when pressing Tab
765- Fixed diff view not updating when terminal is resized
766- Improved memory usage by 3x for large conversations
767- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
768- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
769- Fix thinking mode toggle in /config not persisting correctly
770- Improve UI for file creation permission dialog
771
772## 2.0.69
773
774- Minor bugfixes
775
776## 2.0.68
777
778- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
779- Fixed a bug where disallowed MCP tools were visible to the model
780- Fixed an issue where steering messages could be lost while a subagent is working
781- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
782- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
783- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
784
785## 2.0.67
786
787- Thinking mode is now enabled by default for Opus 4.5
788- Thinking mode configuration has moved to /config
789- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
790- Show reason why autoupdater is disabled in `/doctor`
791- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
792- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
793- Fixed scroll position resetting after deleting a permission rule in `/permissions`
794- 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
795- Fixed `claude install --force` not bypassing stale lock files
796- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
797- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
798
799## 2.0.65
800
801- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
802- Added context window information to status line input
803- Added `fileSuggestion` setting for custom `@` file search commands
804- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
805- Fixed prompt not being saved to history when aborting a query with Escape
806- Fixed Read tool image handling to identify format from bytes instead of file extension
807
808## 2.0.64
809
810- Made auto-compacting instant
811- Agents and bash commands can run asynchronously and send messages to wake up the main agent
812- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
813- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
814- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
815- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
816- Fixed auto-loading .env when using native installer
817- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
818- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
819- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
820- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
821- Bedrock: Improve efficiency of token counting
822- Bedrock: Add support for `aws login` AWS Management Console credentials
823- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
824
825## 2.0.62
826
827- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
828- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
829- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
830- Fixed slash command selection not working when multiple commands share the same name
831- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
832- Fixed running versions getting removed because lock file incorrectly going stale
833- Fixed IDE diff tab not closing when rejecting file changes
834
835## 2.0.61
836
837- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
838
839## 2.0.60
840
841- Added background agent support. Agents run in the background while you work
842- Added --disable-slash-commands CLI flag to disable all slash commands
843- Added model name to "Co-Authored-By" commit messages
844- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
845- Updated Fetch to skip summarization for pre-approved websites
846- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
847
848## 2.0.59
849
850- Added --agent CLI flag to override the agent setting for the current session
851- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
852- VS Code: Fixed .claude.json config file being read from incorrect location
853
854## 2.0.58
855
856- Pro users now have access to Opus 4.5 as part of their subscription!
857- Fixed timer duration showing "11m 60s" instead of "12m 0s"
858- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
859
860## 2.0.57
861
862- Added feedback input when rejecting plans, allowing users to tell Claude what to change
863- VSCode: Added streaming message support for real-time response display
864
865## 2.0.56
866
867- Added setting to enable/disable terminal progress bar (OSC 9;4)
868- 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.
869
870## 2.0.55
871
872- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
873- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
874- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
875- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
876
877## 2.0.54
878
879- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
880- Fix issue with excessive iTerm notifications
881
882## 2.0.52
883
884- Fixed duplicate message display when starting Claude with a command line argument
885- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
886- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
887- Permit some uses of `$!` in bash commands
888
889## 2.0.51
890
891- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
892- Introducing Claude Code for Desktop: https://claude.com/download
893- 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
894- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
895- Plan Mode now builds more precise plans and executes more thoroughly
896- Usage limit notifications now easier to understand
897- Switched `/usage` back to "% used"
898- Fixed handling of thinking errors
899- Fixed performance regression
900
901## 2.0.50
902
903- Fixed bug preventing calling MCP tools that have nested references in their input schemas
904- Silenced a noisy but harmless error during upgrades
905- Improved ultrathink text display
906- Improved clarity of 5-hour session limit warning message
907
908## 2.0.49
909
910- Added readline-style ctrl-y for pasting deleted text
911- Improved clarity of usage limit warning message
912- Fixed handling of subagent permissions
913
914## 2.0.47
915
916- Improved error messages and validation for `claude --teleport`
917- Improved error handling in `/usage`
918- Fixed race condition with history entry not getting logged at exit
919- Fixed Vertex AI configuration not being applied from `settings.json`
920
921## 2.0.46
922
923- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
924
925## 2.0.45
926
927- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
928- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
929- Send background tasks to Claude Code on the web by starting a message with `&`
930
931## 2.0.43
932
933- Added `permissionMode` field for custom agents
934- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
935- Added skills frontmatter field to declare skills to auto-load for subagents
936- Added the `SubagentStart` hook event
937- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
938- Fixed duplicate rendering of some messages in the UI
939- Fixed some visual flickers
940- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
941
942## 2.0.42
943
944- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
945
946## 2.0.41
947
948- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
949- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
950- Fixed incorrect labeling of user settings vs project settings in command descriptions
951- Fixed crash when plugin command hooks timeout during execution
952- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
953- Fixed broken security documentation links in trust dialogs and onboarding
954- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
955- ctrl-r history search landing on a slash command no longer cancels the search
956- SDK: Support custom timeouts for hooks
957- Allow more safe git commands to run without approval
958- Plugins: Added support for sharing and installing output styles
959- Teleporting a session from web will automatically set the upstream branch
960
961## 2.0.37
962
963- Fixed how idleness is computed for notifications
964- Hooks: Added matcher values for Notification hook events
965- Output Styles: Added `keep-coding-instructions` option to frontmatter
966
967## 2.0.36
968
969- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
970- Fixed queued messages being incorrectly executed as bash commands
971- Fixed input being lost when typing while a queued message is processed
972
973## 2.0.35
974
975- Improve fuzzy search results when searching commands
976- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
977- 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
978- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
979- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
980
981## 2.0.34
982
983- VSCode Extension: Added setting to configure the initial permission mode for new conversations
984- Improved file path suggestion performance with native Rust-based fuzzy finder
985- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
986- Fixed memory crash when reading or writing large files (especially base64-encoded images)
987
988## 2.0.33
989
990- Native binary installs now launch quicker.
991- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
992- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
993
994## 2.0.32
995
996- Un-deprecate output styles based on community feedback
997- Added `companyAnnouncements` setting for displaying announcements on startup
998- Fixed hook progress messages not updating correctly during PostToolUse hook execution
999
1000## 2.0.31
1001
1002- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
1003- Vertex: add support for Web Search on supported models
1004- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1005- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
1006- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
1007- Fixed plugin uninstall not removing plugins
1008
1009## 2.0.30
1010
1011- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
1012- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
1013- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
1014- Added prompt-based stop hooks
1015- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1016- Enabled SSE MCP servers on native build
1017- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
1018- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
1019- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
1020- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
1021- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
1022- Fixed bug that saved session permissions to local settings
1023- Fixed MCP tools not being available to sub-agents
1024- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
1025- Fixed delay when navigating through typeahead suggestions with arrow keys
1026- VSCode: Restored selection indicator in input footer showing current file or code selection status
1027
1028## 2.0.28
1029
1030- Plan mode: introduced new Plan subagent
1031- Subagents: claude can now choose to resume subagents
1032- Subagents: claude can dynamically choose the model used by its subagents
1033- SDK: added --max-budget-usd flag
1034- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
1035- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
1036- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
1037- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
1038- Various other bug fixes
1039
1040## 2.0.27
1041
1042- New UI for permission prompts
1043- Added current branch filtering and search to session resume screen for easier navigation
1044- Fixed directory @-mention causing "No assistant message found" error
1045- VSCode Extension: Add config setting to include .gitignored files in file searches
1046- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
1047
1048## 2.0.25
1049
1050- 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
1051
1052## 2.0.24
1053
1054- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
1055- Claude Code Web: Support for Web -> CLI teleport
1056- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
1057- Bedrock: Display awsAuthRefresh output when auth is required
1058
1059## 2.0.22
1060
1061- Fixed content layout shift when scrolling through slash commands
1062- IDE: Add toggle to enable/disable thinking.
1063- Fix bug causing duplicate permission prompts with parallel tool calls
1064- Add support for enterprise managed MCP allowlist and denylist
1065
1066## 2.0.21
1067
1068- Support MCP `structuredContent` field in tool responses
1069- Added an interactive question tool
1070- Claude will now ask you questions more often in plan mode
1071- Added Haiku 4.5 as a model option for Pro users
1072- Fixed an issue where queued commands don't have access to previous messages' output
1073
1074## 2.0.20
1075
1076- Added support for Claude Skills
1077
1078## 2.0.19
1079
1080- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
1081- Fixed a bug where Haiku was unnecessarily called in print mode
1082
1083## 2.0.17
1084
1085- Added Haiku 4.5 to model selector!
1086- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
1087- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
1088- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
1089- OTEL: support HTTP_PROXY and HTTPS_PROXY
1090- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
1091
1092## 2.0.15
1093
1094- Fixed bug with resuming where previously created files needed to be read again before writing
1095- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
1096
1097## 2.0.14
1098
1099- Fix @-mentioning MCP servers to toggle them on/off
1100- Improve permission checks for bash with inline env vars
1101- Fix ultrathink + thinking toggle
1102- Reduce unnecessary logins
1103- Document --system-prompt
1104- Several improvements to rendering
1105- Plugins UI polish
1106
1107## 2.0.13
1108
1109- Fixed `/plugin` not working on native build
1110
1111## 2.0.12
1112
1113- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
1114- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
1115- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
1116- `/plugin validate` command for validating plugin structure and configuration
1117- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
1118- Plugin documentation available at https://code.claude.com/docs/en/plugins
1119- Comprehensive error messages and diagnostics via `/doctor` command
1120- Avoid flickering in `/model` selector
1121- Improvements to `/help`
1122- Avoid mentioning hooks in `/resume` summaries
1123- Changes to the "verbose" setting in `/config` now persist across sessions
1124
1125## 2.0.11
1126
1127- Reduced system prompt size by 1.4k tokens
1128- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
1129- Fixed Opus fallback rate limit errors appearing incorrectly
1130- Fixed /add-dir command selecting wrong default tab
1131
1132## 2.0.10
1133
1134- Rewrote terminal renderer for buttery smooth UI
1135- Enable/disable MCP servers by @mentioning, or in /mcp
1136- Added tab completion for shell commands in bash mode
1137- PreToolUse hooks can now modify tool inputs
1138- Press Ctrl-G to edit your prompt in your system's configured text editor
1139- Fixes for bash permission checks with environment variables in the command
1140
1141## 2.0.9
1142
1143- Fix regression where bash backgrounding stopped working
1144
1145## 2.0.8
1146
1147- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
1148- IDE: Add drag-and-drop support for files and folders in chat
1149- /context: Fix counting for thinking blocks
1150- Improve message rendering for users with light themes on dark terminals
1151- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
1152
1153## 2.0.5
1154
1155- IDE: Fix IME unintended message submission with Enter and Tab
1156- IDE: Add "Open in Terminal" link in login screen
1157- Fix unhandled OAuth expiration 401 API errors
1158- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
1159
1160## 2.0.1
1161
1162- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
1163- Various bug fixes and presentation improvements
1164
1165## 2.0.0
1166
1167- New native VS Code extension
1168- Fresh coat of paint throughout the whole app
1169- /rewind a conversation to undo code changes
1170- /usage command to see plan limits
1171- Tab to toggle thinking (sticky across sessions)
1172- Ctrl-R to search history
1173- Unshipped claude config command
1174- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1175- SDK: The Claude Code SDK is now the Claude Agent SDK
1176- Add subagents dynamically with `--agents` flag
1177
1178## 1.0.126
1179
1180- Enable /context command for Bedrock and Vertex
1181- Add mTLS support for HTTP-based OpenTelemetry exporters
1182
1183## 1.0.124
1184
1185- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1186- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1187- No longer inform Claude of the list of allowed tools when permission is denied
1188- Fixed security vulnerability in Bash tool permission checks
1189- Improved VSCode extension performance for large files
1190
1191## 1.0.123
1192
1193- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1194- Fixed thinking mode triggering on negation phrases like "don't think"
1195- Fixed rendering performance degradation during token streaming
1196- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1197- Enhanced BashTool environment snapshot logging
1198- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1199- Migrated --debug logging to a file, to enable easy tailing & filtering
1200
1201## 1.0.120
1202
1203- Fix input lag during typing, especially noticeable with large prompts
1204- Improved VSCode extension command registry and sessions dialog user experience
1205- Enhanced sessions dialog responsiveness and visual feedback
1206- Fixed IDE compatibility issue by removing worktree support check
1207- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1208
1209## 1.0.119
1210
1211- Fix Windows issue where process visually freezes on entering interactive mode
1212- Support dynamic headers for MCP servers via headersHelper configuration
1213- Fix thinking mode not working in headless sessions
1214- Fix slash commands now properly update allowed tools instead of replacing them
1215
1216## 1.0.117
1217
1218- Add Ctrl-R history search to recall previous commands like bash/zsh
1219- Fix input lag while typing, especially on Windows
1220- Add sed command to auto-allowed commands in acceptEdits mode
1221- Fix Windows PATH comparison to be case-insensitive for drive letters
1222- Add permissions management hint to /add-dir output
1223
1224## 1.0.115
1225
1226- Improve thinking mode display with enhanced visual effects
1227- Type /t to temporarily disable thinking mode in your prompt
1228- Improve path validation for glob and grep tools
1229- Show condensed output for post-tool hooks to reduce visual clutter
1230- Fix visual feedback when loading state completes
1231- Improve UI consistency for permission request dialogs
1232
1233## 1.0.113
1234
1235- Deprecated piped input in interactive mode
1236- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1237
1238## 1.0.112
1239
1240- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1241- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1242- Hooks: Added systemMessage support for SessionEnd hooks
1243- Added `spinnerTipsEnabled` setting to disable spinner tips
1244- IDE: Various improvements and bug fixes
1245
1246## 1.0.111
1247
1248- /model now validates provided model names
1249- Fixed Bash tool crashes caused by malformed shell syntax parsing
1250
1251## 1.0.110
1252
1253- /terminal-setup command now supports WezTerm
1254- MCP: OAuth tokens now proactively refresh before expiration
1255- Fixed reliability issues with background Bash processes
1256
1257## 1.0.109
1258
1259- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1260
1261## 1.0.106
1262
1263- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1264
1265## 1.0.97
1266
1267- Settings: /doctor now validates permission rule syntax and suggests corrections
1268
1269## 1.0.94
1270
1271- Vertex: add support for global endpoints for supported models
1272- /memory command now allows direct editing of all imported memory files
1273- SDK: Add custom tools as callbacks
1274- Added /todos command to list current todo items
1275
1276## 1.0.93
1277
1278- Windows: Add alt + v shortcut for pasting images from clipboard
1279- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1280
1281## 1.0.90
1282
1283- Settings file changes take effect immediately - no restart required
1284
1285## 1.0.88
1286
1287- Fixed issue causing "OAuth authentication is currently not supported"
1288- Status line input now includes `exceeds_200k_tokens`
1289- Fixed incorrect usage tracking in /cost.
1290- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1291- Bedrock: Updated default Sonnet model to Sonnet 4
1292
1293## 1.0.86
1294
1295- Added /context to help users self-serve debug context issues
1296- SDK: Added UUID support for all SDK messages
1297- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1298
1299## 1.0.85
1300
1301- Status line input now includes session cost info
1302- Hooks: Introduced SessionEnd hook
1303
1304## 1.0.84
1305
1306- Fix tool_use/tool_result id mismatch error when network is unstable
1307- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1308- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1309- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1310
1311## 1.0.83
1312
1313- @-mention: Support files with spaces in path
1314- New shimmering spinner
1315
1316## 1.0.82
1317
1318- SDK: Add request cancellation support
1319- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1320- Settings: Validation prevents invalid fields in .claude/settings.json files
1321- MCP: Improve tool name consistency
1322- Bash: Fix crash when Claude tries to automatically read large files
1323
1324## 1.0.81
1325
1326- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1327- Agents: Fix custom agent loading when agent files are unparsable
1328
1329## 1.0.80
1330
1331- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1332
1333## 1.0.77
1334
1335- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1336- SDK: Add session support and permission denial tracking
1337- Fix token limit errors in conversation summarization
1338- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1339
1340## 1.0.73
1341
1342- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1343- MCP: Press Esc to cancel OAuth authentication flows
1344- Bash: Improved command validation and reduced false security warnings
1345- UI: Enhanced spinner animations and status line visual hierarchy
1346- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1347
1348## 1.0.72
1349
1350- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1351
1352## 1.0.71
1353
1354- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1355- Customizable status line: add your terminal prompt to Claude Code with /statusline
1356
1357## 1.0.70
1358
1359- Performance: Optimized message rendering for better performance with large contexts
1360- Windows: Fixed native file search, ripgrep, and subagent functionality
1361- Added support for @-mentions in slash command arguments
1362
1363## 1.0.69
1364
1365- Upgraded Opus to version 4.1
1366
1367## 1.0.68
1368
1369- Fix incorrect model names being used for certain commands like `/pr-comments`
1370- 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.
1371- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1372- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1373- SDK: Added canUseTool callback support for tool confirmation
1374- Added `disableAllHooks` setting
1375- Improved file suggestions performance in large repos
1376
1377## 1.0.65
1378
1379- IDE: Fixed connection stability issues and error handling for diagnostics
1380- Windows: Fixed shell environment setup for users without .bashrc files
1381
1382## 1.0.64
1383
1384- Agents: Added model customization support - you can now specify which model an agent should use
1385- Agents: Fixed unintended access to the recursive agent tool
1386- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1387- SDK: Fixed user input tracking across multi-turn conversations
1388- Added hidden files to file search and @-mention suggestions
1389
1390## 1.0.63
1391
1392- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1393
1394## 1.0.62
1395
1396- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1397- Hooks: Added SessionStart hook for new session initialization
1398- /add-dir command now supports typeahead for directory paths
1399- Improved network connectivity check reliability
1400
1401## 1.0.61
1402
1403- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1404- Settings: Added `--settings` flag to load settings from a JSON file
1405- Settings: Fixed resolution of settings files paths that are symlinks
1406- OTEL: Fixed reporting of wrong organization after authentication changes
1407- Slash commands: Fixed permissions checking for allowed-tools with Bash
1408- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1409- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1410- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1411
1412## 1.0.60
1413
1414- You can now create custom subagents for specialized tasks! Run /agents to get started
1415
1416## 1.0.59
1417
1418- SDK: Added tool confirmation support with canUseTool callback
1419- SDK: Allow specifying env for spawned process
1420- Hooks: Exposed PermissionDecision to hooks (including "ask")
1421- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1422- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1423
1424## 1.0.58
1425
1426- Added support for reading PDFs
1427- MCP: Improved server health status display in 'claude mcp list'
1428- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1429
1430## 1.0.57
1431
1432- Added support for specifying a model in slash commands
1433- Improved permission messages to help Claude understand allowed tools
1434- Fix: Remove trailing newlines from bash output in terminal wrapping
1435
1436## 1.0.56
1437
1438- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1439- Fixes for WSL IDE detection
1440- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1441
1442## 1.0.55
1443
1444- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1445- Windows: fixed Ctrl+Z crash
1446- SDK: Added ability to capture error logging
1447- Add --system-prompt-file option to override system prompt in print mode
1448
1449## 1.0.54
1450
1451- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1452- Custom slash commands: Added argument-hint to frontmatter
1453- Windows: OAuth uses port 45454 and properly constructs browser URL
1454- Windows: mode switching now uses alt + m, and plan mode renders properly
1455- Shell: Switch to in-memory shell snapshot to fix file-related errors
1456
1457## 1.0.53
1458
1459- Updated @-mention file truncation from 100 lines to 2000 lines
1460- 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).
1461
1462## 1.0.52
1463
1464- Added support for MCP server instructions
1465
1466## 1.0.51
1467
1468- Added support for native Windows (requires Git for Windows)
1469- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1470- Settings: /doctor can now help you identify and fix invalid setting files
1471- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1472- Increased auto-compact warning threshold from 60% to 80%
1473- Fixed an issue with handling user directories with spaces for shell snapshots
1474- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1475- Custom slash commands: Fixed user-level commands in subdirectories
1476- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1477
1478## 1.0.48
1479
1480- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1481- Added progress messages to Bash tool based on the last 5 lines of command output
1482- Added expanding variables support for MCP server configuration
1483- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1484- Improved IDE extension path handling when Claude Code runs in WSL
1485- Hooks: Added a PreCompact hook
1486- Vim mode: Added c, f/F, t/T
1487
1488## 1.0.45
1489
1490- Redesigned Search (Grep) tool with new tool input parameters and features
1491- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1492- Fixed config file corruption issue by enforcing atomic writes
1493- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1494- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1495- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1496
1497## 1.0.44
1498
1499- New /export command lets you quickly export a conversation for sharing
1500- MCP: resource_link tool results are now supported
1501- MCP: tool annotations and tool titles now display in /mcp view
1502- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1503
1504## 1.0.43
1505
1506- Fixed a bug where the theme selector was saving excessively
1507- Hooks: Added EPIPE system error handling
1508
1509## 1.0.42
1510
1511- Added tilde (`~`) expansion support to `/add-dir` command
1512
1513## 1.0.41
1514
1515- Hooks: Split Stop hook triggering into Stop and SubagentStop
1516- Hooks: Enabled optional timeout configuration for each command
1517- Hooks: Added "hook_event_name" to hook input
1518- Fixed a bug where MCP tools would display twice in tool list
1519- New tool parameters JSON for Bash tool in `tool_decision` event
1520
1521## 1.0.40
1522
1523- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1524
1525## 1.0.39
1526
1527- New Active Time metric in OpenTelemetry logging
1528
1529## 1.0.38
1530
1531- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1532
1533## 1.0.37
1534
1535- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1536
1537## 1.0.36
1538
1539- Web search now takes today's date into context
1540- Fixed a bug where stdio MCP servers were not terminating properly on exit
1541
1542## 1.0.35
1543
1544- Added support for MCP OAuth Authorization Server discovery
1545
1546## 1.0.34
1547
1548- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1549
1550## 1.0.33
1551
1552- Improved logging functionality with session ID support
1553- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1554- Improvements to plan mode
1555
1556## 1.0.32
1557
1558- Updated loopback config for litellm
1559- Added forceLoginMethod setting to bypass login selection screen
1560
1561## 1.0.31
1562
1563- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1564
1565## 1.0.30
1566
1567- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1568- Improved file path autocomplete with filename matching
1569- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1570- Enhanced jq regex support for complex filters with pipes and select
1571
1572## 1.0.29
1573
1574- Improved CJK character support in cursor navigation and rendering
1575
1576## 1.0.28
1577
1578- Slash commands: Fix selector display during history navigation
1579- Resizes images before upload to prevent API size limit errors
1580- Added XDG_CONFIG_HOME support to configuration directory
1581- Performance optimizations for memory usage
1582- New attributes (terminal.type, language) in OpenTelemetry logging
1583
1584## 1.0.27
1585
1586- Streamable HTTP MCP servers are now supported
1587- Remote MCP servers (SSE and HTTP) now support OAuth
1588- MCP resources can now be @-mentioned
1589- /resume slash command to switch conversations within Claude Code
1590
1591## 1.0.25
1592
1593- Slash commands: moved "project" and "user" prefixes to descriptions
1594- Slash commands: improved reliability for command discovery
1595- Improved support for Ghostty
1596- Improved web search reliability
1597
1598## 1.0.24
1599
1600- Improved /mcp output
1601- Fixed a bug where settings arrays got overwritten instead of merged
1602
1603## 1.0.23
1604
1605- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1606- Released Python SDK: pip install claude-code-sdk to get started
1607
1608## 1.0.22
1609
1610- SDK: Renamed `total_cost` to `total_cost_usd`
1611
1612## 1.0.21
1613
1614- Improved editing of files with tab-based indentation
1615- Fix for tool_use without matching tool_result errors
1616- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1617
1618## 1.0.18
1619
1620- Added --add-dir CLI argument for specifying additional working directories
1621- Added streaming input support without require -p flag
1622- Improved startup performance and session storage performance
1623- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1624- Added detailed MCP server tools display (/mcp)
1625- MCP authentication and permission improvements
1626- Added auto-reconnection for MCP SSE connections on disconnect
1627- Fixed issue where pasted content was lost when dialogs appeared
1628
1629## 1.0.17
1630
1631- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1632- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1633- MCP server list UI improvements
1634- Update Claude Code process title to display "claude" instead of "node"
1635
1636## 1.0.11
1637
1638- Claude Code can now also be used with a Claude Pro subscription
1639- Added /upgrade for smoother switching to Claude Max plans
1640- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1641- Improved shell configuration error handling
1642- Improved todo list handling during compaction
1643
1644## 1.0.10
1645
1646- Added markdown table support
1647- Improved streaming performance
1648
1649## 1.0.8
1650
1651- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1652- Increased default otel interval from 1s -> 5s
1653- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1654- Fixed a regression where search tools unnecessarily asked for permissions
1655- Added support for triggering thinking non-English languages
1656- Improved compacting UI
1657
1658## 1.0.7
1659
1660- Renamed /allowed-tools -> /permissions
1661- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1662- Deprecated claude config commands in favor of editing settings.json
1663- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1664- Improved error handling for /install-github-app
1665- Bugfixes, UI polish, and tool reliability improvements
1666
1667## 1.0.6
1668
1669- Improved edit reliability for tab-indented files
1670- Respect CLAUDE_CONFIG_DIR everywhere
1671- Reduced unnecessary tool permission prompts
1672- Added support for symlinks in @file typeahead
1673- Bugfixes, UI polish, and tool reliability improvements
1674
1675## 1.0.4
1676
1677- Fixed a bug where MCP tool errors weren't being parsed correctly
1678
1679## 1.0.1
1680
1681- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1682- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1683- Updated documentation links and OAuth process descriptions
1684
1685## 1.0.0
1686
1687- Claude Code is now generally available
1688- Introducing Sonnet 4 and Opus 4 models
1689
1690## 0.2.125
1691
1692- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1693- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1694
1695## 0.2.117
1696
1697- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1698- Introduced settings.cleanupPeriodDays
1699- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1700- Introduced --debug mode
1701
1702## 0.2.108
1703
1704- You can now send messages to Claude while it works to steer Claude in real-time
1705- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1706- Fixed a bug where thinking was not working in -p mode
1707- Fixed a regression in /cost reporting
1708- Deprecated MCP wizard interface in favor of other MCP commands
1709- Lots of other bugfixes and improvements
1710
1711## 0.2.107
1712
1713- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1714
1715## 0.2.106
1716
1717- MCP SSE server configs can now specify custom headers
1718- Fixed a bug where MCP permission prompt didn't always show correctly
1719
1720## 0.2.105
1721
1722- Claude can now search the web
1723- Moved system & account status to /status
1724- Added word movement keybindings for Vim
1725- Improved latency for startup, todo tool, and file edits
1726
1727## 0.2.102
1728
1729- Improved thinking triggering reliability
1730- Improved @mention reliability for images and folders
1731- You can now paste multiple large chunks into one prompt
1732
1733## 0.2.100
1734
1735- Fixed a crash caused by a stack overflow error
1736- Made db storage optional; missing db support disables --continue and --resume
1737
1738## 0.2.98
1739
1740- Fixed an issue where auto-compact was running twice
1741
1742## 0.2.96
1743
1744- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1745
1746## 0.2.93
1747
1748- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1749- Claude now has access to a Todo list that helps it stay on track and be more organized
1750
1751## 0.2.82
1752
1753- Added support for --disallowedTools
1754- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1755
1756## 0.2.75
1757
1758- Hit Enter to queue up additional messages while Claude is working
1759- Drag in or copy/paste image files directly into the prompt
1760- @-mention files to directly add them to context
1761- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1762- Improved performance for filename auto-complete
1763
1764## 0.2.74
1765
1766- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1767- Task tool can now perform writes and run bash commands
1768
1769## 0.2.72
1770
1771- Updated spinner to indicate tokens loaded and tool usage
1772
1773## 0.2.70
1774
1775- Network commands like curl are now available for Claude to use
1776- Claude can now run multiple web queries in parallel
1777- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1778
1779## 0.2.69
1780
1781- Fixed UI glitches with improved Select component behavior
1782- Enhanced terminal output display with better text truncation logic
1783
1784## 0.2.67
1785
1786- Shared project permission rules can be saved in .claude/settings.json
1787
1788## 0.2.66
1789
1790- Print mode (-p) now supports streaming output via --output-format=stream-json
1791- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1792
1793## 0.2.63
1794
1795- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1796
1797## 0.2.61
1798
1799- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1800- Enhanced image detection for more reliable clipboard paste functionality
1801- Fixed an issue where ESC key could crash the conversation history selector
1802
1803## 0.2.59
1804
1805- Copy+paste images directly into your prompt
1806- Improved progress indicators for bash and fetch tools
1807- Bugfixes for non-interactive mode (-p)
1808
1809## 0.2.54
1810
1811- Quickly add to Memory by starting your message with '#'
1812- Press ctrl+r to see full output for long tool results
1813- Added support for MCP SSE transport
1814
1815## 0.2.53
1816
1817- New web fetch tool lets Claude view URLs that you paste in
1818- Fixed a bug with JPEG detection
1819
1820## 0.2.50
1821
1822- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1823
1824## 0.2.49
1825
1826- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1827
1828## 0.2.47
1829
1830- Press Tab to auto-complete file and folder names
1831- Press Shift + Tab to toggle auto-accept for file edits
1832- Automatic conversation compaction for infinite conversation length (toggle with /config)
1833
1834## 0.2.44
1835
1836- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1837
1838## 0.2.41
1839
1840- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1841- MCP server startup no longer blocks the app from starting up
1842
1843## 0.2.37
1844
1845- New /release-notes command lets you view release notes at any time
1846- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1847
1848## 0.2.36
1849
1850- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1851- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1852
1853## 0.2.34
1854
1855- Vim bindings for text input - enable with /vim or /config
1856
1857## 0.2.32
1858
1859- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1860- Fix for some PersistentShell issues
1861
1862## 0.2.31
1863
1864- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1865- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1866
1867## 0.2.30
1868
1869- Added ANSI color theme for better terminal compatibility
1870- Fixed issue where slash command arguments weren't being sent properly
1871- (Mac-only) API keys are now stored in macOS Keychain
1872
1873## 0.2.26
1874
1875- New /approved-tools command for managing tool permissions
1876- Word-level diff display for improved code readability
1877- Fuzzy matching for slash commands
1878
1879## 0.2.21
1880
1881- Fuzzy matching for /commands