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 9582ad4 · on Mar 5, 2026 History
2,069 lines · 125.8 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.69
4
5- Added the `/claude-api` skill for building applications with the Claude API and Anthropic SDK
6- Added Ctrl+U on an empty bash prompt (`!`) to exit bash mode, matching `escape` and `backspace`
7- Added numeric keypad support for selecting options in Claude's interview questions (previously only the number row above QWERTY worked)
8- Added optional name argument to `/remote-control` and `claude remote-control` (`/remote-control My Project` or `--name "My Project"`) to set a custom session title visible in claude.ai/code
9- Added Voice STT support for 10 new languages (20 total) — Russian, Polish, Turkish, Dutch, Ukrainian, Greek, Czech, Danish, Swedish, Norwegian
10- Added effort level display (e.g., "with low effort") to the logo and spinner, making it easier to see which effort setting is active
11- Added agent name display in terminal title when using `claude --agent`
12- Added `sandbox.enableWeakerNetworkIsolation` setting (macOS only) to allow Go programs like `gh`, `gcloud`, and `terraform` to verify TLS certificates when using a custom MITM proxy with `httpProxyPort`
13- Added `includeGitInstructions` setting (and `CLAUDE_CODE_DISABLE_GIT_INSTRUCTIONS` env var) to remove built-in commit and PR workflow instructions from Claude's system prompt
14- Added `/reload-plugins` command to activate pending plugin changes without restarting
15- Added a one-time startup prompt suggesting Claude Code Desktop on macOS and Windows (max 3 showings, dismissible)
16- Added `${CLAUDE_SKILL_DIR}` variable for skills to reference their own directory in SKILL.md content
17- Added `InstructionsLoaded` hook event that fires when CLAUDE.md or `.claude/rules/*.md` files are loaded into context
18- Added `agent_id` (for subagents) and `agent_type` (for subagents and `--agent`) to hook events
19- Added `worktree` field to status line hook commands with name, path, branch, and original repo directory when running in a `--worktree` session
20- Added `pluginTrustMessage` in managed settings to append organization-specific context to the plugin trust warning shown before installation
21- Added policy limit fetching (e.g., remote control restrictions) for Team plan OAuth users, not just Enterprise
22- Added `pathPattern` to `strictKnownMarketplaces` for regex-matching file/directory marketplace sources alongside `hostPattern` restrictions
23- Added plugin source type `git-subdir` to point to a subdirectory within a git repo
24- Added `oauth.authServerMetadataUrl` config option for MCP servers to specify a custom OAuth metadata discovery URL when standard discovery fails
25- Fixed a security issue where nested skill discovery could load skills from gitignored directories like `node_modules`
26- Fixed trust dialog silently enabling all `.mcp.json` servers on first run. You'll now see the per-server approval dialog as expected
27- Fixed `claude remote-control` crashing immediately on npm installs with "bad option: --sdk-url" (anthropics/claude-code#28334)
28- Fixed `--model claude-opus-4-0` and `--model claude-opus-4-1` resolving to deprecated Opus versions instead of current
29- Fixed macOS keychain corruption when using multiple OAuth MCP servers. Large OAuth metadata blobs could overflow the `security -i` stdin buffer, silently leaving stale credentials behind and causing repeated `/login` prompts.
30- Fixed `.credentials.json` losing `subscriptionType` (showing "Claude API" instead of "Claude Pro"/"Claude Max") when the profile endpoint transiently fails during token refresh (anthropics/claude-code#30185)
31- Fixed ghost dotfiles (`.bashrc`, `HEAD`, etc.) appearing as untracked files in the working directory after sandboxed Bash commands on Linux
32- Fixed Shift+Enter printing `[27;2;13~` instead of inserting a newline in Ghostty over SSH
33- Fixed stash (Ctrl+S) being cleared when submitting a message while Claude is working
34- Fixed ctrl+o (transcript toggle) freezing for many seconds in long sessions with lots of file edits
35- Fixed plan mode feedback input not supporting multi-line text entry (backslash+Enter and Shift+Enter now insert newlines)
36- Fixed cursor not moving down into blank lines at the top of the input box
37- Fixed `/stats` crash when transcript files contain entries with missing or malformed timestamps
38- Fixed a brief hang after a streaming error on long sessions (the transcript was being fully rewritten to drop one line; it is now truncated in place)
39- Fixed `--setting-sources user` not blocking dynamically discovered project skills
40- Fixed duplicate CLAUDE.md, slash commands, agents, and rules when running from a worktree nested inside its main repo (e.g. `claude -w`)
41- Fixed plugin Stop/SessionEnd/etc hooks not firing after any `/plugin` operation
42- Fixed plugin hooks being silently dropped when two plugins use the same `${CLAUDE_PLUGIN_ROOT}/...` command template
43- Fixed memory leak in long-running SDK/CCR sessions where conversation messages were retained unnecessarily
44- Fixed API 400 errors in forked agents (autocompact, summarization) when resuming sessions that were interrupted mid-tool-batch
45- Fixed "unexpected tool_use_id found in tool_result blocks" error when resuming conversations that start with an orphaned tool result
46- Fixed teammates accidentally spawning nested teammates via the Agent tool's `name` parameter
47- Fixed `CLAUDE_CODE_MAX_OUTPUT_TOKENS` being ignored during conversation compaction
48- Fixed `/compact` summary rendering as a user bubble in SDK consumers (Claude Code Remote web UI, VSCode extension)
49- Fixed voice space bar getting stuck after a failed voice activation (module loading race, cold GrowthBook)
50- Fixed worktree file copy on Windows
51- Fixed global `.claude` folder detection on Windows
52- Fixed symlink bypass where writing new files through a symlinked parent directory could escape the working directory in `acceptEdits` mode
53- Fixed sandbox prompting users to approve non-allowed domains when `allowManagedDomainsOnly` is enabled in managed settings — non-allowed domains are now blocked automatically with no bypass
54- Fixed interactive tools (e.g., `AskUserQuestion`) being silently auto-allowed when listed in a skill's allowed-tools, bypassing the permission prompt and running with empty answers
55- Fixed multi-GB memory spike when committing with large untracked binary files in the working tree
56- Fixed Escape not interrupting a running turn when the input box has draft text. Use Up arrow to pull queued messages back for editing, or Ctrl+U to clear the input line.
57- Fixed Android app crash when running local slash commands (`/voice`, `/cost`) in Remote Control sessions
58- Fixed a memory leak where old message array versions accumulated in React Compiler `memoCache` over long sessions
59- Fixed a memory leak where REPL render scopes accumulated over long sessions (~35MB over 1000 turns)
60- Fixed memory retention in in-process teammates where the parent's full conversation history was pinned for the teammate's lifetime, preventing GC after `/clear` or auto-compact
61- Fixed a memory leak in interactive mode where hook events could accumulate unboundedly during long sessions
62- Fixed hang when `--mcp-config` points to a corrupted file
63- Fixed slow startup when many skills/plugins are installed
64- Fixed `cd <outside-dir> && <cmd>` permission prompt to surface the chained command instead of only showing "Yes, allow reading from <dir>/"
65- Fixed conditional `.claude/rules/*.md` files (with `paths:` frontmatter) and nested CLAUDE.md files not loading in print mode (`claude -p`)
66- Fixed `/clear` not fully clearing all session caches, reducing memory retention in long sessions
67- Fixed terminal flicker caused by animated elements at the scrollback boundary
68- Fixed UI frame drops on macOS when using MCP servers with OAuth (regression from 2.1.x)
69- Fixed occasional frame stalls during typing caused by synchronous debug log flushes
70- Fixed `TeammateIdle` and `TaskCompleted` hooks to support `{"continue": false, "stopReason": "..."}` to stop the teammate, matching `Stop` hook behavior
71- Fixed `WorktreeCreate` and `WorktreeRemove` plugin hooks being silently ignored
72- Fixed skill descriptions with colons (e.g., "Triggers include: X, Y, Z") failing to load from SKILL.md frontmatter
73- Fixed project skills without a `description:` frontmatter field not appearing in Claude's available skills list
74- Fixed `/context` showing identical token counts for all MCP tools from a server
75- Fixed literal `nul` file creation on Windows when the model uses CMD-style `2>nul` redirection in Git Bash
76- Fixed extra blank lines appearing below each tool call in the expanded subagent transcript view (Ctrl+O)
77- Fixed Tab/arrow keys not cycling Settings tabs when `/config` search box is focused but empty
78- Fixed service key OAuth sessions (CCR containers) spamming `[ERROR]` logs with 403s from profile-scoped endpoints
79- Fixed inconsistent color for "Remote Control active" status indicator
80- Fixed Voice waveform cursor covering the first suffix letter when dictating mid-input
81- Fixed Voice input showing all 5 spaces during warmup instead of capping at ~2 (aligning with the "keep holding…" hint)
82- Improved spinner performance by isolating the 50ms animation loop from the surrounding shell, reducing render and CPU overhead during turns
83- Improved UI rendering performance in native binaries with React Compiler
84- Improved `--worktree` startup by eliminating a git subprocess on the startup path
85- Improved macOS startup by eliminating redundant settings-file reloads when managed settings resolve
86- Improved macOS startup for Claude.ai enterprise/team users by skipping an unnecessary keychain lookup
87- Improved MCP `-p` startup by pipelining claude.ai config fetch with local connections and using a concurrency pool instead of sequential batching
88- Improved voice startup by removing imperceptible warmup pulse animations that were causing re-render stutter
89- Improved MCP binary content handling: tools returning PDFs, Office documents, or audio now save decoded bytes to disk with the correct file extension instead of dumping raw base64 into the conversation context. WebFetch also saves binary responses alongside its summary.
90- Improved memory usage in long sessions by stabilizing `onSubmit` across message updates
91- Improved LSP tool rendering and memory context building to no longer read entire files
92- Improved session upload and memory sync to avoid reading large files into memory before size/binary checks
93- Improved file operation performance by avoiding reading file contents for existence checks (6 sites)
94- Improved documentation to clarify that `--append-system-prompt-file` and `--system-prompt-file` work in interactive mode (the docs previously said print mode only)
95- Reduced baseline memory by ~16MB by deferring Yoga WASM preloading
96- Reduced memory footprint for SDK and CCR sessions using stream-json output
97- Reduced memory usage when resuming large sessions (including compacted history)
98- Reduced token usage on multi-agent tasks with more concise subagent final reports
99- Changed Sonnet 4.5 users on Pro/Max/Team Premium to be automatically migrated to Sonnet 4.6
100- Changed the `/resume` picker to show your most recent prompt instead of the first one. This also resolves some titles appearing as `(session)`.
101- Changed claude.ai MCP connector failures to show a notification instead of silently disappearing from the tool list
102- Changed example command suggestions to be generated deterministically instead of calling Haiku
103- Changed resuming after compaction to no longer produce a preamble recap before continuing
104- [SDK] Changed task creation to no longer require the `activeForm` field — the spinner falls back to the task subject
105- [VSCode] Added compaction display as a collapsible "Compacted chat" card with the summary inside
106- [VSCode] The permission mode picker now respects `permissions.disableBypassPermissionsMode` from your effective Claude Code settings (including managed/policy settings) — when set to `disable`, bypass permissions mode is hidden from the picker
107- [VSCode] Fixed RTL text (Arabic, Hebrew, Persian) rendering reversed in the chat panel (regression in v2.1.63)
108
109## 2.1.68
110
111- Opus 4.6 now defaults to medium effort for Max and Team subscribers. Medium effort works well for most tasks — it's the sweet spot between speed and thoroughness. You can change this anytime with `/model`
112- Re-introduced the "ultrathink" keyword to enable high effort for the next turn
113- Removed Opus 4 and 4.1 from Claude Code on the first-party API — users with these models pinned are automatically moved to Opus 4.6
114
115## 2.1.66
116
117- Reduced spurious error logging
118
119## 2.1.63
120
121- Added `/simplify` and `/batch` bundled slash commands
122- Fixed local slash command output like /cost appearing as user-sent messages instead of system messages in the UI
123- Project configs & auto memory now shared across git worktrees of the same repository
124- Added `ENABLE_CLAUDEAI_MCP_SERVERS=false` env var to opt out from making claude.ai MCP servers available
125- Improved `/model` command to show the currently active model in the slash command menu
126- Added HTTP hooks, which can POST JSON to a URL and receive JSON instead of running a shell command
127- Fixed listener leak in bridge polling loop
128- Fixed listener leak in MCP OAuth flow cleanup
129- Added manual URL paste fallback during MCP OAuth authentication. If the automatic localhost redirect doesn't work, you can paste the callback URL to complete authentication.
130- Fixed memory leak when navigating hooks configuration menu
131- Fixed listener leak in interactive permission handler during auto-approvals
132- Fixed file count cache ignoring glob ignore patterns
133- Fixed memory leak in bash command prefix cache
134- Fixed MCP tool/resource cache leak on server reconnect
135- Fixed IDE host IP detection cache incorrectly sharing results across ports
136- Fixed WebSocket listener leak on transport reconnect
137- Fixed memory leak in git root detection cache that could cause unbounded growth in long-running sessions
138- Fixed memory leak in JSON parsing cache that grew unbounded over long sessions
139- VSCode: Fixed remote sessions not appearing in conversation history
140- Fixed a race condition in the REPL bridge where new messages could arrive at the server interleaved with historical messages during the initial connection flush, causing message ordering issues.
141- Fixed memory leak where long-running teammates retained all messages in AppState even after conversation compaction
142- Fixed a memory leak where MCP server fetch caches were not cleared on disconnect, causing growing memory usage with servers that reconnect frequently
143- Improved memory usage in long sessions with subagents by stripping heavy progress message payloads during context compaction
144- Added "Always copy full response" option to the `/copy` picker. When selected, future `/copy` commands will skip the code block picker and copy the full response directly.
145- VSCode: Added session rename and remove actions to the sessions list
146- Fixed `/clear` not resetting cached skills, which could cause stale skill content to persist in the new conversation
147
148## 2.1.62
149
150- Fixed prompt suggestion cache regression that reduced cache hit rates
151
152## 2.1.61
153
154- Fixed concurrent writes corrupting config file on Windows
155
156## 2.1.59
157
158- Claude automatically saves useful context to auto-memory. Manage with /memory
159- Added `/copy` command to show an interactive picker when code blocks are present, allowing selection of individual code blocks or the full response.
160- Improved "always allow" prefix suggestions for compound bash commands (e.g. `cd /tmp && git fetch && git push`) to compute smarter per-subcommand prefixes instead of treating the whole command as one
161- Improved ordering of short task lists
162- Improved memory usage in multi-agent sessions by releasing completed subagent task state
163- Fixed MCP OAuth token refresh race condition when running multiple Claude Code instances simultaneously
164- Fixed shell commands not showing a clear error message when the working directory has been deleted
165- Fixed config file corruption that could wipe authentication when multiple Claude Code instances ran simultaneously
166
167## 2.1.58
168
169- Expand Remote Control to more users
170
171## 2.1.56
172
173- VS Code: Fixed another cause of "command 'claude-vscode.editor.openLast' not found" crashes
174
175## 2.1.55
176
177- Fixed BashTool failing on Windows with EINVAL error
178
179## 2.1.53
180
181- Fixed a UI flicker where user input would briefly disappear after submission before the message rendered
182- Fixed bulk agent kill (ctrl+f) to send a single aggregate notification instead of one per agent, and to properly clear the command queue
183- Fixed graceful shutdown sometimes leaving stale sessions when using Remote Control by parallelizing teardown network calls
184- Fixed `--worktree` sometimes being ignored on first launch
185- Fixed a panic ("switch on corrupted value") on Windows
186- Fixed a crash that could occur when spawning many processes on Windows
187- Fixed a crash in the WebAssembly interpreter on Linux x64 & Windows x64
188- Fixed a crash that sometimes occurred after 2 minutes on Windows ARM64
189
190## 2.1.52
191
192- VS Code: Fixed extension crash on Windows ("command 'claude-vscode.editor.openLast' not found")
193
194## 2.1.51
195
196- Added `claude remote-control` subcommand for external builds, enabling local environment serving for all users.
197- Updated plugin marketplace default git timeout from 30s to 120s and added `CLAUDE_CODE_PLUGIN_GIT_TIMEOUT_MS` to configure.
198- Added support for custom npm registries and specific version pinning when installing plugins from npm sources
199- 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`.
200- Fixed a security issue where `statusLine` and `fileSuggestion` hook commands could execute without workspace trust acceptance in interactive mode.
201- Tool results larger than 50K characters are now persisted to disk (previously 100K). This reduces context window usage and improves conversation longevity.
202- 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.
203- 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.
204- Fixed slash command autocomplete crashing when a plugin's SKILL.md description is a YAML array or other non-string type
205- 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.
206- Managed settings can now be set via macOS plist or Windows Registry. Learn more at https://code.claude.com/docs/en/settings#settings-files
207
208## 2.1.50
209
210- Added support for `startupTimeout` configuration for LSP servers
211- Added `WorktreeCreate` and `WorktreeRemove` hook events, enabling custom VCS setup and teardown when agent worktree isolation creates or removes worktrees.
212- 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.
213- Linux: Fixed native modules not loading on systems with glibc older than 2.30 (e.g., RHEL 8)
214- Fixed memory leak in agent teams where completed teammate tasks were never garbage collected from session state
215- Fixed `CLAUDE_CODE_SIMPLE` to fully strip down skills, session memory, custom agents, and CLAUDE.md token counting
216- Fixed `/mcp reconnect` freezing the CLI when given a server name that doesn't exist
217- Fixed memory leak where completed task state objects were never removed from AppState
218- Added support for `isolation: worktree` in agent definitions, allowing agents to declaratively run in isolated git worktrees.
219- `CLAUDE_CODE_SIMPLE` mode now also disables MCP tools, attachments, hooks, and CLAUDE.md file loading for a fully minimal experience.
220- Fixed bug where MCP tools were not discovered when tool search is enabled and a prompt is passed in as a launch argument
221- Improved memory usage during long sessions by clearing internal caches after compaction
222- Added `claude agents` CLI command to list all configured agents
223- Improved memory usage during long sessions by clearing large tool results after they have been processed
224- Fixed a memory leak where LSP diagnostic data was never cleaned up after delivery, causing unbounded memory growth in long sessions
225- Fixed a memory leak where completed task output was not freed from memory, reducing memory usage in long sessions with many tasks
226- Improved startup performance for headless mode (`-p` flag) by deferring Yoga WASM and UI component imports
227- Fixed prompt suggestion cache regression that reduced cache hit rates
228- Fixed unbounded memory growth in long sessions by capping file history snapshots
229- Added `CLAUDE_CODE_DISABLE_1M_CONTEXT` environment variable to disable 1M context window support
230- Opus 4.6 (fast mode) now includes the full 1M context window
231- VSCode: Added `/extra-usage` command support in VS Code sessions
232- Fixed memory leak where TaskOutput retained recent lines after cleanup
233- Fixed memory leak in CircularBuffer where cleared items were retained in the backing array
234- Fixed memory leak in shell command execution where ChildProcess and AbortController references were retained after cleanup
235
236## 2.1.49
237
238- Improved MCP OAuth authentication with step-up auth support and discovery caching, reducing redundant network requests during server connections
239- Added `--worktree` (`-w`) flag to start Claude in an isolated git worktree
240- Subagents support `isolation: "worktree"` for working in a temporary git worktree
241- Added Ctrl+F keybinding to kill background agents (two-press confirmation)
242- Agent definitions support `background: true` to always run as a background task
243- Plugins can ship `settings.json` for default configuration
244- Fixed file-not-found errors to suggest corrected paths when the model drops the repo folder
245- 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.
246- Fixed prompt suggestion cache regression that reduced cache hit rates.
247- Fixed `plugin enable` and `plugin disable` to auto-detect the correct scope when `--scope` is not specified, instead of always defaulting to user scope
248- Simple mode (`CLAUDE_CODE_SIMPLE`) now includes the file edit tool in addition to the Bash tool, allowing direct file editing in simple mode.
249- Permission suggestions are now populated when safety checks trigger an ask response, enabling SDK consumers to display permission options
250- 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.
251- Fixed verbose mode not updating thinking block display when toggled via `/config` — memo comparators now correctly detect verbose changes
252- Fixed unbounded WASM memory growth during long sessions by periodically resetting the tree-sitter parser
253- Fixed potential rendering issues caused by stale yoga layout references
254- Improved performance in non-interactive mode (`-p`) by skipping unnecessary API calls during startup
255- Improved performance by caching authentication failures for HTTP and SSE MCP servers, avoiding repeated connection attempts to servers requiring auth
256- Fixed unbounded memory growth during long-running sessions caused by Yoga WASM linear memory never shrinking
257- SDK model info now includes `supportsEffort`, `supportedEffortLevels`, and `supportsAdaptiveThinking` fields so consumers can discover model capabilities.
258- Added `ConfigChange` hook event that fires when configuration files change during a session, enabling enterprise security auditing and optional blocking of settings changes.
259- Improved startup performance by caching MCP auth failures to avoid redundant connection attempts
260- Improved startup performance by reducing HTTP calls for analytics token counting
261- Improved startup performance by batching MCP tool token counting into a single API call
262- Fixed `disableAllHooks` setting to respect managed settings hierarchy — non-managed settings can no longer disable managed hooks set by policy (#26637)
263- 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.
264- 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
265
266## 2.1.47
267
268- Fixed FileWriteTool line counting to preserve intentional trailing blank lines instead of stripping them with `trimEnd()`.
269- 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.
270- 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.
271- 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.
272- 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)
273- Improved memory usage in long-running sessions by releasing API stream buffers, agent context, and skill state after use
274- Improved startup performance by deferring SessionStart hook execution, reducing time-to-interactive by ~500ms.
275- Fixed an issue where bash tool output was silently discarded on Windows when using MSYS2 or Cygwin shells.
276- 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.
277- Improved memory usage by trimming agent task message history after tasks complete
278- Improved memory usage during long agent sessions by eliminating O(n²) message accumulation in progress updates
279- Fixed the bash permission classifier to validate that returned match descriptions correspond to actual input rules, preventing hallucinated descriptions from incorrectly granting permissions
280- Fixed user-defined agents only loading one file on NFS/FUSE filesystems that report zero inodes (anthropics/claude-code#26044)
281- Fixed plugin agent skills silently failing to load when referenced by bare name instead of fully-qualified plugin name (anthropics/claude-code#25834)
282- Search patterns in collapsed tool results are now displayed in quotes for clarity
283- Windows: Fixed CWD tracking temp files never being cleaned up, causing them to accumulate indefinitely (anthropics/claude-code#17600)
284- 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.
285- 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.
286- Simplified teammate navigation to use only Shift+Down (with wrapping) instead of both Shift+Up and Shift+Down.
287- 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.
288- 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.
289- Fixed custom session titles set via `/rename` being lost after resuming a conversation (anthropics/claude-code#23610)
290- Fixed collapsed read/search hint text overflowing on narrow terminals by truncating from the start.
291- 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.
292- 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)
293- Fixed MCP servers not appearing in the MCP Management Dialog after deferred loading
294- Fixed session name persisting in status bar after `/clear` command (anthropics/claude-code#26082)
295- 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)
296- Fixed /resume silently dropping sessions when the first message exceeds 16KB or uses array-format content (anthropics/claude-code#25721)
297- Added `chat:newline` keybinding action for configurable multi-line input (anthropics/claude-code#26075)
298- Added `added_dirs` to the statusline JSON `workspace` section, exposing directories added via `/add-dir` to external scripts (anthropics/claude-code#26096)
299- Fixed `claude doctor` misclassifying mise and asdf-managed installations as native installs (anthropics/claude-code#26033)
300- Fixed zsh heredoc failing with "read-only file system" error in sandboxed commands (anthropics/claude-code#25990)
301- Fixed agent progress indicator showing inflated tool use count (anthropics/claude-code#26023)
302- Fixed image pasting not working on WSL2 systems where Windows copies images as BMP format (anthropics/claude-code#25935)
303- Fixed background agent results returning raw transcript data instead of the agent's final answer (anthropics/claude-code#26012)
304- Fixed Warp terminal incorrectly prompting for Shift+Enter setup when it supports it natively (anthropics/claude-code#25957)
305- Fixed CJK wide characters causing misaligned timestamps and layout elements in the TUI (anthropics/claude-code#26084)
306- Fixed custom agent `model` field in `.claude/agents/*.md` being ignored when spawning team teammates (anthropics/claude-code#26064)
307- Fixed plan mode being lost after context compaction, causing the model to switch from planning to implementation mode (anthropics/claude-code#26061)
308- Fixed `alwaysThinkingEnabled: true` in settings.json not enabling thinking mode on Bedrock and Vertex providers (anthropics/claude-code#26074)
309- Fixed `tool_decision` OTel telemetry event not being emitted in headless/SDK mode (anthropics/claude-code#26059)
310- Fixed session name being lost after context compaction — renamed sessions now preserve their custom title through compaction (anthropics/claude-code#26121)
311- Increased initial session count in resume picker from 10 to 50 for faster session discovery (anthropics/claude-code#26123)
312- Windows: fixed worktree session matching when drive letter casing differs (anthropics/claude-code#26123)
313- Fixed `/resume <session-id>` failing to find sessions whose first message exceeds 16KB (anthropics/claude-code#25920)
314- Fixed "Always allow" on multiline bash commands creating invalid permission patterns that corrupt settings (anthropics/claude-code#25909)
315- 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)
316- 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)
317- Fixed read-only git commands triggering FSEvents file watcher loops on macOS by adding --no-optional-locks flag (anthropics/claude-code#25750)
318- 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)
319- Fixed non-interactive subcommands like `claude doctor` and `claude plugin validate` being blocked inside nested Claude sessions (anthropics/claude-code#25803)
320- Windows: Fixed the same CLAUDE.md file being loaded twice when drive letter casing differs between paths (anthropics/claude-code#25756)
321- Fixed inline code spans in markdown being incorrectly parsed as bash commands (anthropics/claude-code#25792)
322- Fixed teammate spinners not respecting custom spinnerVerbs from settings (anthropics/claude-code#25748)
323- Fixed shell commands permanently failing after a command deletes its own working directory (anthropics/claude-code#26136)
324- Fixed hooks (PreToolUse, PostToolUse) silently failing to execute on Windows by using Git Bash instead of cmd.exe (anthropics/claude-code#25981)
325- Fixed LSP `findReferences` and other location-based operations returning results from gitignored files (e.g., `node_modules/`, `venv/`) (anthropics/claude-code#26051)
326- Moved config backup files from home directory root to `~/.claude/backups/` to reduce home directory clutter (anthropics/claude-code#26130)
327- Fixed sessions with large first prompts (>16KB) disappearing from the /resume list (anthropics/claude-code#26140)
328- Fixed shell functions with double-underscore prefixes (e.g., `__git_ps1`) not being preserved across shell sessions (anthropics/claude-code#25824)
329- Fixed spinner showing "0 tokens" counter before any tokens have been received (anthropics/claude-code#26105)
330- VSCode: Fixed conversation messages appearing dimmed while the AskUserQuestion dialog is open (anthropics/claude-code#26078)
331- 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)
332- Fixed Right Alt key leaving visible `[25~` escape sequence residue in the input field on Windows/Git Bash terminals (anthropics/claude-code#25943)
333- The `/rename` command now updates the terminal tab title by default (anthropics/claude-code#25789)
334- 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)
335- Fixed OSC 8 hyperlinks only being clickable on the first line when link text wraps across multiple terminal lines.
336
337## 2.1.46
338
339- Fixed orphaned CC processes after terminal disconnect on macOS
340- Added support for using claude.ai MCP connectors in Claude Code
341
342## 2.1.45
343
344- Added support for Claude Sonnet 4.6
345- Added support for reading `enabledPlugins` and `extraKnownMarketplaces` from `--add-dir` directories
346- 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
347- Added `SDKRateLimitInfo` and `SDKRateLimitEvent` types to the SDK, enabling consumers to receive rate limit status updates including utilization, reset times, and overage information
348- Fixed Agent Teams teammates failing on Bedrock, Vertex, and Foundry by propagating API provider environment variables to tmux-spawned processes (anthropics/claude-code#23561)
349- Fixed sandbox "operation not permitted" errors when writing temporary files on macOS by using the correct per-user temp directory (anthropics/claude-code#21654)
350- Fixed Task tool (backgrounded agents) crashing with a `ReferenceError` on completion (anthropics/claude-code#22087)
351- Fixed autocomplete suggestions not being accepted on Enter when images are pasted in the input
352- Fixed skills invoked by subagents incorrectly appearing in main session context after compaction
353- Fixed excessive `.claude.json.backup` files accumulating on every startup
354- Fixed plugin-provided commands, agents, and hooks not being available immediately after installation without requiring a restart
355- Improved startup performance by removing eager loading of session history for stats caching
356- Improved memory usage for shell commands that produce large output — RSS no longer grows unboundedly with command output size
357- Improved collapsed read/search groups to show the current file or search pattern being processed beneath the summary line while active
358- [VSCode] Improved permission destination choice (project/user/session) to persist across sessions
359
360## 2.1.44
361
362- Fixed ENAMETOOLONG errors for deeply-nested directory paths
363- Fixed auth refresh errors
364
365## 2.1.43
366
367- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
368- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
369- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
370
371## 2.1.42
372
373- Improved startup performance by deferring Zod schema construction
374- Improved prompt cache hit rates by moving date out of system prompt
375- Added one-time Opus 4.6 effort callout for eligible users
376- Fixed /resume showing interrupt messages as session titles
377- Fixed image dimension limit errors to suggest /compact
378
379## 2.1.41
380
381- Added guard against launching Claude Code inside another Claude Code session
382- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
383- Fixed a crash when MCP tools return image content during streaming
384- Fixed /resume session previews showing raw XML tags instead of readable command names
385- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
386- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
387- Fixed hook blocking errors (exit code 2) not showing stderr to the user
388- Added `speed` attribute to OTel events and trace spans for fast mode visibility
389- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
390- Added Windows ARM64 (win32-arm64) native binary support
391- Improved `/rename` to auto-generate session name from conversation context when called without arguments
392- Improved narrow terminal layout for prompt footer
393- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
394- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
395- Fixed background task notifications not being delivered in streaming Agent SDK mode
396- Fixed cursor jumping to end on each keystroke in classifier rule input
397- Fixed markdown link display text being dropped for raw URL
398- Fixed auto-compact failure error notifications being shown to users
399- Fixed permission wait time being included in subagent elapsed time display
400- Fixed proactive ticks firing while in plan mode
401- Fixed clear stale permission rules when settings change on disk
402- Fixed hook blocking errors showing stderr content in UI
403
404## 2.1.39
405
406- Improved terminal rendering performance
407- Fixed fatal errors being swallowed instead of displayed
408- Fixed process hanging after session close
409- Fixed character loss at terminal screen boundary
410- Fixed blank lines in verbose transcript view
411
412## 2.1.38
413
414- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
415- Fixed Tab key queueing slash commands instead of autocompleting
416- Fixed bash permission matching for commands using environment variable wrappers
417- Fixed text between tool uses disappearing when not using streaming
418- Fixed duplicate sessions when resuming in VS Code extension
419- Improved heredoc delimiter parsing to prevent command smuggling
420- Blocked writes to `.claude/skills` directory in sandbox mode
421
422## 2.1.37
423
424- Fixed an issue where /fast was not immediately available after enabling /extra-usage
425
426## 2.1.36
427
428- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
429
430## 2.1.34
431
432- Fixed a crash when agent teams setting changed between renders
433- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
434
435## 2.1.33
436
437- Fixed agent teammate sessions in tmux to send and receive messages
438- Fixed warnings about agent teams not being available on your current plan
439- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
440- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
441- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
442- Added plugin name to skill descriptions and `/skills` menu for better discoverability
443- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
444- 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
445- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
446- 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
447- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
448- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
449- Errors from invalid managed settings are now surfaced
450- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
451- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
452- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
453
454## 2.1.32
455
456- Claude Opus 4.6 is now available!
457- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
458- Claude now automatically records and recalls memories as it works
459- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
460- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
461- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
462- Updated --resume to re-use --agent value specified in previous conversation by default.
463- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
464- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
465- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
466- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
467- VSCode: Added spinner when loading past conversations list
468
469## 2.1.31
470
471- Added session resume hint on exit, showing how to continue your conversation later
472- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
473- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
474- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
475- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
476- 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
477- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
478- 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
479- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
480- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
481- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
482
483## 2.1.30
484
485- 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.
486- 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`.
487- Added `/debug` for Claude to help troubleshoot the current session
488- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
489- Added token count, tool uses, and duration metrics to Task tool results
490- Added reduced motion mode to the config
491- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
492- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
493- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
494- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
495- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
496- Fixed permission dialogs stealing focus while actively typing
497- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
498- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
499- 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
500- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
501- Changed `/model` to execute immediately instead of being queued
502- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
503- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
504
505## 2.1.29
506
507- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
508
509## 2.1.27
510
511- Added tool call failures and denials to debug logs
512- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
513- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
514- Sessions are now automatically linked to PRs when created via `gh pr create`
515- Fixed /context command not displaying colored output
516- Fixed status bar duplicating background task indicator when PR status was shown
517- Windows: Fixed bash command execution failing for users with `.bashrc` files
518- Windows: Fixed console windows flashing when spawning child processes
519- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
520
521## 2.1.25
522
523- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
524
525## 2.1.23
526
527- Added customizable spinner verbs setting (`spinnerVerbs`)
528- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
529- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
530- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
531- Fixed pending async hooks not being cancelled when headless streaming sessions ended
532- Fixed tab completion not updating the input field when accepting a suggestion
533- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
534- Improved terminal rendering performance with optimized screen data layout
535- Changed Bash commands to show timeout duration alongside elapsed time
536- Changed merged pull requests to show a purple status indicator in the prompt footer
537- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
538
539## 2.1.22
540
541- Fixed structured outputs for non-interactive (-p) mode
542
543## 2.1.21
544
545- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
546- Fixed shell completion cache files being truncated on exit
547- Fixed API errors when resuming sessions that were interrupted during tool execution
548- Fixed auto-compact triggering too early on models with large output token limits
549- Fixed task IDs potentially being reused after deletion
550- Fixed file search not working in VS Code extension on Windows
551- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
552- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
553- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
554- [VSCode] Fixed message action buttons having incorrect background colors
555
556## 2.1.20
557
558- Added arrow key history navigation in vim normal mode when cursor cannot move further
559- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
560- 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
561- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
562- Added ability to delete tasks via the `TaskUpdate` tool
563- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
564- Fixed agents sometimes ignoring user messages sent while actively working on a task
565- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
566- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
567- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
568- Fixed draft prompt being lost when pressing UP arrow to navigate command history
569- Fixed ghost text flickering when typing slash commands mid-input
570- Fixed marketplace source removal not properly deleting settings
571- Fixed duplicate output in some commands like `/context`
572- Fixed task list sometimes showing outside the main conversation view
573- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
574- Fixed crashes when cancelling tool use
575- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
576- Improved thinking status text with a subtle shimmer animation
577- Improved task list to dynamically adjust visible items based on terminal height
578- Improved fork conversation hint to show how to resume the original session
579- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
580- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
581- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
582- Changed the `/copy` command to be available to all users
583- Changed background agents to prompt for tool permissions before launching
584- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
585- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
586
587## 2.1.19
588
589- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
590- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
591- Fixed crashes on processors without AVX instruction support
592- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
593- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
594- Fixed resuming sessions by custom title not working when run from a different directory
595- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
596- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
597- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
598- Fixed file write preview omitting empty lines
599- Changed skills without additional permissions or hooks to be allowed without requiring approval
600- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
601- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
602- [VSCode] Enabled session forking and rewind functionality for all users
603
604## 2.1.18
605
606- 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
607
608## 2.1.17
609
610- Fixed crashes on processors without AVX instruction support
611
612## 2.1.16
613
614- Added new task management system, including new capabilities like dependency tracking
615- [VSCode] Added native plugin management support
616- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
617- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
618- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
619- Fixed session titles on the resume screen not respecting the user's language setting
620- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
621
622## 2.1.15
623
624- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
625- Improved UI rendering performance with React Compiler
626- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
627- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
628
629## 2.1.14
630
631- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
632- Added search to installed plugins list - type to filter by name or description
633- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
634- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
635- Fixed memory issues that could cause crashes when running parallel subagents
636- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
637- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
638- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
639- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
640- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
641- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
642- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
643- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
644- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
645- Improved backspace to delete pasted text as a single token instead of one character at a time
646- [VSCode] Added `/usage` command to display current plan usage
647
648## 2.1.12
649
650- Fixed message rendering bug
651
652## 2.1.11
653
654- Fixed excessive MCP connection requests for HTTP/SSE transports
655
656## 2.1.10
657
658- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
659- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
660- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
661- Improved startup to capture keystrokes typed before the REPL is fully ready
662- Improved file suggestions to show as removable attachments instead of inserting text when accepted
663- [VSCode] Added install count display to plugin listings
664- [VSCode] Added trust warning when installing plugins
665
666## 2.1.9
667
668- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
669- Added `plansDirectory` setting to customize where plan files are stored
670- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
671- Added session URL attribution to commits and PRs created from web sessions
672- Added support for `PreToolUse` hooks to return `additionalContext` to the model
673- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
674- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
675- Fixed MCP server reconnection hanging when cached connection promise never resolves
676- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
677
678## 2.1.7
679
680- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
681- Added ability to provide feedback when accepting permission prompts
682- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
683- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
684- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
685- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
686- 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)
687- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
688- Fixed terminal title animation jitter by using fixed-width braille characters
689- Fixed plugins with git submodules not being fully initialized when installed
690- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
691- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
692- 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.
693- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
694- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
695
696## 2.1.6
697
698- Added search functionality to `/config` command for quickly filtering settings
699- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
700- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
701- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
702- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
703- Added an error display when the editor fails during Ctrl+G
704- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
705- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
706- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
707- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
708- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
709- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
710- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
711- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
712- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
713- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
714- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
715- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
716- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
717- Improved @ autocomplete with icons for different suggestion types and single-line formatting
718- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
719- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
720- Changed terminal title to "Claude Code" on startup for better window identification
721- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
722- [VSCode] Fixed usage indicator not updating after manual compact
723
724## 2.1.5
725
726- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
727
728## 2.1.4
729
730- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
731- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
732
733## 2.1.3
734
735- Merged slash commands and skills, simplifying the mental model with no change in behavior
736- Added release channel (`stable` or `latest`) toggle to `/config`
737- 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
738- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
739- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
740- Fixed mismatch between background task count in status bar and items shown in tasks dialog
741- Fixed sub-agents using the wrong model during conversation compaction
742- Fixed web search in sub-agents using incorrect model
743- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
744- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
745- Improved slash command suggestion readability by truncating long descriptions to 2 lines
746- Changed tool hook execution timeout from 60 seconds to 10 minutes
747- [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)
748
749## 2.1.2
750
751- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
752- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
753- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
754- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
755- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
756- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
757- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
758- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
759- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
760- Fixed updates incorrectly claiming another installation is in progress
761- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
762- Fixed remote session URL and teleport being broken when using `/tasks` command
763- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
764- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
765- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
766- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
767- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
768- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
769- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
770- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
771- [SDK] Changed minimum zod peer dependency to ^4.0.0
772- [VSCode] Fixed usage display not updating after manual compact
773
774## 2.1.0
775
776- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
777- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
778- Added support for `agent` field in skills to specify agent type for execution
779- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
780- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
781- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
782- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
783- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
784- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
785- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
786- Fixed Esc key with queued prompts to only move them to input without canceling the running task
787- Reduced permission prompts for complex bash commands
788- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
789- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
790- Fixed plugin path resolution for file-based marketplace sources
791- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
792- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
793- Fixed Claude in Chrome support for WSL environments
794- Fixed Windows native installer silently failing when executable creation fails
795- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
796- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
797- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
798- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
799- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
800- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
801- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
802- Added hooks support for skill and slash command frontmatter
803- 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
804- Added `/plan` command shortcut to enable plan mode directly from the prompt
805- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
806- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
807- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
808- Added support for `once: true` config for hooks
809- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
810- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
811- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
812- Added left/right arrow key navigation for cycling through tabs in dialogs
813- Added real-time thinking block display in Ctrl+O transcript mode
814- Added filepath to full output in background bash task details dialog
815- Added Skills as a separate category in the context visualization
816- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
817- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
818- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
819- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
820- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
821- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
822- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
823- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
824- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
825- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
826- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
827- Fixed commands with `$()` command substitution failing with parse errors
828- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
829- 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)
830- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
831- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
832- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
833- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
834- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
835- Fixed subagents sometimes not inheriting the parent's model by default
836- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
837- Fixed duplicate Bash commands appearing in permission request option labels
838- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
839- Fixed background task completion notifications to appear proactively with bullet point
840- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
841- Fixed cursor disappearing after dismissing permission dialogs
842- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
843- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
844- Fixed images being silently dropped when queueing messages while backgrounding a task
845- Fixed large pasted images failing with "Image was too large" error
846- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
847- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
848- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
849- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
850- Fixed spinner token counter to properly accumulate tokens from subagents during execution
851- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
852- Fixed race condition where LSP tool could return "no server available" during startup
853- Fixed feedback submission hanging indefinitely when network requests timeout
854- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
855- Fixed hook success message showing trailing colon when hook has no output
856- Multiple optimizations to improve startup performance
857- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
858- Improved performance when reading Jupyter notebooks with many cells
859- Improved reliability for piped input like `cat refactor.md | claude`
860- Improved reliability for AskQuestion tool
861- Improved sed in-place edit commands to render as file edits with diff preview
862- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
863- Improved compaction reliability
864- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
865- Improved skills to show progress while executing, displaying tool uses as they happen
866- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
867- Improved skill suggestions to prioritize recently and frequently used skills
868- Improved spinner feedback when waiting for the first response token
869- Improved token count display in spinner to include tokens from background agents
870- Improved incremental output for async agents to give the main thread more control and visibility
871- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
872- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
873- Improved macOS screenshot paste reliability with TIFF format support
874- Improved `/stats` output
875- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
876- Changed "Interrupted" message color from red to grey for a less alarming appearance
877- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
878- Removed underline styling from image reference links
879- [SDK] Changed minimum zod peer dependency to ^4.0.0
880- [VSCode] Added currently selected model name to the context menu
881- [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")
882- [VSCode] Fixed paragraph breaks not rendering in markdown content
883- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
884- [Windows] Fixed issue with improper rendering
885
886## 2.0.76
887
888- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
889
890## 2.0.75
891
892- Minor bugfixes
893
894## 2.0.74
895
896- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
897- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
898- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
899- Added syntax highlighting info to theme picker
900- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
901- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
902- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
903- Fixed a potential crash when syntax highlighting isn't initialized correctly
904- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
905- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
906- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
907- [Windows] Fixed issue with improper rendering
908- [VSCode] Added gift tag pictogram for year-end promotion message
909
910## 2.0.73
911
912- Added clickable `[Image #N]` links that open attached images in the default viewer
913- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
914- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
915- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
916- Fixed slow input history cycling and race condition that could overwrite text after message submission
917- Improved `/theme` command to open theme picker directly
918- Improved theme picker UI
919- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
920- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
921
922## 2.0.72
923
924- 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
925- Reduced terminal flickering
926- Added scannable QR code to mobile app tip for quick app downloads
927- Added loading indicator when resuming conversations for better feedback
928- Fixed `/context` command not respecting custom system prompts in non-interactive mode
929- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
930- Improved @ mention file suggestion speed (~3× faster in git repositories)
931- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
932- Improved settings validation errors to be more prominent
933- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
934
935## 2.0.71
936
937- Added /config toggle to enable/disable prompt suggestions
938- Added `/settings` as an alias for the `/config` command
939- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
940- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
941- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
942- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
943- New syntax highlighting engine for native build
944
945## 2.0.70
946
947- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
948- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
949- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
950- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
951- Fixed input being cleared when processing queued commands while the user was typing
952- Fixed prompt suggestions replacing typed input when pressing Tab
953- Fixed diff view not updating when terminal is resized
954- Improved memory usage by 3x for large conversations
955- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
956- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
957- Fix thinking mode toggle in /config not persisting correctly
958- Improve UI for file creation permission dialog
959
960## 2.0.69
961
962- Minor bugfixes
963
964## 2.0.68
965
966- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
967- Fixed a bug where disallowed MCP tools were visible to the model
968- Fixed an issue where steering messages could be lost while a subagent is working
969- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
970- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
971- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
972
973## 2.0.67
974
975- Thinking mode is now enabled by default for Opus 4.5
976- Thinking mode configuration has moved to /config
977- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
978- Show reason why autoupdater is disabled in `/doctor`
979- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
980- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
981- Fixed scroll position resetting after deleting a permission rule in `/permissions`
982- 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
983- Fixed `claude install --force` not bypassing stale lock files
984- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
985- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
986
987## 2.0.65
988
989- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
990- Added context window information to status line input
991- Added `fileSuggestion` setting for custom `@` file search commands
992- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
993- Fixed prompt not being saved to history when aborting a query with Escape
994- Fixed Read tool image handling to identify format from bytes instead of file extension
995
996## 2.0.64
997
998- Made auto-compacting instant
999- Agents and bash commands can run asynchronously and send messages to wake up the main agent
1000- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
1001- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
1002- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
1003- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
1004- Fixed auto-loading .env when using native installer
1005- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
1006- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
1007- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
1008- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
1009- Bedrock: Improve efficiency of token counting
1010- Bedrock: Add support for `aws login` AWS Management Console credentials
1011- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
1012
1013## 2.0.62
1014
1015- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
1016- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
1017- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
1018- Fixed slash command selection not working when multiple commands share the same name
1019- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
1020- Fixed running versions getting removed because lock file incorrectly going stale
1021- Fixed IDE diff tab not closing when rejecting file changes
1022
1023## 2.0.61
1024
1025- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
1026
1027## 2.0.60
1028
1029- Added background agent support. Agents run in the background while you work
1030- Added --disable-slash-commands CLI flag to disable all slash commands
1031- Added model name to "Co-Authored-By" commit messages
1032- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
1033- Updated Fetch to skip summarization for pre-approved websites
1034- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
1035
1036## 2.0.59
1037
1038- Added --agent CLI flag to override the agent setting for the current session
1039- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
1040- VS Code: Fixed .claude.json config file being read from incorrect location
1041
1042## 2.0.58
1043
1044- Pro users now have access to Opus 4.5 as part of their subscription!
1045- Fixed timer duration showing "11m 60s" instead of "12m 0s"
1046- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
1047
1048## 2.0.57
1049
1050- Added feedback input when rejecting plans, allowing users to tell Claude what to change
1051- VSCode: Added streaming message support for real-time response display
1052
1053## 2.0.56
1054
1055- Added setting to enable/disable terminal progress bar (OSC 9;4)
1056- 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.
1057
1058## 2.0.55
1059
1060- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
1061- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
1062- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
1063- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
1064
1065## 2.0.54
1066
1067- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
1068- Fix issue with excessive iTerm notifications
1069
1070## 2.0.52
1071
1072- Fixed duplicate message display when starting Claude with a command line argument
1073- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
1074- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
1075- Permit some uses of `$!` in bash commands
1076
1077## 2.0.51
1078
1079- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
1080- Introducing Claude Code for Desktop: https://claude.com/download
1081- 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
1082- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
1083- Plan Mode now builds more precise plans and executes more thoroughly
1084- Usage limit notifications now easier to understand
1085- Switched `/usage` back to "% used"
1086- Fixed handling of thinking errors
1087- Fixed performance regression
1088
1089## 2.0.50
1090
1091- Fixed bug preventing calling MCP tools that have nested references in their input schemas
1092- Silenced a noisy but harmless error during upgrades
1093- Improved ultrathink text display
1094- Improved clarity of 5-hour session limit warning message
1095
1096## 2.0.49
1097
1098- Added readline-style ctrl-y for pasting deleted text
1099- Improved clarity of usage limit warning message
1100- Fixed handling of subagent permissions
1101
1102## 2.0.47
1103
1104- Improved error messages and validation for `claude --teleport`
1105- Improved error handling in `/usage`
1106- Fixed race condition with history entry not getting logged at exit
1107- Fixed Vertex AI configuration not being applied from `settings.json`
1108
1109## 2.0.46
1110
1111- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
1112
1113## 2.0.45
1114
1115- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
1116- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
1117- Send background tasks to Claude Code on the web by starting a message with `&`
1118
1119## 2.0.43
1120
1121- Added `permissionMode` field for custom agents
1122- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
1123- Added skills frontmatter field to declare skills to auto-load for subagents
1124- Added the `SubagentStart` hook event
1125- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
1126- Fixed duplicate rendering of some messages in the UI
1127- Fixed some visual flickers
1128- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
1129
1130## 2.0.42
1131
1132- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
1133
1134## 2.0.41
1135
1136- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
1137- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
1138- Fixed incorrect labeling of user settings vs project settings in command descriptions
1139- Fixed crash when plugin command hooks timeout during execution
1140- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
1141- Fixed broken security documentation links in trust dialogs and onboarding
1142- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
1143- ctrl-r history search landing on a slash command no longer cancels the search
1144- SDK: Support custom timeouts for hooks
1145- Allow more safe git commands to run without approval
1146- Plugins: Added support for sharing and installing output styles
1147- Teleporting a session from web will automatically set the upstream branch
1148
1149## 2.0.37
1150
1151- Fixed how idleness is computed for notifications
1152- Hooks: Added matcher values for Notification hook events
1153- Output Styles: Added `keep-coding-instructions` option to frontmatter
1154
1155## 2.0.36
1156
1157- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
1158- Fixed queued messages being incorrectly executed as bash commands
1159- Fixed input being lost when typing while a queued message is processed
1160
1161## 2.0.35
1162
1163- Improve fuzzy search results when searching commands
1164- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
1165- 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
1166- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
1167- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
1168
1169## 2.0.34
1170
1171- VSCode Extension: Added setting to configure the initial permission mode for new conversations
1172- Improved file path suggestion performance with native Rust-based fuzzy finder
1173- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
1174- Fixed memory crash when reading or writing large files (especially base64-encoded images)
1175
1176## 2.0.33
1177
1178- Native binary installs now launch quicker.
1179- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
1180- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
1181
1182## 2.0.32
1183
1184- Un-deprecate output styles based on community feedback
1185- Added `companyAnnouncements` setting for displaying announcements on startup
1186- Fixed hook progress messages not updating correctly during PostToolUse hook execution
1187
1188## 2.0.31
1189
1190- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
1191- Vertex: add support for Web Search on supported models
1192- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1193- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
1194- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
1195- Fixed plugin uninstall not removing plugins
1196
1197## 2.0.30
1198
1199- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
1200- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
1201- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
1202- Added prompt-based stop hooks
1203- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1204- Enabled SSE MCP servers on native build
1205- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
1206- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
1207- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
1208- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
1209- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
1210- Fixed bug that saved session permissions to local settings
1211- Fixed MCP tools not being available to sub-agents
1212- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
1213- Fixed delay when navigating through typeahead suggestions with arrow keys
1214- VSCode: Restored selection indicator in input footer showing current file or code selection status
1215
1216## 2.0.28
1217
1218- Plan mode: introduced new Plan subagent
1219- Subagents: claude can now choose to resume subagents
1220- Subagents: claude can dynamically choose the model used by its subagents
1221- SDK: added --max-budget-usd flag
1222- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
1223- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
1224- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
1225- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
1226- Various other bug fixes
1227
1228## 2.0.27
1229
1230- New UI for permission prompts
1231- Added current branch filtering and search to session resume screen for easier navigation
1232- Fixed directory @-mention causing "No assistant message found" error
1233- VSCode Extension: Add config setting to include .gitignored files in file searches
1234- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
1235
1236## 2.0.25
1237
1238- 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
1239
1240## 2.0.24
1241
1242- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
1243- Claude Code Web: Support for Web -> CLI teleport
1244- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
1245- Bedrock: Display awsAuthRefresh output when auth is required
1246
1247## 2.0.22
1248
1249- Fixed content layout shift when scrolling through slash commands
1250- IDE: Add toggle to enable/disable thinking.
1251- Fix bug causing duplicate permission prompts with parallel tool calls
1252- Add support for enterprise managed MCP allowlist and denylist
1253
1254## 2.0.21
1255
1256- Support MCP `structuredContent` field in tool responses
1257- Added an interactive question tool
1258- Claude will now ask you questions more often in plan mode
1259- Added Haiku 4.5 as a model option for Pro users
1260- Fixed an issue where queued commands don't have access to previous messages' output
1261
1262## 2.0.20
1263
1264- Added support for Claude Skills
1265
1266## 2.0.19
1267
1268- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
1269- Fixed a bug where Haiku was unnecessarily called in print mode
1270
1271## 2.0.17
1272
1273- Added Haiku 4.5 to model selector!
1274- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
1275- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
1276- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
1277- OTEL: support HTTP_PROXY and HTTPS_PROXY
1278- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
1279
1280## 2.0.15
1281
1282- Fixed bug with resuming where previously created files needed to be read again before writing
1283- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
1284
1285## 2.0.14
1286
1287- Fix @-mentioning MCP servers to toggle them on/off
1288- Improve permission checks for bash with inline env vars
1289- Fix ultrathink + thinking toggle
1290- Reduce unnecessary logins
1291- Document --system-prompt
1292- Several improvements to rendering
1293- Plugins UI polish
1294
1295## 2.0.13
1296
1297- Fixed `/plugin` not working on native build
1298
1299## 2.0.12
1300
1301- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
1302- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
1303- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
1304- `/plugin validate` command for validating plugin structure and configuration
1305- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
1306- Plugin documentation available at https://code.claude.com/docs/en/plugins
1307- Comprehensive error messages and diagnostics via `/doctor` command
1308- Avoid flickering in `/model` selector
1309- Improvements to `/help`
1310- Avoid mentioning hooks in `/resume` summaries
1311- Changes to the "verbose" setting in `/config` now persist across sessions
1312
1313## 2.0.11
1314
1315- Reduced system prompt size by 1.4k tokens
1316- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
1317- Fixed Opus fallback rate limit errors appearing incorrectly
1318- Fixed /add-dir command selecting wrong default tab
1319
1320## 2.0.10
1321
1322- Rewrote terminal renderer for buttery smooth UI
1323- Enable/disable MCP servers by @mentioning, or in /mcp
1324- Added tab completion for shell commands in bash mode
1325- PreToolUse hooks can now modify tool inputs
1326- Press Ctrl-G to edit your prompt in your system's configured text editor
1327- Fixes for bash permission checks with environment variables in the command
1328
1329## 2.0.9
1330
1331- Fix regression where bash backgrounding stopped working
1332
1333## 2.0.8
1334
1335- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
1336- IDE: Add drag-and-drop support for files and folders in chat
1337- /context: Fix counting for thinking blocks
1338- Improve message rendering for users with light themes on dark terminals
1339- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
1340
1341## 2.0.5
1342
1343- IDE: Fix IME unintended message submission with Enter and Tab
1344- IDE: Add "Open in Terminal" link in login screen
1345- Fix unhandled OAuth expiration 401 API errors
1346- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
1347
1348## 2.0.1
1349
1350- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
1351- Various bug fixes and presentation improvements
1352
1353## 2.0.0
1354
1355- New native VS Code extension
1356- Fresh coat of paint throughout the whole app
1357- /rewind a conversation to undo code changes
1358- /usage command to see plan limits
1359- Tab to toggle thinking (sticky across sessions)
1360- Ctrl-R to search history
1361- Unshipped claude config command
1362- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1363- SDK: The Claude Code SDK is now the Claude Agent SDK
1364- Add subagents dynamically with `--agents` flag
1365
1366## 1.0.126
1367
1368- Enable /context command for Bedrock and Vertex
1369- Add mTLS support for HTTP-based OpenTelemetry exporters
1370
1371## 1.0.124
1372
1373- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1374- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1375- No longer inform Claude of the list of allowed tools when permission is denied
1376- Fixed security vulnerability in Bash tool permission checks
1377- Improved VSCode extension performance for large files
1378
1379## 1.0.123
1380
1381- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1382- Fixed thinking mode triggering on negation phrases like "don't think"
1383- Fixed rendering performance degradation during token streaming
1384- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1385- Enhanced BashTool environment snapshot logging
1386- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1387- Migrated --debug logging to a file, to enable easy tailing & filtering
1388
1389## 1.0.120
1390
1391- Fix input lag during typing, especially noticeable with large prompts
1392- Improved VSCode extension command registry and sessions dialog user experience
1393- Enhanced sessions dialog responsiveness and visual feedback
1394- Fixed IDE compatibility issue by removing worktree support check
1395- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1396
1397## 1.0.119
1398
1399- Fix Windows issue where process visually freezes on entering interactive mode
1400- Support dynamic headers for MCP servers via headersHelper configuration
1401- Fix thinking mode not working in headless sessions
1402- Fix slash commands now properly update allowed tools instead of replacing them
1403
1404## 1.0.117
1405
1406- Add Ctrl-R history search to recall previous commands like bash/zsh
1407- Fix input lag while typing, especially on Windows
1408- Add sed command to auto-allowed commands in acceptEdits mode
1409- Fix Windows PATH comparison to be case-insensitive for drive letters
1410- Add permissions management hint to /add-dir output
1411
1412## 1.0.115
1413
1414- Improve thinking mode display with enhanced visual effects
1415- Type /t to temporarily disable thinking mode in your prompt
1416- Improve path validation for glob and grep tools
1417- Show condensed output for post-tool hooks to reduce visual clutter
1418- Fix visual feedback when loading state completes
1419- Improve UI consistency for permission request dialogs
1420
1421## 1.0.113
1422
1423- Deprecated piped input in interactive mode
1424- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1425
1426## 1.0.112
1427
1428- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1429- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1430- Hooks: Added systemMessage support for SessionEnd hooks
1431- Added `spinnerTipsEnabled` setting to disable spinner tips
1432- IDE: Various improvements and bug fixes
1433
1434## 1.0.111
1435
1436- /model now validates provided model names
1437- Fixed Bash tool crashes caused by malformed shell syntax parsing
1438
1439## 1.0.110
1440
1441- /terminal-setup command now supports WezTerm
1442- MCP: OAuth tokens now proactively refresh before expiration
1443- Fixed reliability issues with background Bash processes
1444
1445## 1.0.109
1446
1447- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1448
1449## 1.0.106
1450
1451- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1452
1453## 1.0.97
1454
1455- Settings: /doctor now validates permission rule syntax and suggests corrections
1456
1457## 1.0.94
1458
1459- Vertex: add support for global endpoints for supported models
1460- /memory command now allows direct editing of all imported memory files
1461- SDK: Add custom tools as callbacks
1462- Added /todos command to list current todo items
1463
1464## 1.0.93
1465
1466- Windows: Add alt + v shortcut for pasting images from clipboard
1467- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1468
1469## 1.0.90
1470
1471- Settings file changes take effect immediately - no restart required
1472
1473## 1.0.88
1474
1475- Fixed issue causing "OAuth authentication is currently not supported"
1476- Status line input now includes `exceeds_200k_tokens`
1477- Fixed incorrect usage tracking in /cost.
1478- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1479- Bedrock: Updated default Sonnet model to Sonnet 4
1480
1481## 1.0.86
1482
1483- Added /context to help users self-serve debug context issues
1484- SDK: Added UUID support for all SDK messages
1485- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1486
1487## 1.0.85
1488
1489- Status line input now includes session cost info
1490- Hooks: Introduced SessionEnd hook
1491
1492## 1.0.84
1493
1494- Fix tool_use/tool_result id mismatch error when network is unstable
1495- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1496- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1497- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1498
1499## 1.0.83
1500
1501- @-mention: Support files with spaces in path
1502- New shimmering spinner
1503
1504## 1.0.82
1505
1506- SDK: Add request cancellation support
1507- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1508- Settings: Validation prevents invalid fields in .claude/settings.json files
1509- MCP: Improve tool name consistency
1510- Bash: Fix crash when Claude tries to automatically read large files
1511
1512## 1.0.81
1513
1514- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1515- Agents: Fix custom agent loading when agent files are unparsable
1516
1517## 1.0.80
1518
1519- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1520
1521## 1.0.77
1522
1523- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1524- SDK: Add session support and permission denial tracking
1525- Fix token limit errors in conversation summarization
1526- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1527
1528## 1.0.73
1529
1530- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1531- MCP: Press Esc to cancel OAuth authentication flows
1532- Bash: Improved command validation and reduced false security warnings
1533- UI: Enhanced spinner animations and status line visual hierarchy
1534- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1535
1536## 1.0.72
1537
1538- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1539
1540## 1.0.71
1541
1542- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1543- Customizable status line: add your terminal prompt to Claude Code with /statusline
1544
1545## 1.0.70
1546
1547- Performance: Optimized message rendering for better performance with large contexts
1548- Windows: Fixed native file search, ripgrep, and subagent functionality
1549- Added support for @-mentions in slash command arguments
1550
1551## 1.0.69
1552
1553- Upgraded Opus to version 4.1
1554
1555## 1.0.68
1556
1557- Fix incorrect model names being used for certain commands like `/pr-comments`
1558- 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.
1559- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1560- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1561- SDK: Added canUseTool callback support for tool confirmation
1562- Added `disableAllHooks` setting
1563- Improved file suggestions performance in large repos
1564
1565## 1.0.65
1566
1567- IDE: Fixed connection stability issues and error handling for diagnostics
1568- Windows: Fixed shell environment setup for users without .bashrc files
1569
1570## 1.0.64
1571
1572- Agents: Added model customization support - you can now specify which model an agent should use
1573- Agents: Fixed unintended access to the recursive agent tool
1574- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1575- SDK: Fixed user input tracking across multi-turn conversations
1576- Added hidden files to file search and @-mention suggestions
1577
1578## 1.0.63
1579
1580- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1581
1582## 1.0.62
1583
1584- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1585- Hooks: Added SessionStart hook for new session initialization
1586- /add-dir command now supports typeahead for directory paths
1587- Improved network connectivity check reliability
1588
1589## 1.0.61
1590
1591- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1592- Settings: Added `--settings` flag to load settings from a JSON file
1593- Settings: Fixed resolution of settings files paths that are symlinks
1594- OTEL: Fixed reporting of wrong organization after authentication changes
1595- Slash commands: Fixed permissions checking for allowed-tools with Bash
1596- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1597- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1598- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1599
1600## 1.0.60
1601
1602- You can now create custom subagents for specialized tasks! Run /agents to get started
1603
1604## 1.0.59
1605
1606- SDK: Added tool confirmation support with canUseTool callback
1607- SDK: Allow specifying env for spawned process
1608- Hooks: Exposed PermissionDecision to hooks (including "ask")
1609- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1610- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1611
1612## 1.0.58
1613
1614- Added support for reading PDFs
1615- MCP: Improved server health status display in 'claude mcp list'
1616- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1617
1618## 1.0.57
1619
1620- Added support for specifying a model in slash commands
1621- Improved permission messages to help Claude understand allowed tools
1622- Fix: Remove trailing newlines from bash output in terminal wrapping
1623
1624## 1.0.56
1625
1626- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1627- Fixes for WSL IDE detection
1628- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1629
1630## 1.0.55
1631
1632- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1633- Windows: fixed Ctrl+Z crash
1634- SDK: Added ability to capture error logging
1635- Add --system-prompt-file option to override system prompt in print mode
1636
1637## 1.0.54
1638
1639- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1640- Custom slash commands: Added argument-hint to frontmatter
1641- Windows: OAuth uses port 45454 and properly constructs browser URL
1642- Windows: mode switching now uses alt + m, and plan mode renders properly
1643- Shell: Switch to in-memory shell snapshot to fix file-related errors
1644
1645## 1.0.53
1646
1647- Updated @-mention file truncation from 100 lines to 2000 lines
1648- 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).
1649
1650## 1.0.52
1651
1652- Added support for MCP server instructions
1653
1654## 1.0.51
1655
1656- Added support for native Windows (requires Git for Windows)
1657- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1658- Settings: /doctor can now help you identify and fix invalid setting files
1659- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1660- Increased auto-compact warning threshold from 60% to 80%
1661- Fixed an issue with handling user directories with spaces for shell snapshots
1662- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1663- Custom slash commands: Fixed user-level commands in subdirectories
1664- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1665
1666## 1.0.48
1667
1668- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1669- Added progress messages to Bash tool based on the last 5 lines of command output
1670- Added expanding variables support for MCP server configuration
1671- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1672- Improved IDE extension path handling when Claude Code runs in WSL
1673- Hooks: Added a PreCompact hook
1674- Vim mode: Added c, f/F, t/T
1675
1676## 1.0.45
1677
1678- Redesigned Search (Grep) tool with new tool input parameters and features
1679- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1680- Fixed config file corruption issue by enforcing atomic writes
1681- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1682- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1683- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1684
1685## 1.0.44
1686
1687- New /export command lets you quickly export a conversation for sharing
1688- MCP: resource_link tool results are now supported
1689- MCP: tool annotations and tool titles now display in /mcp view
1690- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1691
1692## 1.0.43
1693
1694- Fixed a bug where the theme selector was saving excessively
1695- Hooks: Added EPIPE system error handling
1696
1697## 1.0.42
1698
1699- Added tilde (`~`) expansion support to `/add-dir` command
1700
1701## 1.0.41
1702
1703- Hooks: Split Stop hook triggering into Stop and SubagentStop
1704- Hooks: Enabled optional timeout configuration for each command
1705- Hooks: Added "hook_event_name" to hook input
1706- Fixed a bug where MCP tools would display twice in tool list
1707- New tool parameters JSON for Bash tool in `tool_decision` event
1708
1709## 1.0.40
1710
1711- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1712
1713## 1.0.39
1714
1715- New Active Time metric in OpenTelemetry logging
1716
1717## 1.0.38
1718
1719- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1720
1721## 1.0.37
1722
1723- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1724
1725## 1.0.36
1726
1727- Web search now takes today's date into context
1728- Fixed a bug where stdio MCP servers were not terminating properly on exit
1729
1730## 1.0.35
1731
1732- Added support for MCP OAuth Authorization Server discovery
1733
1734## 1.0.34
1735
1736- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1737
1738## 1.0.33
1739
1740- Improved logging functionality with session ID support
1741- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1742- Improvements to plan mode
1743
1744## 1.0.32
1745
1746- Updated loopback config for litellm
1747- Added forceLoginMethod setting to bypass login selection screen
1748
1749## 1.0.31
1750
1751- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1752
1753## 1.0.30
1754
1755- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1756- Improved file path autocomplete with filename matching
1757- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1758- Enhanced jq regex support for complex filters with pipes and select
1759
1760## 1.0.29
1761
1762- Improved CJK character support in cursor navigation and rendering
1763
1764## 1.0.28
1765
1766- Slash commands: Fix selector display during history navigation
1767- Resizes images before upload to prevent API size limit errors
1768- Added XDG_CONFIG_HOME support to configuration directory
1769- Performance optimizations for memory usage
1770- New attributes (terminal.type, language) in OpenTelemetry logging
1771
1772## 1.0.27
1773
1774- Streamable HTTP MCP servers are now supported
1775- Remote MCP servers (SSE and HTTP) now support OAuth
1776- MCP resources can now be @-mentioned
1777- /resume slash command to switch conversations within Claude Code
1778
1779## 1.0.25
1780
1781- Slash commands: moved "project" and "user" prefixes to descriptions
1782- Slash commands: improved reliability for command discovery
1783- Improved support for Ghostty
1784- Improved web search reliability
1785
1786## 1.0.24
1787
1788- Improved /mcp output
1789- Fixed a bug where settings arrays got overwritten instead of merged
1790
1791## 1.0.23
1792
1793- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1794- Released Python SDK: pip install claude-code-sdk to get started
1795
1796## 1.0.22
1797
1798- SDK: Renamed `total_cost` to `total_cost_usd`
1799
1800## 1.0.21
1801
1802- Improved editing of files with tab-based indentation
1803- Fix for tool_use without matching tool_result errors
1804- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1805
1806## 1.0.18
1807
1808- Added --add-dir CLI argument for specifying additional working directories
1809- Added streaming input support without require -p flag
1810- Improved startup performance and session storage performance
1811- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1812- Added detailed MCP server tools display (/mcp)
1813- MCP authentication and permission improvements
1814- Added auto-reconnection for MCP SSE connections on disconnect
1815- Fixed issue where pasted content was lost when dialogs appeared
1816
1817## 1.0.17
1818
1819- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1820- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1821- MCP server list UI improvements
1822- Update Claude Code process title to display "claude" instead of "node"
1823
1824## 1.0.11
1825
1826- Claude Code can now also be used with a Claude Pro subscription
1827- Added /upgrade for smoother switching to Claude Max plans
1828- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1829- Improved shell configuration error handling
1830- Improved todo list handling during compaction
1831
1832## 1.0.10
1833
1834- Added markdown table support
1835- Improved streaming performance
1836
1837## 1.0.8
1838
1839- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1840- Increased default otel interval from 1s -> 5s
1841- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1842- Fixed a regression where search tools unnecessarily asked for permissions
1843- Added support for triggering thinking non-English languages
1844- Improved compacting UI
1845
1846## 1.0.7
1847
1848- Renamed /allowed-tools -> /permissions
1849- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1850- Deprecated claude config commands in favor of editing settings.json
1851- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1852- Improved error handling for /install-github-app
1853- Bugfixes, UI polish, and tool reliability improvements
1854
1855## 1.0.6
1856
1857- Improved edit reliability for tab-indented files
1858- Respect CLAUDE_CONFIG_DIR everywhere
1859- Reduced unnecessary tool permission prompts
1860- Added support for symlinks in @file typeahead
1861- Bugfixes, UI polish, and tool reliability improvements
1862
1863## 1.0.4
1864
1865- Fixed a bug where MCP tool errors weren't being parsed correctly
1866
1867## 1.0.1
1868
1869- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1870- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1871- Updated documentation links and OAuth process descriptions
1872
1873## 1.0.0
1874
1875- Claude Code is now generally available
1876- Introducing Sonnet 4 and Opus 4 models
1877
1878## 0.2.125
1879
1880- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1881- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1882
1883## 0.2.117
1884
1885- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1886- Introduced settings.cleanupPeriodDays
1887- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1888- Introduced --debug mode
1889
1890## 0.2.108
1891
1892- You can now send messages to Claude while it works to steer Claude in real-time
1893- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1894- Fixed a bug where thinking was not working in -p mode
1895- Fixed a regression in /cost reporting
1896- Deprecated MCP wizard interface in favor of other MCP commands
1897- Lots of other bugfixes and improvements
1898
1899## 0.2.107
1900
1901- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1902
1903## 0.2.106
1904
1905- MCP SSE server configs can now specify custom headers
1906- Fixed a bug where MCP permission prompt didn't always show correctly
1907
1908## 0.2.105
1909
1910- Claude can now search the web
1911- Moved system & account status to /status
1912- Added word movement keybindings for Vim
1913- Improved latency for startup, todo tool, and file edits
1914
1915## 0.2.102
1916
1917- Improved thinking triggering reliability
1918- Improved @mention reliability for images and folders
1919- You can now paste multiple large chunks into one prompt
1920
1921## 0.2.100
1922
1923- Fixed a crash caused by a stack overflow error
1924- Made db storage optional; missing db support disables --continue and --resume
1925
1926## 0.2.98
1927
1928- Fixed an issue where auto-compact was running twice
1929
1930## 0.2.96
1931
1932- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1933
1934## 0.2.93
1935
1936- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1937- Claude now has access to a Todo list that helps it stay on track and be more organized
1938
1939## 0.2.82
1940
1941- Added support for --disallowedTools
1942- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1943
1944## 0.2.75
1945
1946- Hit Enter to queue up additional messages while Claude is working
1947- Drag in or copy/paste image files directly into the prompt
1948- @-mention files to directly add them to context
1949- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1950- Improved performance for filename auto-complete
1951
1952## 0.2.74
1953
1954- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1955- Task tool can now perform writes and run bash commands
1956
1957## 0.2.72
1958
1959- Updated spinner to indicate tokens loaded and tool usage
1960
1961## 0.2.70
1962
1963- Network commands like curl are now available for Claude to use
1964- Claude can now run multiple web queries in parallel
1965- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1966
1967## 0.2.69
1968
1969- Fixed UI glitches with improved Select component behavior
1970- Enhanced terminal output display with better text truncation logic
1971
1972## 0.2.67
1973
1974- Shared project permission rules can be saved in .claude/settings.json
1975
1976## 0.2.66
1977
1978- Print mode (-p) now supports streaming output via --output-format=stream-json
1979- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1980
1981## 0.2.63
1982
1983- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1984
1985## 0.2.61
1986
1987- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1988- Enhanced image detection for more reliable clipboard paste functionality
1989- Fixed an issue where ESC key could crash the conversation history selector
1990
1991## 0.2.59
1992
1993- Copy+paste images directly into your prompt
1994- Improved progress indicators for bash and fetch tools
1995- Bugfixes for non-interactive mode (-p)
1996
1997## 0.2.54
1998
1999- Quickly add to Memory by starting your message with '#'
2000- Press ctrl+r to see full output for long tool results
2001- Added support for MCP SSE transport
2002
2003## 0.2.53
2004
2005- New web fetch tool lets Claude view URLs that you paste in
2006- Fixed a bug with JPEG detection
2007
2008## 0.2.50
2009
2010- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
2011
2012## 0.2.49
2013
2014- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
2015
2016## 0.2.47
2017
2018- Press Tab to auto-complete file and folder names
2019- Press Shift + Tab to toggle auto-accept for file edits
2020- Automatic conversation compaction for infinite conversation length (toggle with /config)
2021
2022## 0.2.44
2023
2024- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
2025
2026## 0.2.41
2027
2028- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
2029- MCP server startup no longer blocks the app from starting up
2030
2031## 0.2.37
2032
2033- New /release-notes command lets you view release notes at any time
2034- `claude config add/remove` commands now accept multiple values separated by commas or spaces
2035
2036## 0.2.36
2037
2038- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
2039- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
2040
2041## 0.2.34
2042
2043- Vim bindings for text input - enable with /vim or /config
2044
2045## 0.2.32
2046
2047- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
2048- Fix for some PersistentShell issues
2049
2050## 0.2.31
2051
2052- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
2053- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
2054
2055## 0.2.30
2056
2057- Added ANSI color theme for better terminal compatibility
2058- Fixed issue where slash command arguments weren't being sent properly
2059- (Mac-only) API keys are now stored in macOS Keychain
2060
2061## 0.2.26
2062
2063- New /approved-tools command for managing tool permissions
2064- Word-level diff display for improved code readability
2065- Fuzzy matching for slash commands
2066
2067## 0.2.21
2068
2069- Fuzzy matching for /commands