Experiment 01 · primitives
Every primitive, every version.
All 395 new tools, commands, settings, env vars, hooks, and flags introduced across 63 Claude Code versions, each traced to the changelog line that shipped it and linked to its full version walkthrough. Changelog-derived, published for mining.
CapabilityKeybindingMCP annotationSDK capabilitySDK-typesSettingStatus line JSON fieldUI enhancementUI indicatorbehavior/settingcapabilitycapability/SDKcommandcommand aliascommand availabilitycommand capabilitycommand enhancementcommand flagcommand syntaxcommand/settingenv varenv var / capabilityenv-varenv-var/capabilityenv-var/skill-referenceenvironment variableenvironment-variablefeaturefieldflagflag-extensionflag/capabilityflag/commandflag/keyflag/optionfrontmatter fieldfrontmatter keyshookhook capabilityhook eventhook fieldhook parameterhook returnhook typehook-fieldhook-return-valuehook-typeinfrastructurekeybindingmanaged settingmanifest fieldobservability-flagoutput featureproject\query capabilitysettingsetting/capabilitysetting/frontmattersetting/tracingshortcuttooltool parametervariable
TeammateIdle hook event 2.1.33 Fires when a teammate agent enters idle state in multi-agent workflows TaskCompleted hook event 2.1.33 Fires when a task completes in multi-agent workflows Task(agent_type) syntax capability 2.1.33 Restricts which sub-agents can be spawned by specifying agent type in agent "tools" frontmatter memory frontmatter field setting 2.1.33 Enables persistent memory for agents with scope options (`user`, `project`, or `local`) Plugin name in skill menu UI enhancement 2.1.33 Shows plugin name alongside skill descriptions in `/skills` menu for discoverability claude auth login command 2.1.41 Initiates interactive authentication flow claude auth status command 2.1.41 Reports current authentication state claude auth logout command 2.1.41 Clears authentication credentials speed attribute (OTel) setting/tracing 2.1.41 Marks fast mode visibility in traces and spans Windows ARM64 support capability 2.1.41 Native binary for win32-arm64 architecture /rename auto-generation command enhancement 2.1.41 Session renaming from conversation context without args Opus 4.6 effort callout setting/capability 2.1.42 Displays one-time notification to eligible users about Opus 4.6 availability Claude Sonnet 4.6 capability 2.1.45 Model support for Claude Sonnet 4.6 --add-dir plugins/marketplaces flag-extension 2.1.45 Extended `--add-dir` to read `enabledPlugins` and `extraKnownMarketplaces` from directories spinnerTipsOverride setting 2.1.45 Customize spinner tips with custom tip array; optionally exclude default tips SDKRateLimitInfo / SDKRateLimitEvent SDK-types 2.1.45 SDK types enabling consumers to receive rate limit status updates including utilization, reset times, and overage information claude.ai MCP connectors capability 2.1.46 Use Model Context Protocol connectors configured in claude.ai within Claude Code CLI ctrl+f background agent kill keybinding 2.1.47 Terminate all background agents at once while allowing manual agent lifecycle control last_assistant_message hook 2.1.47 Access the final assistant response text in Stop and SubagentStop hook inputs without parsing transcripts chat:newline keybinding action keybinding 2.1.47 Configure custom keybindings for multi-line input control added_dirs statusline field setting 2.1.47 Expose workspace directories added via `/add-dir` to external scripts and dashboards --worktree / -w flag flag/command 2.1.49 Start Claude in an isolated git worktree isolation: "worktree" setting 2.1.49 Run subagent in a temporary git worktree Ctrl+F keybinding keybinding 2.1.49 Two-press confirmation to kill all background agents background: true in agent definitions setting 2.1.49 Always run an agent as a background task Plugin settings.json shipping capability 2.1.49 Plugins can provide default configuration File edit tool in simple mode capability 2.1.49 Simple mode now includes the file edit tool alongside Bash ConfigChange hook event hook 2.1.49 Fire when configuration files change during a session SDK model capability fields capability/SDK 2.1.49 Discover model support for effort, adaptive thinking isolation: worktree (agent definitions) capability 2.1.50 Agent definitions can declaratively request running in an isolated git worktree, giving each agent its own checkout for safe parallel fan-out without manual worktree management. WorktreeCreate / WorktreeRemove hooks hook 2.1.50 Two new hook events fire when agent worktree isolation creates or removes a worktree, enabling custom VCS setup/teardown (deps install, env priming, cleanup) around isolated agent runs. claude agents (CLI) command 2.1.50 New CLI subcommand that lists all configured agents. CLAUDE_CODE_DISABLE_1M_CONTEXT env-var 2.1.50 Environment variable to disable the 1M context window support. startupTimeout (LSP config) setting 2.1.50 New configuration option to set the startup timeout for LSP servers. /extra-usage in VS Code command 2.1.50 The /extra-usage command is now supported within VS Code sessions, extending that command to the VSCode surface. claude remote-control command 2.1.51 Subcommand for external builds, enabling local environment serving for all users CLAUDE_CODE_PLUGIN_GIT_TIMEOUT_MS env-var 2.1.51 Configurable git timeout for plugin marketplace operations (default increased from 30s to 120s) Custom npm registries + version pinning capability 2.1.51 Support for custom npm registries and specific version pinning when installing plugins from npm sources BashTool login shell behavior setting 2.1.51 Skip login shell (`-l` flag) by default when shell snapshot is available (previously required `CLAUDE_BASH_NO_LOGIN=true`) CLAUDE_CODE_ACCOUNT_UUID env-var 2.1.51 Synchronously provide account UUID for SDK callers, eliminating race condition in early telemetry CLAUDE_CODE_USER_EMAIL env-var 2.1.51 Synchronously provide user email for SDK callers, eliminating race condition in early telemetry CLAUDE_CODE_ORGANIZATION_UUID env-var 2.1.51 Synchronously provide organization UUID for SDK callers, eliminating race condition in early telemetry Macros/Windows Registry managed settings capability 2.1.51 Managed settings can now be set via macOS plist or Windows Registry /memory command 2.1.59 Manage auto-saved context from sessions /copy command 2.1.59 Interactive picker for code blocks, select individual blocks or full response per-subcommand "always allow" capability 2.1.59 Smarter prefix suggestions for compound bash commands that handle subcommands independently /simplify command 2.1.63 Code review for reuse, simplification, and efficiency improvements /batch command 2.1.63 Batch processing capabilities for repeating tasks ENABLE_CLAUDEAI_MCP_SERVERS=false env-var 2.1.63 Opt out from making claude.ai MCP servers available HTTP hooks capability 2.1.63 POST JSON to external URLs and receive JSON responses instead of shell commands /copy "Always copy full response" setting 2.1.63 Skip code block picker and copy full response directly on future `/copy` invocations /model menu display capability 2.1.63 Show the currently active model in the slash command menu Manual URL paste fallback (OAuth) capability 2.1.63 Paste callback URL to complete MCP OAuth authentication if automatic localhost redirect fails /claude-api skill command 2.1.69 Built-in skill for building applications with the Claude API and Anthropic SDK --name arg for /remote-control flag 2.1.69 Sets a custom session title visible in claude.ai/code sandbox.enableWeakerNetworkIsolation setting 2.1.69 Allows Go programs like `gh`, `gcloud`, `terraform` to verify TLS certs when using a custom MITM proxy with `httpProxyPort` (macOS only) includeGitInstructions setting 2.1.69 Removes built-in commit and PR workflow instructions from Claude's system prompt CLAUDE_CODE_DISABLE_GIT_INSTRUCTIONS env-var 2.1.69 Env-var equivalent of `includeGitInstructions` for removing built-in git instructions from system prompt /reload-plugins command 2.1.69 Activates pending plugin changes without restarting the session ${CLAUDE_SKILL_DIR} variable 2.1.69 Allows skills to reference their own directory in SKILL.md content InstructionsLoaded hook event hook 2.1.69 Fires when CLAUDE.md or `.claude/rules/*.md` files are loaded into context agent_id / agent_type hook fields capability 2.1.69 Adds subagent identity (`agent_id`) and agent type (`agent_type`) to hook events, enabling per-agent routing and audit worktree field on status-line hooks capability 2.1.69 Exposes worktree name, path, branch, and original repo directory in hook commands when running in a `--worktree` session pluginTrustMessage setting 2.1.69 Managed setting to append organization-specific context to the plugin trust warning before installation pathPattern in strictKnownMarketplaces setting 2.1.69 Enables regex-matching of file/directory marketplace sources alongside existing `hostPattern` restrictions git-subdir plugin source type capability 2.1.69 Points a plugin source to a subdirectory within a git repo oauth.authServerMetadataUrl setting 2.1.69 MCP config option to specify a custom OAuth metadata discovery URL when standard discovery fails /loop command 2.1.71 New slash command that runs a prompt or slash command on a recurring interval, e.g. `/loop 5m check the deploy`. cron scheduling tools tool 2.1.71 New tools that schedule recurring prompts to fire within a session — an agent-callable recurrence/scheduling surface relevant to orchestration. voice:pushToTalk setting 2.1.71 New rebindable keybinding in keybindings.json for the voice activation key (default: space), supporting modifier+letter combos like meta+k. bash auto-approval allowlist additions (fmt, comm, cmp, numfmt, expr, test, printf, getconf, seq, tsort, pr) capability 2.1.71 Expands the set of bash commands auto-approved without a permission prompt, adding 11 standard text/number/test utilities. ExitWorktree tool 2.1.72 Leave an `EnterWorktree` session and return to the parent worktree CLAUDE_CODE_DISABLE_CRON env-var 2.1.72 Immediately stop scheduled cron jobs mid-session model parameter on Agent tool capability 2.1.72 Per-invocation model override when spawning subagents /plan <description> optional arg command 2.1.72 Enter plan mode and immediately start with a pre-seeded description /copy → w key flag/key 2.1.72 Write the focused selection directly to a file, bypassing clipboard (SSH-safe) lsof, pgrep, tput, ss, fd, fdfind auto-approval capability 2.1.72 Six read-only tools added to bash auto-approval allowlist — no permission prompt CLAUDE.md HTML comment hiding capability 2.1.72 `<!-- ... -->` comments in CLAUDE.md are hidden from Claude on auto-inject; still visible via Read tool vscode://anthropic.claude-code/open URI handler capability 2.1.72 Open a new Claude Code VSCode tab programmatically with optional `prompt` and `session` query params Effort levels simplified to low/medium/high + /effort auto command/setting 2.1.72 Removed `max` tier; new symbols ○ ◐ ●; `/effort auto` resets to default modelOverrides setting 2.1.73 Map model picker entries to custom provider model IDs (e.g. Bedrock inference profile ARNs) /context (enhanced) command 2.1.74 Identifies context-heavy tools, memory bloat, and capacity warnings with specific optimization tips autoMemoryDirectory setting 2.1.74 Configure a custom directory for auto-memory storage CLAUDE_CODE_SESSIONEND_HOOKS_TIMEOUT_MS env-var 2.1.74 Configure timeout duration for `SessionEnd` hooks on exit /color command 2.1.75 Set a prompt-bar color for your session Opus 4.6 1M context (Max/Team/Enterprise) capability 2.1.75 Expand default reasoning window from the standard context to 1M tokens Session name display on prompt bar setting 2.1.75 Show the current session name on the prompt bar when using `/rename` Memory last-modified timestamps capability 2.1.75 Track and display when memory files were last edited, helping Claude reason about freshness Hook source display in permission prompts capability 2.1.75 Show which system component (settings/plugin/skill) is requesting permission MCP elicitation capability 2.1.76 MCP servers can request structured input mid-task via interactive dialog (form fields or browser URL) Elicitation hook hook 2.1.76 Fires when an MCP server issues an elicitation request; allows agent/host to intercept and override ElicitationResult hook hook 2.1.76 Fires after an elicitation response is resolved; allows agent/host to intercept and override before sending back to MCP server -n / --name <name> flag 2.1.76 Sets a display name for the session at startup worktree.sparsePaths setting 2.1.76 For `claude --worktree` in large monorepos: check out only the listed directories via git sparse-checkout PostCompact hook hook 2.1.76 Fires after compaction completes /effort command 2.1.76 Sets model effort level interactively feedbackSurveyRate setting 2.1.76 Enterprise admins can configure the session quality survey sample rate allowRead setting setting 2.1.77 Re-allows read access within regions marked with `denyRead` in sandbox filesystem rules, enabling fine-grained access recovery /copy N command 2.1.77 Copies the Nth-latest assistant response to clipboard (indexed, not just the last one) /branch command 2.1.77 Renames `/fork` to `/branch` as primary command (backward compatible alias remains) SendMessage({to: agentId}) agent continuation tool 2.1.77 Resume a spawned agent without the deprecated `resume` parameter, now auto-resumes in background StopFailure hook event 2.1.78 Fires when turn ends due to API error (rate limit, auth failure, etc.) ${CLAUDE_PLUGIN_DATA} environment variable 2.1.78 Plugin persistent state that survives plugin updates; `/plugin uninstall` prompts before deleting it effort, maxTurns, disallowedTools frontmatter keys 2.1.78 Configuration support for plugin-shipped agents ANTHROPIC_CUSTOM_MODEL_OPTION environment variable 2.1.78 Add custom entry to `/model` picker with optional `_NAME` and `_DESCRIPTION` suffixed vars for display line-by-line response streaming capability 2.1.78 Response text now streams line-by-line as it's generated tmux terminal notifications passthrough capability 2.1.78 Terminal notifications (iTerm2/Kitty/Ghostty popups, progress bar) reach outer terminal when inside tmux with `set -g allow-passthrough on` --console flag flag 2.1.79 Enables Anthropic Console (API billing) authentication in `claude auth login` "Show turn duration" toggle setting 2.1.79 Displays the duration of each turn in `/config` menu output CLAUDE_CODE_PLUGIN_SEED_DIR multi-path env-var 2.1.79 Accepts multiple seed directories separated by platform delimiter (`:` or `;`) /remote-control command command 2.1.79 Bridges CLI session to claude.ai/code to continue work on browser or phone AI-generated session titles capability 2.1.79 Session tabs automatically generate titles based on first message rate_limits in statusline scripts setting 2.1.80 Display Claude.ai rate limit usage (5-hour and 7-day windows with `used_percentage` and `resets_at`) source: 'settings' for plugins setting 2.1.80 Declare plugin entries inline in settings.json instead of external marketplace CLI tool usage detection capability 2.1.80 Plugin tips now detect CLI tool usage in addition to file pattern matching effort frontmatter setting/frontmatter 2.1.80 Skills and slash commands can specify `effort` in frontmatter to override the model effort level when invoked --channels flag (MCP) flag/capability 2.1.80 Allow MCP servers to push messages into your session (research preview) --bare flag 2.1.81 Scripted `-p` calls without hooks, LSP, plugin sync, or interactive auth; requires `ANTHROPIC_API_KEY` or `apiKeyHelper` --channels capability 2.1.81 Permission relay for channel servers to forward tool approval prompts to your phone showClearContextOnPlanAccept setting 2.1.81 Plan mode setting to restore the "clear context" option (default hidden in 2.1.81) managed-settings.d/ setting 2.1.83 Drop-in directory alongside `managed-settings.json`, letting separate teams deploy independent policy fragments that merge alphabetically CwdChanged hook event hook 2.1.83 Fires when the working directory changes, enabling reactive environment management (e.g., direnv) FileChanged hook event hook 2.1.83 Fires when a tracked file changes, enabling reactive environment management sandbox.failIfUnavailable setting 2.1.83 Exit with an error when sandbox is enabled but cannot start, instead of running unsandboxed disableDeepLinkRegistration setting 2.1.83 Prevent `claude-cli://` protocol handler registration CLAUDE_CODE_SUBPROCESS_ENV_SCRUB=1 env-var 2.1.83 Strip Anthropic and cloud provider credentials from subprocess environments (Bash tool, hooks, MCP stdio servers) Transcript search (/ in Ctrl+O mode) capability 2.1.83 Press `/` in transcript mode to search, `n`/`N` to step through matches Ctrl+X Ctrl+E keybinding flag 2.1.83 Alias for opening the external editor (readline-native binding; `Ctrl+G` still works) [Image #N] chip on paste capability 2.1.83 Pasted images insert a positional chip at the cursor so you can reference them by number in your prompt initialPrompt frontmatter capability 2.1.83 Agents can declare `initialPrompt` in frontmatter to auto-submit a first turn chat:killAgents / chat:fastMode rebindable capability 2.1.83 Both are now rebindable via `~/.claude/keybindings.json` CLAUDE_CODE_DISABLE_NONSTREAMING_FALLBACK env-var 2.1.83 Disable the non-streaming fallback when streaming fails Plugin manifest.userConfig / sensitive: true capability 2.1.83 Plugins can prompt for configuration at enable time, with `sensitive: true` values stored in keychain (macOS) or protected credentials file Clipboard-pasted image on-disk path reference capability 2.1.83 Claude can reference the on-disk path of clipboard-pasted images for file operations Ctrl+L full redraw flag 2.1.83 Clears the screen and forces a full redraw — use to recover when Cmd+K leaves the UI partially blank [VSCode] /rewind picker command 2.1.83 Esc-twice or `/rewind` opens a keyboard-navigable rewind picker PowerShell tool (Windows opt-in) tool 2.1.84 Executes PowerShell commands on Windows, parallel to Bash on Unix ANTHROPIC_DEFAULT_{OPUS,SONNET,HAIKU}_MODEL_SUPPORTS env-var 2.1.84 Overrides effort/thinking capability detection for pinned default models on Bedrock/Vertex/Foundry ANTHROPIC_DEFAULT_{OPUS,SONNET,HAIKU}_MODEL_NAME / _DESCRIPTION env-var 2.1.84 Customizes the `/model` picker label for third-party model deployments CLAUDE_STREAM_IDLE_TIMEOUT_MS env-var 2.1.84 Configures the streaming idle watchdog threshold (default 90s) TaskCreated hook hook 2.1.84 Fires when a task is created via `TaskCreate` WorktreeCreate HTTP hook output (hookSpecificOutput.worktreePath) hook 2.1.84 Allows an HTTP hook to return the created worktree path in `hookSpecificOutput.worktreePath` allowedChannelPlugins managed setting setting 2.1.84 Team/enterprise admins can define a channel plugin allowlist paths: YAML list in rules/skills frontmatter capability 2.1.84 Rules and skills `paths:` frontmatter now accepts a YAML list of globs instead of a single string CLAUDE_CODE_MCP_SERVER_NAME env-var 2.1.85 Names the MCP server in headersHelper scripts, allowing one helper to serve multiple servers CLAUDE_CODE_MCP_SERVER_URL env-var 2.1.85 Routes the MCP server URL in headersHelper scripts, allowing one helper to serve multiple servers Hook if field setting 2.1.85 Conditionally filters when hooks run using permission rule syntax (e.g., `Bash(git *)`) to reduce process spawning Transcript timestamp markers capability 2.1.85 Scheduled tasks (`/loop`, `CronCreate`) now emit timestamp markers in transcripts to track execution Deep link query expansion capability 2.1.85 `claude-cli://open?q=…` queries now support up to 5,000 characters with scroll-to-review warning for long pre-filled prompts PreToolUse updatedInput hook 2.1.85 PreToolUse hooks can return `updatedInput` alongside `permissionDecision: "allow"` to satisfy `AskUserQuestion` without interactive prompts, enabling headless integrations X-Claude-Code-Session-Id header capability 2.1.86 Adds session-aware header to all API requests for proxy aggregation without body parsing .jj and .sl VCS exclusions setting 2.1.86 Extends VCS directory blacklist to skip Jujutsu and Sapling metadata during file operations "defer" permission decision hook-return-value 2.1.89 Allows PreToolUse hooks to pause tool execution and resume later with `-p --resume` for re-evaluation CLAUDE_CODE_NO_FLICKER=1 environment-variable 2.1.89 Opts into flicker-free alt-screen rendering with virtualized scrollback for cleaner terminal output PermissionDenied hook hook-type 2.1.89 Fires after auto mode classifier denials; can return `{retry: true}` to signal model should retry the denied command Named subagents in @ mention capability 2.1.89 Typeahead suggestions now include named subagents for direct mention and reference MCP_CONNECTION_NONBLOCKING=true environment-variable 2.1.89 For `-p` mode, skips MCP connection wait entirely and bounds `--mcp-config` server connections at 5s instead of blocking /buddy command command 2.1.89 April 1st feature: hatch a small creature that watches you code /powerup command 2.1.90 Interactive lessons teaching Claude Code features with animated demos CLAUDE_CODE_PLUGIN_KEEP_MARKETPLACE_ON_FAILURE env-var 2.1.90 Keep the existing marketplace cache when `git pull` fails, useful in offline environments .husky protection setting 2.1.90 `.husky` added to protected directories (acceptEdits mode) _meta["anthropic/maxResultSizeChars"] MCP annotation 2.1.91 Persist larger tool results (up to 500K) without truncation disableSkillShellExecution setting 2.1.91 Disable inline shell execution in skills, custom slash commands, and plugin commands Multi-line deep links capability 2.1.91 Support newlines in `claude-cli://open?q=` URLs via `%0A` encoding Plugin bin/ executables capability 2.1.91 Plugins can ship executables under `bin/` and invoke them as bare commands from Bash forceRemoteSettingsRefresh setting 2.1.92 Blocks startup until remote managed settings are freshly fetched; exits if fetch fails (fail-closed policy) Bedrock setup wizard capability 2.1.92 Interactive flow for AWS Bedrock auth, region config, credential verification, and model pinning, accessible from login screen /cost per-model breakdown command enhancement 2.1.92 Subscription users now see per-model and cache-hit breakdown in `/cost` output /release-notes version picker command enhancement 2.1.92 `/release-notes` is now an interactive version picker instead of static output --remote-control-session-name-prefix flag 2.1.92 Override hostname-based default prefix for Remote Control session names (default: `myhost-graceful-unicorn` format) Prompt cache expiry footer hint (Pro) capability 2.1.92 Pro users see footer hint when returning to session after prompt cache expires, showing roughly how many tokens next turn will send uncached CLAUDE_CODE_USE_MANTLE=1 env-var 2.1.94 Enable Amazon Bedrock powered by Mantle for inference /effort default shift command/setting 2.1.94 Default effort level now medium→high for API-key, Bedrock, Vertex, Foundry, Team, and Enterprise users keep-coding-instructions frontmatter field 2.1.94 Plugin output style control for preserving coding instructions in plugin-rendered output hookSpecificOutput.sessionTitle hook parameter 2.1.94 Set session title from UserPromptSubmit hook via `UserPromptSubmit` hook specification Plugin skill frontmatter name precedence behavior/setting 2.1.94 Plugin skills declared via `"skills": ["./"]` now use frontmatter `name` instead of directory basename for stable invocation name Slack MCP compact header output feature 2.1.94 Compact `Slacked #channel` header with clickable channel link in send-message tool calls Ctrl+O focus toggle Keybinding 2.1.97 Toggles focus view in NO_FLICKER mode showing prompt, one-line tool summary with edit diffstats, and final response refreshInterval Setting 2.1.97 Re-runs the status line command every N seconds workspace.git_worktree Status line JSON field 2.1.97 Populated in status line JSON input when current directory is inside a linked git worktree ● N running indicator UI indicator 2.1.97 Shows count of live subagent instances in `/agents` next to agent types Cedar syntax highlighting Capability 2.1.97 Syntax highlighting for Cedar policy files (`.cedar`, `.cedarpolicy` extensions) CJK sentence punctuation completion Capability 2.1.97 Slash command and `@`-mention completion triggers after CJK sentence punctuation without requiring space Monitor tool tool 2.1.98 Stream events from background scripts line-by-line CLAUDE_CODE_PERFORCE_MODE env-var 2.1.98 Causes Edit/Write/NotebookEdit to fail on read-only files with a `p4 edit` hint instead of silently overwriting them CLAUDE_CODE_SUBPROCESS_ENV_SCRUB + PID namespace isolation env-var/capability 2.1.98 Enables subprocess sandboxing with PID namespace isolation on Linux CLAUDE_CODE_SCRIPT_CAPS env-var 2.1.98 Limits per-session script invocations --exclude-dynamic-system-prompt-sections flag 2.1.98 Print mode flag for improved cross-user prompt caching by excluding dynamic sections workspace.git_worktree (re-listed) capability 2.1.98 **NOT new in 2.1.98** — first shipped in **2.1.97** (identical status-line JSON field). 2.1.98 re-lists it with a near-identical changelog line; demoted to an "already shipped" annotation per RECONCILIATION.md (mirrors the `/effort xhigh` handling in 2.1.154). TRACEPARENT env var in Bash subprocesses env-var 2.1.98 W3C TRACEPARENT injected into Bash tool subprocesses when OTEL tracing enabled, so child spans parent correctly Google Vertex AI setup wizard capability 2.1.98 Interactive wizard from the login screen for GCP auth, project/region config, credential verification, and model pinning LSP clientInfo identification capability 2.1.98 Claude Code now identifies itself to language servers via `clientInfo` in the initialize request /team-onboarding command 2.1.101 Generate a teammate ramp-up guide from your local Claude Code usage CLAUDE_CODE_CERT_STORE=bundled env-var 2.1.101 Override to use only bundled CAs instead of OS CA certificate store trust /ultraplan auto-environment capability 2.1.101 Remote-session features now auto-create a default cloud environment instead of requiring web setup first OTEL_LOG_USER_PROMPTS env-var 2.1.101 Beta tracing can now be configured to log user prompts OTEL_LOG_TOOL_DETAILS env-var 2.1.101 Beta tracing can now be configured to log tool details OTEL_LOG_TOOL_CONTENT env-var 2.1.101 Beta tracing can now be configured to log tool content EnterWorktree path parameter tool parameter 2.1.105 Switch into an existing worktree of the current repository PreCompact hook decision:block hook capability 2.1.105 Block compaction by exiting with code 2 or returning `{"decision":"block"}` Plugin monitors manifest key manifest field 2.1.105 Auto-arm background monitors at session start or on skill invoke /proactive command command alias 2.1.105 Alias for `/loop` ENABLE_PROMPT_CACHING_1H env-var 2.1.108 Opt into 1-hour prompt cache TTL on API key, Bedrock, Vertex, and Foundry FORCE_PROMPT_CACHING_5M env-var 2.1.108 Force 5-minute TTL for prompt caching /recap command 2.1.108 Manually invoke session recap; provides context when returning CLAUDE_CODE_ENABLE_AWAY_SUMMARY env-var 2.1.108 Force recap feature when telemetry is disabled Model slash-command discovery capability 2.1.108 The model can now discover and invoke built-in slash commands via the Skill tool /undo command 2.1.108 Alias for `/rewind` /tui command command 2.1.110 Switch to flicker-free rendering in the same conversation tui setting setting 2.1.110 Configure fullscreen rendering mode push notification tool tool 2.1.110 Claude can send mobile push notifications when Remote Control and "Push when Claude decides" config are enabled /focus command command 2.1.110 Toggle focus view separately from transcript view autoScrollEnabled config setting 2.1.110 Disable conversation auto-scroll in fullscreen mode Ctrl+G editor context option setting/capability 2.1.110 Show Claude's last response as commented context in external editor --resume/--continue task resurrection flag/capability 2.1.110 Resurrect unexpired scheduled tasks on session resume Remote Control slash-command support capability 2.1.110 `/context`, `/exit`, and `/reload-plugins` now work from Remote Control (mobile/web) clients TRACEPARENT/TRACESTATE env support env-var 2.1.110 SDK/headless sessions now read distributed trace linking from environment CLAUDE_CODE_ENABLE_AWAY_SUMMARY env-var env-var 2.1.110 Opt out of session recap for users with telemetry disabled /ultrareview command 2.1.111 New slash command that runs comprehensive code review in the cloud using parallel multi-agent analysis and critique; with no args reviews the current branch, or `/ultrareview <PR#>` fetches and reviews a specific GitHub PR. /less-permission-prompts command 2.1.111 New skill/command that scans transcripts for common read-only Bash and MCP tool calls and proposes a prioritized allowlist for `.claude/settings.json`. xhigh capability 2.1.111 New effort level for Opus 4.7 sitting between `high` and `max`, selectable via `/effort`, `--effort`, and the model picker; other models fall back to `high`. PowerShell tool tool 2.1.111 New (progressively rolling out) PowerShell tool on Windows, opt-in/out via `CLAUDE_CODE_USE_POWERSHELL_TOOL`; on Linux/macOS enable with `CLAUDE_CODE_USE_POWERSHELL_TOOL=1` (requires `pwsh` on PATH). The env var is the new control surface for this tool. OTEL_LOG_RAW_API_BODIES env-var 2.1.111 New environment variable that emits full API request and response bodies as OpenTelemetry log events for debugging. Auto (match terminal) theme setting 2.1.111 New theme option that matches the terminal's dark/light mode, selectable from `/theme`. Low-importance UI setting primitive. sandbox.network.deniedDomains setting 2.1.113 Block specific domains even when a broader `allowedDomains` wildcard would permit them Native Claude Code binary infrastructure 2.1.113 CLI spawns native per-platform binary instead of bundled JavaScript Fullscreen Shift+↑/↓ selection scrolling capability 2.1.113 Shift+↑/↓ now scrolls viewport when extending selection past visible edge Ctrl+A / Ctrl+E multiline readline behavior keybinding 2.1.113 Move to start/end of current logical line in multiline input, matching readline Windows Ctrl+Backspace word deletion keybinding 2.1.113 Delete previous word on Windows URL clickability across wrapped lines capability 2.1.113 Long URLs stay clickable when wrapped in terminals with OSC 8 hyperlinks /loop Esc cancellation command enhancement 2.1.113 Pressing Esc cancels pending `/loop` wakeups and shows "Claude resuming /loop wakeup" for clarity /extra-usage on Remote Control command availability 2.1.113 `/extra-usage` now works from Remote Control (mobile/web) clients @-file autocomplete on Remote Control query capability 2.1.113 Remote Control clients can query `@`-file autocomplete suggestions /ultrareview parallelized launch command enhancement 2.1.113 Faster `/ultrareview` launch with parallelized checks, diffstat in dialog, and animated state vim visual mode (v, V) capability 2.1.118 Select text and apply operators in visual and visual-line mode with selection and visual feedback /theme + ~/.claude/themes/ + plugin themes capability 2.1.118 Create, switch between, hand-edit named custom themes; plugins can ship themes via a `themes/` directory type: "mcp_tool" hooks hook 2.1.118 Hooks can invoke MCP tools directly instead of wrapping them in bash DISABLE_UPDATES env var env-var 2.1.118 Completely block all update paths including manual `claude update` — stricter than `DISABLE_AUTOUPDATER` wslInheritsWindowsSettings policy key setting 2.1.118 WSL on Windows can inherit Windows-side managed settings autoMode "$defaults" token setting 2.1.118 Include `"$defaults"` in `autoMode.allow`, `autoMode.soft_deny`, or `autoMode.environment` to add custom rules alongside the built-in list claude plugin tag command command 2.1.118 Create release git tags for plugins with version validation /config settings persistence setting 2.1.119 Saves theme, editor mode, verbose, etc. to `~/.claude/settings.json` with project/local/policy override precedence prUrlTemplate setting 2.1.119 Points the footer PR badge at a custom code-review URL instead of github.com CLAUDE_CODE_HIDE_CWD env-var 2.1.119 Hides the working directory in the startup logo --from-pr (extended) flag 2.1.119 Accepts GitLab merge-request, Bitbucket pull-request, and GitHub Enterprise PR URLs, not just github.com --print with frontmatter gates flag 2.1.119 `--print` mode now honors `tools:` and `disallowedTools:` frontmatter, matching interactive-mode behavior --agent <name> permissionMode flag 2.1.119 `--agent <name>` now honors the agent definition's `permissionMode` for built-in agents PowerShell auto-approval capability 2.1.119 PowerShell tool commands can now be auto-approved in permission mode, matching Bash behavior PostToolUse / PostToolUseFailure duration_ms hook 2.1.119 Hook inputs now include `duration_ms` (tool execution time, excluding permission prompts and PreToolUse hooks) blockedMarketplaces enforcement setting 2.1.119 Security: now correctly enforces `hostPattern` and `pathPattern` entries Status line effort.level / thinking.enabled hook 2.1.119 stdin JSON output now includes `effort.level` and `thinking.enabled` claude ultrareview [target] command 2.1.120 Run `/ultrareview` non-interactively from CI or scripts with stdout output and structured exit codes ${CLAUDE_EFFORT} env-var/skill-reference 2.1.120 Reference the current effort level within skill content for dynamic skill behavior AI_AGENT environment variable env-var 2.1.120 Automatic subprocess environment variable for attribution of `gh` traffic to Claude Code PowerShell shell fallback (Windows) capability 2.1.120 Windows no longer requires Git for Windows (Git Bash); Claude Code uses PowerShell as shell tool when Git Bash absent alwaysLoad (MCP config) setting 2.1.121 MCP servers with this flag skip tool-search deferral; all tools are immediately available claude plugin prune command 2.1.121 Removes orphaned auto-installed plugin dependencies; `plugin uninstall --prune` cascades /skills search box feature 2.1.121 Type-to-filter search box for finding skills in long lists without scrolling hookSpecificOutput.updatedToolOutput (PostToolUse hooks) hook-field 2.1.121 PostToolUse hooks can now replace tool output for all tools, not just MCP-only CLAUDE_CODE_FORK_SUBAGENT=1 (non-interactive) env-var 2.1.121 SDK and `claude -p` now support fork-subagent in non-interactive sessions mcp_authenticate with redirectUri SDK capability 2.1.121 SDK MCP authentication now supports custom scheme completion and claude.ai connectors OTEL_LOG_USER_PROMPTS gating observability-flag 2.1.121 OpenTelemetry spans now optionally log `user_system_prompt` when gated behind this flag ANTHROPIC_BEDROCK_SERVICE_TIER environment variable 2.1.122 Select Bedrock service tier (`default`, `flex`, or `priority`), sent as the `X-Amzn-Bedrock-Service-Tier` header /resume PR URL search command capability 2.1.122 Paste a GitHub/GitLab/Bitbucket PR URL into `/resume` search box to find the session that created that PR --plugin-url <url> flag 2.1.129 Fetch and use a plugin `.zip` archive from a URL for the current session only CLAUDE_CODE_FORCE_SYNC_OUTPUT=1 env-var 2.1.129 Force-enable synchronized output on terminals where auto-detection fails (e.g. Emacs `eat`) CLAUDE_CODE_PACKAGE_MANAGER_AUTO_UPDATE env-var 2.1.129 Enable automatic background upgrade prompts for Homebrew or WinGet installations CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY=1 env-var 2.1.129 Opt-in to Gateway `/v1/models` discovery for the `/model` picker (now requires explicit opt-in, was automatic in 2.1.126–2.1.128) skillOverrides setting (modes: off, user-invocable-only, name-only) setting 2.1.129 Control skill visibility: `off` hides from model and `/`, `user-invocable-only` hides from model only, `name-only` collapses description Ctrl+R history picker search scope toggle capability 2.1.129 Ctrl+R now defaults to all prompts across all projects; press Ctrl+S to narrow to current project/session Plugin manifest experimental namespace setting 2.1.129 Declare `themes` and `monitors` under `"experimental": { ... }` in plugin manifests (top-level still works but warns) CLAUDE_CODE_SESSION_ID env-var 2.1.132 Passed to Bash tool subprocess environment; matches the `session_id` in hook payloads, allowing bash commands to self-identify their session CLAUDE_CODE_DISABLE_ALTERNATE_SCREEN=1 env-var 2.1.132 Disables the fullscreen alternate-screen renderer, keeping the conversation in the terminal's native scrollback instead of a separate screen worktree.baseRef setting 2.1.133 Choose whether `--worktree`, `EnterWorktree`, and agent-isolation worktrees branch from `origin/<default>` or local `HEAD` sandbox.bwrapPath setting 2.1.133 Specify custom bubblewrap binary location (Linux/WSL) sandbox.socatPath setting 2.1.133 Specify custom socat binary location (Linux/WSL) parentSettingsBehavior setting 2.1.133 Admin-tier key to opt SDK `managedSettings` (parent tier) into policy merge (`'first-wins'` or `'merge'`) effort.level (hooks) capability 2.1.133 Hooks now receive active effort level via JSON input field $CLAUDE_EFFORT env-var 2.1.133 Bash tool commands can read the active effort level CLAUDE_CODE_ENABLE_FEEDBACK_SURVEY_FOR_OTEL env-var 2.1.136 Re-enables the session quality survey for enterprises capturing responses through OpenTelemetry settings.autoMode.hard_deny setting 2.1.136 Defines auto mode classifier rules that block unconditionally regardless of user intent or allow exceptions agent view (claude agents) command 2.1.139 New Research-Preview command and capability: a unified list of every Claude Code session (running, blocked-on-you, or done). The headline primitive for agent orchestration — the single control surface for monitoring a fan-out of sessions. /goal command command 2.1.139 New command that sets a completion condition and keeps Claude working across turns until met; works in interactive, -p, and Remote Control with a live overlay panel. Directly relevant to autonomous per-agent persistence in orchestrated runs. /scroll-speed command command 2.1.139 New command to tune mouse wheel scroll speed with a live preview. claude plugin details <name> command 2.1.139 New subcommand showing a plugin's component inventory and projected per-session token cost. transcript view navigation shortcuts capability 2.1.139 New transcript-view keyboard navigation: `?` for shortcuts, `{`/`}` to jump between user prompts, `v` to toggle the shortcut panel. hook args: string[] field (exec form) hook 2.1.139 New hook config field that spawns the command directly without a shell, so path placeholders never need quoting. hook continueOnBlock config option hook 2.1.139 New PostToolUse hook config option: set true to feed the hook's rejection reason back to Claude and continue the turn. CLAUDE_PROJECT_DIR for MCP stdio servers + ${CLAUDE_PROJECT_DIR} in plugin configs env-var 2.1.139 New surface: MCP stdio servers now receive CLAUDE_PROJECT_DIR in their environment (matching hooks), and plugin configs can interpolate ${CLAUDE_PROJECT_DIR} in commands. New env-var exposure and interpolation point, not a fix. subagent identity headers + OTEL attributes capability 2.1.139 New observability surface for orchestration: subagent API requests carry x-claude-code-agent-id / x-claude-code-parent-agent-id headers, and claude_code.llm_request OTEL spans include agent_id / parent_agent_id attributes — enabling parent/child agent tracing in a fan-out. [VSCode] reopen-closed-session shortcut + claudeCode.enableReopenClosedSessionShortcut setting 2.1.139 New VSCode keybinding (Cmd/Ctrl+Shift+T to reopen most recently closed session tab) plus a new setting to configure it. terminalSequence hook field 2.1.141 Allows hooks to emit desktop notifications, window titles, and bells without a controlling terminal CLAUDE_CODE_PLUGIN_PREFER_HTTPS env-var 2.1.141 Forces GitHub plugin clones over HTTPS instead of SSH for environments without SSH keys ANTHROPIC_WORKSPACE_ID env-var 2.1.141 Scopes workload identity federation tokens to a specific workspace claude agents --cwd <path> command flag 2.1.141 Filters the agent session list to a specific directory /feedback recent sessions capability 2.1.141 `/feedback` can now include sessions from the last 24 hours or 7 days for multi-session issues Rewind "Summarize up to here" command 2.1.141 Compresses earlier context while preserving recent turns in rewind menu Background agent permission preservation capability 2.1.141 Background agents launched via `/bg` or `←←` now preserve the current permission mode instead of reverting to default --add-dir flag flag 2.1.142 Adds directory to dispatched background session configuration --settings flag flag 2.1.142 Configures settings for dispatched background session --mcp-config flag flag 2.1.142 Configures MCP servers for dispatched background session --plugin-dir flag flag 2.1.142 Specifies plugin directory for dispatched background session --permission-mode flag flag 2.1.142 Configures permission behavior for dispatched background session --model flag flag 2.1.142 Sets model for dispatched background session --effort flag flag 2.1.142 Sets reasoning effort level for dispatched background session --dangerously-skip-permissions flag flag 2.1.142 Skips permission prompts for dispatched background session CLAUDE_CODE_OPUS_4_6_FAST_MODE_OVERRIDE env var env-var 2.1.142 Pins fast mode to Opus 4.6 instead of new Opus 4.7 default Root-level SKILL.md skill discovery capability 2.1.142 Plugins with root-level `SKILL.md` and no `skills/` subdirectory are now auto-surfaced as skills LSP server visibility in /plugin capability 2.1.142 `/plugin` details pane and `claude plugin details` now expose LSP servers provided by a plugin /web-setup connection warning capability 2.1.142 `/web-setup` warns before overwriting existing GitHub App connection claude plugin disable with dep enforcement command 2.1.143 Refuses to disable a plugin when another enabled plugin depends on it; provides copy-pasteable chain to disable dependents first claude plugin enable with transitive deps command 2.1.143 Force-enables transitive dependencies when enabling a plugin /plugin marketplace cost projection capability 2.1.143 Shows per-turn and per-invocation token cost estimates in browse pane for plugin context usage worktree.bgIsolation: "none" setting setting 2.1.143 Lets background sessions edit the working copy directly without `EnterWorktree`, for repos where worktrees are impractical CLAUDE_CODE_POWERSHELL_RESPECT_EXECUTION_POLICY env var env-var 2.1.143 Opt-out flag; disables PowerShell `-ExecutionPolicy Bypass` pass-through CLAUDE_CODE_STOP_HOOK_BLOCK_CAP env var env-var 2.1.143 Override for the 8-consecutive-block limit that now caps repeated stop-hook looping claude agents --add-dir flag flag 2.1.143 Apply custom directory to agent dashboard and background sessions dispatched from it claude agents --settings flag flag 2.1.143 Apply custom settings.json to agent dashboard and background sessions dispatched from it claude agents --mcp-config flag flag 2.1.143 Apply custom MCP config to agent dashboard and background sessions dispatched from it claude agents --plugin-dir flag flag 2.1.143 Apply custom plugin directory to agent dashboard and background sessions dispatched from it claude agents --permission-mode flag flag 2.1.143 Set default permission mode for sessions dispatched from agent view claude agents --model flag flag 2.1.143 Set default model for sessions dispatched from agent view claude agents --effort flag flag 2.1.143 Set default effort level for sessions dispatched from agent view claude agents --dangerously-skip-permissions flag flag 2.1.143 Set permission bypass default for sessions dispatched from agent view CLAUDE_CODE_USE_POWERSHELL_TOOL=0 opt-out env-var 2.1.143 Disable PowerShell tool (which is now enabled by default on Windows for Bedrock/Vertex/Foundry) /resume command 2.1.144 Resume background sessions alongside interactive ones, marked with `bg` Elapsed duration in completion notifications capability 2.1.144 Background subagent completion notifications now display elapsed time (e.g. "Agent completed · 3h 2m 5s") /plugin last-updated display capability 2.1.144 The `/plugin` browse and discover panes show when each plugin was last updated /model session-only mode command enhancement 2.1.144 `/model` now changes model for current session only; press `d` to set default for new sessions /usage-credits command 2.1.144 Renamed from `/extra-usage`; old name still works for backward compatibility /bg preserves /add-dir command enhancement 2.1.144 Background sessions now retain directories added via `/add-dir` across detach/reattach /doctor exec-form hint command enhancement 2.1.144 `/doctor` now shows exec-form example when a command hook is missing the `command` field claude agents --json command 2.1.145 List live Claude sessions as JSON for scripting, integration with tmux-resurrect, status bars, and session pickers agent_id OTEL span attribute capability 2.1.145 OTEL trace attribute on `claude_code.tool` spans to track agent identity and enable subagent debugging parent_agent_id OTEL span attribute capability 2.1.145 OTEL trace attribute on `claude_code.tool` spans to track parent-child relationships in subagent spans Status line JSON GitHub context capability 2.1.145 Status line JSON input now includes GitHub repo and PR information when detected for contextual awareness /plugin pre-installation metadata display capability 2.1.145 Discover and Browse screens show a plugin's commands, agents, skills, hooks, and MCP/LSP servers before installation claude agents awaiting-input indicator capability 2.1.145 Terminal tab title now shows the awaiting-input count so an alt-tabbed window tells you when an agent needs attention Mouse hover and click in suggestion list capability 2.1.145 Slash command and @-mention suggestion list now supports mouse hover and click in fullscreen mode background_tasks in Stop/SubagentStop hook input field 2.1.145 Stop and SubagentStop hook input now includes `background_tasks` field for hook-aware task management session_crons in Stop/SubagentStop hook input field 2.1.145 Stop and SubagentStop hook input now includes `session_crons` field for hook-aware cron management /code-review --fix flag 2.1.152 Applies review findings to your working tree after the review, surfacing reuse, simplification, and efficiency suggestions disallowed-tools in frontmatter setting 2.1.152 Remove tools from the model while the skill is active /reload-skills command 2.1.152 Re-scan skill directories without restarting the session SessionStart hook: reloadSkills: true hook return 2.1.152 Re-scan skill directories, making skills installed by the hook available in the same session SessionStart hook: hookSpecificOutput.sessionTitle hook return 2.1.152 Set the session title via hook on startup and resume MessageDisplay hook event hook type 2.1.152 Transform or hide assistant message text as it is displayed pluginSuggestionMarketplaces managed setting 2.1.152 Admin allowlist for org marketplaces whose plugins may be suggested via context-aware tips --scope user\ project\ 2.1.152 local` for `claude plugin marketplace remove` --fallback-model session persistence capability 2.1.152 Switch to fallback model for the rest of the session instead of failing every request Auto mode consent removal capability 2.1.152 Auto mode no longer requires opt-in consent Vim mode: / in NORMAL keybinding 2.1.152 Open reverse history search (like Ctrl+R), matching bash/zsh vi-mode /usage large session files command enhancement 2.1.152 Includes large session files in breakdown; scanned with streaming read for flat memory usage OTEL_METRICS_INCLUDE_ENTRYPOINT=true env var 2.1.152 Include session entrypoint as an OpenTelemetry metric attribute skipLfs flag/option 2.1.153 Skip Git LFS downloads during clone and update for github/git plugin marketplace sources Status line COLUMNS/LINES env vars capability 2.1.153 Status line commands receive terminal width and height so scripts can size output accordingly claude agents autocomplete capability 2.1.153 Autocomplete in dispatch input now suggests native slash commands and bundled skills alongside project skills claude agents PR column format capability 2.1.153 PR column shows `PR #N` for single PR or `N PRs` for multiple /model selection persistence capability 2.1.153 Model selection now saves as default for new sessions; `s` key switches model for current session only dynamic workflows / /workflows command 2.1.154 New capability: ask Claude to create a workflow and it orchestrates work across tens-to-hundreds of background agents; `/workflows` is a new command to view runs. This is the headline new orchestration primitive. /effort xhigh extended to Opus 4.8 capability 2.1.154 **NOT new in 2.1.154.** `xhigh` was introduced in **2.1.111** (for Opus 4.7); here it is merely extended to Opus 4.8 (which now defaults to high effort). Listed as an extension, not a net-new primitive. *(This row was originally mis-classified as "new" — the per-version verifier could not see that xhigh predates this release; see OPUS-THOUGHTS.md §6.)* claude agents ! <command> / claude --bg --exec '<command>' flag 2.1.154 New capability to run a shell command as an attachable/detachable background session, either via `! <command>` inside `claude agents` or via the new `--bg --exec` CLI flags. plugin.json defaultEnabled: false / claude plugin enable setting 2.1.154 New `defaultEnabled: false` setting in plugin.json (or marketplace entry) so plugins ship disabled, plus the new `claude plugin enable` command (and /plugin) to enable them; dependencies of enabled plugins still auto-enable. CLAUDE_CODE_SESSION_ID / CLAUDECODE env vars for stdio MCP subprocesses env-var 2.1.154 Stdio MCP server subprocesses now receive `CLAUDE_CODE_SESSION_ID` and `CLAUDECODE=1` in their environment — a new env-var surface an MCP server can read to detect it's running under Claude Code and key off the session. /chrome → Select browser… capability 2.1.154 New selectable capability to pick which connected browser Claude in Chrome uses, via `/chrome` → "Select browser…" or in-chat when an action runs with multiple connected browsers. Minor. Opus 4.8 (incl. fast mode) model capability 2.1.154 New model available to select/use, including a fast mode on Opus 4.8. Counted as a capability primitive (new model surface); fast-mode pricing/speed detail itself is an extension of the existing fast-mode mechanism. Plugins in .claude/skills directories auto-load capability 2.1.157 Plugins placed in `.claude/skills/` are now automatically discovered and loaded at startup without marketplace registration claude plugin init <name> command 2.1.157 Scaffold a new plugin skeleton in `.claude/skills/` with standard boilerplate Plugin /plugin autocomplete tool 2.1.157 Tab-complete `/plugin` subcommands, installed plugin names, and plugins from known marketplaces agent field in settings.json honored for dispatched sessions setting 2.1.157 The `agent` field in project `settings.json` is now respected when dispatching background sessions; override with `--agent <name>` EnterWorktree mid-session switching tool 2.1.157 `EnterWorktree` can now switch between Claude-managed worktrees without ending the session tool_parameters in telemetry with OTEL_LOG_TOOL_DETAILS=1 env-var 2.1.157 When `OTEL_LOG_TOOL_DETAILS=1` is set, `tool_decision` telemetry events include the actual tool parameters (bash commands, MCP/skill names) for observability Worktrees left unlocked after agent finish capability 2.1.157 Claude-managed worktrees are now left in an unlocked state after the agent finishes, allowing `git worktree remove` and `prune` to clean them up without manual unlock "Workflow keyword trigger" setting in /config setting 2.1.157 New `/config` setting to disable or control the auto-trigger behavior when the word "workflow" appears in a prompt ultracode (dynamic-workflow trigger keyword) capability 2.1.160 The dynamic-workflow trigger keyword was renamed from `workflow` to `ultracode`; the word `workflow` no longer triggers a run, asking in your own words still works, and the keyword is highlighted in violet in the input. A rename of an existing trigger surface, not a new capability — included as the single most primitive-shaped, orchestration-relevant change. grep satisfies read-before-edit capability 2.1.160 Single-file `grep`/`egrep`/`fgrep` commands now satisfy the read-before-edit check, so Edit no longer requires a separate Read after viewing a file with grep. A friction reduction to an existing gate; borderline primitive because it changes what counts as a valid pre-edit read. requiredMinimumVersion / requiredMaximumVersion managed setting 2.1.163 Claude Code refuses to start if its version is outside the allowed range and directs the user to an approved version /plugin list command command 2.1.163 List installed plugins, with `--enabled`/`--disabled` filters "c to copy" on /btw shortcut 2.1.163 Copies the raw markdown answer to the clipboard, preserving formatting when pasted elsewhere hookSpecificOutput.additionalContext for Stop/SubagentStop hook capability 2.1.163 Stop and SubagentStop hooks can return additional context to give Claude feedback and keep the turn going without being labeled a hook error \$ escape syntax in skills command syntax 2.1.163 Escape syntax to include a literal `$` before a digit in command bodies CLAUDE_CODE_SESSION_ID for stdio MCP servers env var / capability 2.1.163 stdio MCP servers now receive the same `CLAUDE_CODE_SESSION_ID` as hooks/Bash on `--resume` fallbackModel setting setting 2.1.166 Configure up to three fallback models tried in order when primary model is overloaded or unavailable --fallback-model flag flag 2.1.166 Apply fallback model routing to interactive sessions (not just background agents) Glob pattern deny rules capability 2.1.166 Use `"*"` in deny rule tool-name position to deny all tools; allow rules reject non-MCP globs MAX_THINKING_TOKENS=0 env-var 2.1.166 Disable thinking on models that think by default via the Claude API --thinking disabled flag 2.1.166 Disable thinking on models with default thinking enabled Per-model thinking toggle setting 2.1.166 Configure thinking per model instead of globally claude update version announcement capability 2.1.166 `claude update` announces target version before downloading claude agents URL filtering capability 2.1.166 Type a URL into the agents list to filter to the session whose first prompt contained it disableBundledSkills / CLAUDE_CODE_DISABLE_BUNDLED_SKILLS setting 2.1.169 A setting plus matching env-var that hides bundled skills, workflows, and built-in slash commands from the model. One feature exposed via two surfaces (settings key + env-var). Most orchestration-relevant primitive: lets a parent strip each worker's default skill/command surface to a controlled allowlist for deterministic fan-out. --safe-mode / CLAUDE_CODE_SAFE_MODE flag 2.1.169 A flag plus matching env-var (one feature, two surfaces) that starts Claude Code with all customizations — CLAUDE.md, plugins, skills, hooks, MCP servers — disabled for troubleshooting. Blunt all-off switch rather than targeted control. /cd command 2.1.169 A slash command to move a session to a new working directory without breaking the prompt cache mid-session. Useful for long-lived orchestrator sessions that need to retarget without a cache-busting restart. claude agents --all (+ id/state fields) flag 2.1.169 Ships inside a fix line, but introduces a genuinely new flag `--all` (include completed sessions) plus new `id` and `state` fields on `claude agents --json`. Counted as a primitive per the task's flag/capability definition. Directly relevant to fan-out observability: enumerating and tracking dispatched background sessions. API_FORCE_IDLE_TIMEOUT env-var 2.1.169 Ships inside a restore-default fix line, but introduces a genuinely new env-var. The 5-minute idle timeout on Vertex/Foundry is a restoration (hardening); the opt-out knob `API_FORCE_IDLE_TIMEOUT=0` is the new primitive surface. Counted per the task's env-var definition.