fix(deps): update minor and patch updates#178
Conversation
![renovate-approve[bot]](https://avatars.githubusercontent.com/in/7394?s=60&v=4){kind=small}
Greptile SummaryThis Renovate-generated PR bumps 27 dependencies with minor and patch updates, the most notable being
Confidence Score: 4/5Safe to merge with awareness of the MCP background-connection behaviour change already flagged in a prior review thread. The package.json — specifically the Important Files Changed
|
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| @@ -45,15 +45,15 @@ | |||
| } | |||
| }, | |||
There was a problem hiding this comment.
In @anthropic-ai/claude-agent-sdk v0.3.142 (included in ~0.3.0), MCP servers now connect in the background by default — sessions start immediately and slow servers are reported as status: "pending" in the init event until ready. Previously the SDK waited up to 5 s for servers to connect before starting the first turn.
The agent configures a single, critical MCP server (@workos/mcp-docs-server) that provides the WorkOS documentation used throughout the integration workflow. Under the new default, the agent can begin its first turn before that server is ready, meaning WorkOS MCP tools may be absent or pending on turn 1. The existing AgentSignals.ERROR_MCP_MISSING guard only fires if the agent explicitly emits the [ERROR-MCP-MISSING] signal string — it will not fire for a pending-state MCP tool failure, so the regression could be silent.
To preserve the previous behaviour, either set MCP_CONNECTION_NONBLOCKING=0 in sdkEnv or add alwaysLoad: true to the workos MCP server entry in src/lib/agent-interface.ts.
Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
renovate
Bot
force-pushed
the
renovate/minor-and-patch-updates
branch
from
fda5371 to
cfd85d1
Compare
This PR contains the following updates:
~0.2.62→~0.3.0^0.78.0→^0.102.01.0.1→1.4.11.0.1→1.5.11.19.13→1.19.141.2.0→1.3.0^0.4.0→^0.7.022.19.7→22.19.205.1.0→5.5.019.2.14→19.2.174.0.18→4.1.84.0.18→4.1.88.7.0→8.13.02.0.0→2.4.0^0.1.0→^0.6.08.0.3→8.0.417.3.1→17.4.2^0.35.0→^0.54.01.50.0→1.69.010.28.2→10.34.119.2.4→19.2.77.7.4→7.8.34.21.0→4.22.413.0.1→13.0.24.1.0→4.1.85.28.0→5.32.02.8.3→2.9.04.3.6→4.4.3Release Notes
anthropics/claude-agent-sdk-typescript (@anthropic-ai/claude-agent-sdk)
v0.3.169Compare Source
usage_EXPERIMENTAL_MAY_CHANGE_DO_NOT_RELY_ON_THIS_API_YET()method onQueryreturning structured session cost, plan rate-limit, and local usage-behaviors datasseoption (SSEOptions) toBrowserQueryOptionsas an alternative towebsocket, for browser SDK consumers who prefer Server-Sent Eventsv0.3.168Compare Source
v0.3.167Compare Source
v0.3.166Compare Source
mcp_set_serverscontrol requestv0.3.165Compare Source
v0.3.163Compare Source
stop_taskcontrol requests now return success when the target task is already gone (not_foundornot_running), so SDK clients can reliably prune stale task chipsclaude-in-chrome) viasetMcpServerswhen the CLI was launched without themadditionalContextinhookSpecificOutput, enabling non-error feedback that continues the turnv0.3.162Compare Source
stop_reason: "refusal"andstop_detailson the assistant message and in session transcripts, so SDK consumers can detect refusals without text-matching the error contentfind/grepsearch in Bash, matching the interactive CLI, instead of always registering the dedicated Grep/Glob tools. To keep the dedicated tools (e.g. to intercept searches viacanUseToolor hooks), name them in thetoolsoption or reference them inallowedToolsv0.3.161Compare Source
initializecontrol request is now idempotent: a secondinitializereturns the same success payload instead of anAlready initializederror.ControlResponsegains an optionalpending_permission_requestsfield, mirroringControlErrorResponseapplyFlagSettingsnow live-appliesagentchanges: switching the active agent (or passingnullto reset) takes effect on the next turn in a running sessionv0.3.160Compare Source
resultmessage instead of hanging the calling processv0.3.159Compare Source
v0.3.158Compare Source
v0.3.157Compare Source
v0.3.156Compare Source
v0.3.154Compare Source
v0.3.153Compare Source
v0.3.152Compare Source
SessionStarthooks can now returnreloadSkills: trueto trigger a skill re-scan, and set the session title viahookSpecificOutput.sessionTitleMessageDisplayhook event that lets hooks transform or hide assistant message text as it is displayedv0.3.150Compare Source
v0.3.149Compare Source
options.envdroppingCLAUDE_AGENT_SDK_VERSION(used forUser-Agentand telemetry) when a custom environment is supplied, and corrected theOptions.envdocs to state that the value replaces the subprocess environment rather than merging withprocess.envv0.3.148Compare Source
v0.3.147Compare Source
v0.3.146Compare Source
v0.3.145Compare Source
v0.3.144Compare Source
StopFailurehooks now reporterror: 'model_not_found'when the selected model doesn't exist or isn't available, instead of the generic'invalid_request'. Theapi_error_statusfield on result messages is now documented.@anthropic-ai/claude-agent-sdk/extractexport forbun build --compileconsumers: import the platform native binary withwith { type: 'file' }, callextractFromBunfs(binPath)to copy it out of the compiled executable's virtual filesystem, and pass the result tooptions.pathToClaudeCodeExecutablev0.3.143Compare Source
@anthropic-ai/sdkand@modelcontextprotocol/sdkare nowpeerDependenciesinstead ofdependencies. Runtime is unaffected (both are bundled); npm/bun/pnpm auto-install them. yarn classic users should add them explicitly for full TypeScript type resolutionv0.3.142Compare Source
unstable_v2_createSession,unstable_v2_resumeSession,unstable_v2_prompt,SDKSession,SDKSessionOptions), deprecated since 0.2.133. Usequery()— pass anAsyncIterable<SDKUserMessage>for multi-turn, oroptions.resumeto continue a session.status: "pending"ininituntil ready. SetMCP_CONNECTION_NONBLOCKING=0to restore the old behavior of waiting up to 5s before the first query, or mark a serveralwaysLoad: trueto require it in turn 1.TaskCreate/TaskUpdate/TaskGet/TaskList) instead ofTodoWrite, deprecated since 0.2.136. Tool consumers should accumulate by task ID instead of replacing a snapshot list.request_id,subagent_type, andtask_descriptionon SDK message types and task system events--sdk-urlsessions now exit non-zero with a stderr diagnostic when the remote transport closes permanently (401/403/404 or WS permanent close), instead of silently exiting 0v0.2.141Compare Source
TaskCreateInput,TaskCreateOutput,TaskGetInput,TaskGetOutput,TaskUpdateInput,TaskUpdateOutput,TaskListInput, andTaskListOutputtypes are now exported from@anthropic-ai/claude-agent-sdk/sdk-toolsand included in theToolInputSchemas/ToolOutputSchemasunions@anthropic-ai/sdkdependency to ^0.93.0v0.2.140Compare Source
v0.2.139Compare Source
v0.2.138Compare Source
v0.2.137Compare Source
v0.2.136Compare Source
resolveSettings()(alpha) to inspect effective merged settings without spawning the Claude CLI; reads MDM (plist/HKLM/HKCU) for parity with CLI startupTodoWritetool — future versions will switch to Task tools (TaskCreate,TaskGet,TaskUpdate,TaskList)v0.2.133Compare Source
unstable_v2_createSession/unstable_v2_resumeSession/unstable_v2_prompt) — usequery()instead'Skill'inallowedTools— use theskillsoption insteadv0.2.132Compare Source
applyFlagSettings()in the TypeScript Agent SDK reference and added support fornullon top-level keys to clear flag-settings overridesv0.2.131Compare Source
v0.2.129Compare Source
v0.2.128Compare Source
v0.2.126originto result messages (SDKResultSuccess/SDKResultError) — forwards the triggering message'sSDKMessageOriginso consumers can distinguish user-prompted results fromtask-notificationfollowupsv0.2.124v0.2.123Compare Source
v0.2.122Compare Source
v0.2.121Compare Source
updatedToolOutputtoPostToolUseHookSpecificOutputfor replacing tool output on all tools.updatedMCPToolOutputis deprecated.v0.2.120Compare Source
skillsoption (string[] | 'all') to control which Skills are loaded into the main session, matching the Python SDKv0.2.119Compare Source
forwardSubagentTextoption to stream subagent text deltas to SDK consumersexcludeDynamicSectionsnow keeps static auto-memory instructions in the cacheable system-prompt block; only the per-user memory directory path and per-machine environment values are relocated to the first user messageSessionStore.append()failures are now retried up to 3 times with short backoff before the batch is dropped andmirror_erroris emittedv0.2.118Compare Source
Options.managedSettingsfor embedders to pass policy-tier settings to the spawned CLI in-memory, honored below IT-controlled managed sourcesv0.2.117Compare Source
v0.2.116Compare Source
v0.2.114Compare Source
v0.2.113Compare Source
sessionStoreoption (alpha) toquery()and session helpers for mirroring session transcripts to external storage, withSessionStore/SessionKey/SessionStoreEntrytypes,InMemorySessionStorereference implementation, andimportSessionToStore()for migrating existing sessionsdeleteSession()for removing a session from disk or aSessionStoreSDKMirrorErrorMessage(subtype: 'mirror_error') to theSDKMessageunion — emitted when asessionStore.append()batch failsoptions.envonce again replacesprocess.envfor the CLI subprocess instead of overlaying it. To add or override individual variables, passenv: { ...process.env, MY_VAR: "x" }titleoption toquery()— sets the session title and skips auto-generationv0.2.112Compare Source
v0.2.111Compare Source
mcp_set_serverscontrol request: remote (http/sse) server entries can now carry per-toolpermission_policyvalues, which are applied to the session's allow/deny rulesstartup()andWarmQueryare now part of the public TypeScript APIoptions.envto overlay the inheritedprocess.envinstead of replacing itv0.2.110Compare Source
unstable_v2_createSessionnot respectingcwd,settingSources, andallowDangerouslySkipPermissionsoptionsshouldQueryfield toSDKUserMessage— set tofalseto append a user message without triggering an assistant turn; fixedshouldQuery: falsemessages incorrectly triggering auto-title generation, prompt suggestions, andUserPromptSubmithooksCLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFICandCLAUDE_CODE_DISABLE_TERMINAL_TITLEv0.2.109Compare Source
v0.2.108Compare Source
SDKStatusnow includes'requesting'; whenincludePartialMessagesis enabled, a{type:'system', subtype:'status', status:'requesting'}message is emitted before each API request in the streamv0.2.107Compare Source
v0.2.105Compare Source
system/memory_recallevent andmemory_pathsonsystem/initfor SDK renderers to surface memory operationserror_max_structured_output_retriesbeing emitted when the final retry attempt succeeded, discarding valid structured outputv0.2.104Compare Source
v0.2.101@anthropic-ai/sdkto^0.81.0and@modelcontextprotocol/sdkto^1.29.0to resolve GHSA-5474-4w2j-mq4c and transitive hono advisoriesawait usingdisposal raced its own cleanup callbackMaxListenersExceededWarningwhen running 11+ concurrentquery()callsv0.2.100v0.2.98Compare Source
v0.2.97Compare Source
v0.2.96Compare Source
v0.2.94Compare Source
getContextUsage()to include agents passed viaoptions.agentsin theagentsbreakdownquery()session endsunhandledRejectionv0.2.92Compare Source
v0.2.91Compare Source
terminal_reasonfield to result messages, exposing why the query loop terminated (completed,aborted_tools,max_turns,blocking_limit, etc.)'auto'to the publicPermissionModetypesandboxoption to defaultfailIfUnavailabletotruewhenenabled: trueis passed —query()will emit an error result and exit if sandbox dependencies are missing, instead of silently running unsandboxed. SetfailIfUnavailable: falseto allow graceful degradation.v0.2.90Compare Source
v0.2.89Compare Source
startup()to pre-warm the CLI subprocess beforequery(), making the first query ~20x faster when startup cost can be paid upfrontincludeSystemMessagesoption togetSessionMessages()to optionally include system messages in session historylistSubagents()andgetSubagentMessages()functions to retrieve subagent conversation history from sessionsincludeHookEventsoption to enable hook lifecycle messages (hook_started,hook_progress,hook_response) for all hook event typesERR_STREAM_WRITE_AFTER_ENDerrors when single-turn queries with SDK MCP servers or hooks have control responses arriving after the result message.describe()metadata being dropped fromcreateSdkMcpServertool schemasside_questionreturning null on resume before the first turn completessettingSourcesempty array causing--setting-sources ""to consume the next CLI flagerror_during_execution,error_max_turns,error_max_budget_usd) to correctly setis_error: truewith descriptive messagesv0.2.87Compare Source
v0.2.86Compare Source
getContextUsage()control method to retrieve a breakdown of context window usage by categorysession_idoptional inSDKUserMessagetype — callers no longer need to provide a session ID when sending user messages, as the SDK assigns one automaticallyanyby adding@anthropic-ai/sdkand@modelcontextprotocol/sdkas dependenciesv0.2.85Compare Source
reloadPlugins()SDK method to reload plugins and receive refreshed commands, agents, and MCP server statuspermissionDecision: "ask"being ignored in SDK modev0.2.84Compare Source
taskBudgetoption for API-side token budget awareness, allowing the model to pace tool use within a token limitenableChannel()method andcapabilitiesfield onMcpServerStatusfor SDK-driven MCP channel activationEffortLeveltype ('low' | 'medium' | 'high' | 'max') for consumers to reference effort values directlyv0.2.83Compare Source
seed_read_statecontrol subtype to seedreadFileStatewith{path, mtime}soEditworks after the originatingReadwas removed from contextsession_state_changedevents to opt-in: setCLAUDE_CODE_EMIT_SESSION_STATE_EVENTS=1to receive themv0.2.81Compare Source
canUseToolnot providing a workingaddRulessuggestion when a write under.claude/skills/{name}/hits the bypass-immune safety checkv0.2.80Compare Source
getSessionMessages()dropping parallel tool results — sessions with parallel tool calls now return all tool_use/tool_result pairsv0.2.79Compare Source
'resume'to theExitReasontype for distinguishing resume-triggered session ends in hooksv0.2.78Compare Source
v0.2.77Compare Source
api_retrysystem messages when retrying transient API errors, exposing attempt count, max retries, delay, and error statusv0.2.76Compare Source
forkSession(sessionId, opts?)for branching conversations from a pointcancel_async_messagecontrol subtype to drop a queued user message by UUID before executionplanFilePathfield toExitPlanModetool input for hooks and SDK consumersSDKElicitationCompleteMessagesystem message for handling MCP server input requests programmaticallyv0.2.75Compare Source
tagandcreatedAtfields toSDKSessionInfogetSessionInfo(sessionId, opts?)for single-session metadata lookupoffsetoption tolistSessionsfor paginationtagSession(sessionId, tag, opts?)for tagging session filesqueued_to_runningstatus toAgentToolOutput— returned whenAgent({resume})targets a still-running agentv0.2.74Compare Source
renameSession(sessionId, title, opts?)for renaming session filesimport typefrom@anthropic-ai/claude-agent-sdk/sdk-toolsfailing under NodeNext/Bundler module resolution (missing exports map entry since v0.2.69)user-invocable: falsebeing included insupportedCommands()and thesystem:initmessage'sslash_commands/skillslistsv0.2.73Compare Source
options.envbeing overridden by the~/.claude/settings.jsonenv block when not usinguseras asettingSourcesoptionv0.2.72Compare Source
agentProgressSummariesoption to enable periodic AI-generated progress summaries for running subagents (foreground and background), emitted ontask_progressevents via the newsummaryfieldgetSettings()appliedsection with runtime-resolvedmodelandeffortvaluestoggleMcpServerandreconnectMcpServerfailing with "Server not found" for servers passed viaquery({mcpServers})v0.2.71Compare Source
v0.2.70Compare Source
type: "http"MCP servers failing with HTTP 406 "Not Acceptable" on Streamable HTTP servers that strictly enforce theAccept: application/json, text/event-streamheaderAgentToolInput.subagent_typeto optional — defaults to thegeneral-purposeagent when omittedv0.2.69Compare Source
toolConfig.askUserQuestion.previewFormatoption to configure the content format ('markdown'or'html') for thepreviewfield on AskUserQuestion tool options. Thepreviewfield andannotationsoutput are now exposed in the public SDK types.supportsFastModefield toModelInfoindicating whether a model supports fast modeagent_id(for subagents) andagent_type(for subagents and--agent) fields to hook eventssdkMcpServersin theinitializecontrol request) getting disconnected when background plugin installation refreshes project MCP configsystem:initandresultevents now emit'Task'as the Agent tool name again (reverted from'Agent', which was an unintentional breaking change in a patch release). The wire name will migrate to'Agent'in the next minor release.updatedPermissionsfrom SDK hosts blocking tool calls with a ZodError; the invalid field is now stripped and a warning is logged instead.getSessionMessages()for large sessions with compacted historyv0.2.68Compare Source
v0.2.66Compare Source
v0.2.64Compare Source
v0.2.63Compare Source
pathToClaudeCodeExecutablefailing when set to a bare command name (e.g.,"claude") that should resolve via PATHsupportedAgents()method to the Query interface to view available subagentsanthropics/anthropic-sdk-typescript (@anthropic-ai/sdk)
v0.102.0Compare Source
Full Changelog: sdk-v0.101.0...sdk-v0.102.0
Features
Bug Fixes
v0.101.0Compare Source
Full Changelog: sdk-v0.100.1...sdk-v0.101.0
Features
Bug Fixes
Chores
Documentation
v0.100.1Compare Source
Full Changelog: sdk-v0.100.0...sdk-v0.100.1
Bug Fixes
Chores
v0.100.0Compare Source
Full Changelog: sdk-v0.99.0...sdk-v0.100.0
Features
Documentation
v0.99.0Compare Source
Full Changelog: sdk-v0.98.1...sdk-v0.99.0
Features
Bug Fixes
Configuration
📅 Schedule: (in timezone UTC)
🚦 Automerge: Enabled.
♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.
👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.
This PR was generated by Mend Renovate. View the repository job log.