[dependabot]

Bumps the prod-dependencies group with 12 updates in the / directory:

Package	From	To
@langchain/anthropic	1.5.11	1.6.0
@langchain/langgraph	1.4.18	1.5.2
@sap-ai-sdk/langchain	2.16.0	2.17.0
@sap-ai-sdk/orchestration	2.16.0	2.17.0
deepagents	1.14.1	1.15.1
marked	18.0.14	18.1.1
zod	4.6.5	4.6.7
@cap-js/sqlite	3.1.1	3.1.2
@sap/cds-mtxs	4.1.1	4.2.0
@cap-js/telemetry	2.0.1	2.1.1
@opentelemetry/exporter-trace-otlp-proto	0.223.0	0.300.0
@sap/cds	10.1.1	10.2.1

Updates @langchain/anthropic from 1.5.11 to 1.6.0

Release notes

Sourced from @​langchain/anthropic's releases.

@​langchain/anthropic@​1.6.0

Minor Changes

• #11986 2caf64d Thanks @​richardscarrott! - feat: stream tool-call args as tool-call-delta slices

  Tool-call and server-tool-call args now stream as tool-call-delta events, which carry only the new slice of the args, instead of block-delta snapshots of the args so far. A large tool call now costs about the same bytes on the wire as the same length of text.

  • This covers the compat bridge (any model that only implements _streamResponseChunks), the OpenAI-compatible converter (OpenAI Completions, Groq, Mistral, IBM, OpenRouter, Perplexity, DeepSeek, Fireworks, Together AI and xAI Completions), OpenAI and xAI Responses, Anthropic (client and server tool calls) and Bedrock Converse. Google and Ollama still send one complete snapshot per call.
  • A tool call's id, name and index go on content-block-start, which carries no args. An id or name that arrives later reaches consumers on content-block-finish.
  • If you read raw events and their block-delta fields.args mid-stream, or write agent stream transformers that do, read tool-call-delta instead. await stream, .toolCalls and the assembled message are unchanged.
  • Upgrade your clients first. Older clients mishandle tool-call-delta. You need @langchain/langgraph-sdk 1.12.6 or later; @langchain/react, @langchain/vue, @langchain/svelte or @langchain/angular 1.2.6 or later; and langchain-core 1.6.13 or later for Python clients.
  • To hold back, pin @langchain/core below 1.3.0. That also keeps you on @langchain/openai, @langchain/anthropic, @langchain/aws and @langchain/xai versions that send snapshots.

@​langchain/anthropic@​1.5.13

Patch Changes

• #11922 d71c542 Thanks @​richardscarrott! - feat(anthropic): declare mid-conversation support in model profiles. Claude Opus 4.8, Opus 5, Opus 5.5, Sonnet 5.5, Haiku 5.5, Fable 5 and 5.1, and Mythos 5 and 5.1 set midConversationSystemMessages and midConversationTools. Sonnet 5 sets system messages only.

  Regenerating the profiles from models.dev also adds Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5, removes Claude Opus 4.1, and lowers Claude Sonnet 4.5's maxInputTokens to 200,000.

@​langchain/anthropic@​1.5.12

Patch Changes

• #11707 4109812 Thanks @​richardscarrott! - feat(anthropic): support mid-conversation tool changes. A SystemMessage can now carry Anthropic's tool_addition and tool_removal blocks, either bare in content or wrapped in a non_standard block, and they reach the provider verbatim. The required beta is added automatically: mid-conversation-tool-changes-2026-07-01 for tools named by reference, or inline-tools-2026-09-15 when a tool_addition defines a tool by value (tool: { type: "tool_definition", definition }).

  System content is now narrowed to what Anthropic accepts there. text, tool_addition and tool_removal are sent; any other block is dropped instead of being forwarded for the provider to reject. A system message left with no content is omitted.

Commits

• 38eb4d8 chore: version packages (#12018)
• dae3b1f chore(core): fill in client versions in tool-call-delta changeset (#12019)
• 2caf64d feat(core,openai,anthropic,aws,xai): stream tool-call args as `tool-call-delt...
• 8a008ea test(core): accept either order of a chunk's stream events in chat model chai...
• f65acc3 chore(deps): restore missing @modelcontextprotocol/core@2.3.0 lockfile entr...
• 6a0ab38 chore: version packages (#12008)
• 654c7c6 chore(deps-dev): bump @​modelcontextprotocol/server from 2.3.0 to 2.3.1 (#12005)
• f096a32 fix(langchain): deliver final answers after todo updates (#11989)
• c690992 fix: update Node.js to October 2026 security release (#12007)
• 0c5e7bc fix(langchain): avoid full-state copies in parallel tool tasks (#11990)
• Additional commits viewable in compare view

Updates @langchain/langgraph from 1.4.18 to 1.5.2

Release notes

Sourced from @​langchain/langgraph's releases.

@​langchain/langgraph-api@​1.5.2

Patch Changes

• Updated dependencies []:
  • @​langchain/langgraph-ui@​1.5.2

@​langchain/langgraph-cli@​1.5.2

Patch Changes

• #2887 8f8eaa8 Thanks @​syachamaneni-lc! - add --agent-id and --agent-environment flags
• Updated dependencies []:
  • @​langchain/langgraph-api@​1.5.2

@​langchain/langgraph-ui@​1.5.2

No changes in this release.

@​langchain/langgraph@​1.5.2

Patch Changes

• #3022 a17f609 Thanks @​richardscarrott! - Depend on @langchain/protocol 0.0.20, the version @langchain/langgraph-sdk uses, so installs resolve a single copy and the protocol event types stay compatible.
• Updated dependencies [6e84570, a17f609]:
  • @​langchain/langgraph-checkpoint@​1.2.1
  • @​langchain/langgraph-sdk@​1.12.6

@​langchain/langgraph-api@​1.5.2-dev.0

Patch Changes

• Updated dependencies []:
  • @​langchain/langgraph-ui@​1.5.2-dev.0

@​langchain/langgraph-cli@​1.5.2-dev.0

Patch Changes

• #2887 8f8eaa8 Thanks @​syachamaneni-lc! - add --agent-id and --agent-environment flags
• Updated dependencies []:
  • @​langchain/langgraph-api@​1.5.2-dev.0

@​langchain/langgraph-ui@​1.5.2-dev.0

No changes in this release.

@​langchain/langgraph-api@​1.5.1

Patch Changes

• #2883 545c7d1 Thanks @​casparb! - Fix langgraph dev --no-reload with the default TypeScript loader. Pass the tsx --clear-screen=false option only in watch mode so Node can start the server when reload is disabled.
• Updated dependencies []:
  • @​langchain/langgraph-ui@​1.5.1

@​langchain/langgraph-cli@​1.5.1

Patch Changes

... (truncated)

Changelog

Sourced from @​langchain/langgraph's changelog.

1.5.2

Patch Changes

• #3022 a17f609 Thanks @​richardscarrott! - Depend on @langchain/protocol 0.0.20, the version @langchain/langgraph-sdk uses, so installs resolve a single copy and the protocol event types stay compatible.
• Updated dependencies [6e84570, a17f609]:
  • @​langchain/langgraph-checkpoint@​1.2.1
  • @​langchain/langgraph-sdk@​1.12.6

1.5.1

Patch Changes

• #3017 61b0049 Thanks @​eliornl! - A field that changes from a DeltaChannel to a plain reducer, for example to MessagesAnnotation, keeps its value on threads written under the DeltaChannel: reads rebuild it from the stored writes instead of returning the empty default, and the next run adds to it instead of replacing it.

• #3027 af7e421 Thanks @​eliornl! - With durability: "exit", resuming parallel interrupts one at a time no longer applies a finished task's output twice: the save at exit keeps only the writes made during that run, instead of storing the ones loaded with the checkpoint again under new indexes.

• #2986 53cc627 Thanks @​byhow! - When one answer in a multi-answer resume fails its interrupt's responseSchema, the other tasks in the step now finish before the ZodError is thrown, instead of being aborted. A task aborted after its side effects kept its answer and ran again on the next resume. The error is unchanged, and the new isInvalidResume(error) tells it apart from other errors, as is_invalid_resume does in langgraph (Python).

• #3038 bc01023 Thanks @​eliornl! - messagesStateReducer and messagesDeltaReducer store message chunks as their non-chunk class, as Python's add_messages does. A graph run with streamMode: "messages" or streamEvents now checkpoints the same AIMessage as invoke, without a duplicate tool_call_chunks copy of every tool call, and an in-place edit to tool_calls (for example a rejected call in humanInTheLoopMiddleware) is no longer undone when the thread is reloaded.

• #3028 ec1854a Thanks @​eliornl! - Resuming after some parallel interrupts were answered no longer reruns the nodes and task() calls that already finished, in a subgraph or when the resume is sent with the id of the thread's latest checkpoint. Finished tasks now rerun only on time travel, or on a resume at a checkpoint the thread has moved past.

• Updated dependencies [2e1fe5b, 0acf88d, 9cc54cc, 4690a5b]:

  • @​langchain/langgraph-sdk@​1.12.5

1.5.0

Minor Changes

• #2961 1631fd8 Thanks @​soarez! - order parallel writes by task path, like langgraph (Python) — one sort key for live execution and every saver, instead of the task-id order #2544 added for DeltaChannel writes

  When two or more tasks in one superstep write the same DeltaChannel, their writes are now applied and replayed in (task_path, task_id, idx) order — the order Python's langgraph#8544 established — instead of task-id order. Task ids are uuid5 hashes that include the checkpoint id, so the old order permuted parallel writers from run to run and disagreed with Python.

  • BaseCheckpointSaver.putWrites takes an optional taskPath (the serialized task path, taskPathStr), and getTuple/list must return pendingWrites in the new writesSortKey order. writesSortKey and compareWritesSortKeys are new exports of @langchain/langgraph-checkpoint; every saver (memory, SQLite, Postgres, Redis incl. ShallowRedisSaver, MongoDB) stores the path and sorts in JS — no database collation is involved.
  • SQLite and Postgres gain a task_path column (migrated automatically on setup(); read-only pre-column SQLite databases keep reading, with their rows keeping the old order). Redis and MongoDB documents gain a task_path field. Rows written before the upgrade keep their old order.
  • Exit-durability resumes replay delta writes in live order (port of langgraph#9114): the resumed superstep's writes interleave with the loaded ones by task path, later supersteps sort after every real write, and writes loaded with the checkpoint are no longer stored twice.
  • bulkUpdateState seals the first update superstep: DeltaChannels touched by the base checkpoint's pending writes (and whose version moved) snapshot on the update checkpoint, so a finished sibling's writes and the update's can't replay in the wrong order (partial port of langgraph#8548).
  • bulkUpdateState/updateState update tasks carry (__interrupt__, i) task paths, as in langgraph#9128, so several updates in one superstep writing the same DeltaChannel replay in the order given.
  • Send task paths become [PUSH, i, false] and node-error-handler paths become [*failedTask.path[:3], "node_error_handler", false] (the failed task's first three path elements), matching Python's shapes — langgraph_path task metadata changes accordingly, and task ids of functional task() calls under a Send or handler task change with them. The SDK's Send-task detection accepts both the old and new shapes.

  Upgrade notes:

  • PostgresSaver: run setup() once after upgrading, from one process, before the new version serves traffic. The task_path column comes from the first new migration since the saver shipped, and until it runs, reads and writes fail with column "task_path" does not exist. Several processes running setup() at once on a database that still needs the migration can fail with a duplicate key error on checkpoint_migrations.
  • Third-party checkpoint savers remain source-compatible but must now persist the taskPath they are given and return pendingWrites in writesSortKey order; the conformance suite's new base tests enforce this. A saver that ignores the new argument replays in task-id order and can rebuild a different DeltaChannel value than the run produced.
  • An incomplete superstep containing stored delta writes that must interleave with resumed or retried task writes is affected by the upgrade, whether the incompleteness came from an interrupt, a task error, a crash, or a drain: finish such threads with the old runtime before upgrading writers. A single interrupted node with no finished delta-writing siblings is safe to resume. The same applies to resuming work that includes functional task() calls under a Send or error-handler task, whose child task ids change.
  • @langchain/langgraph-checkpoint 1.2.0 must be released before any package requiring its new exports; the consumers' minimum ranges are raised accordingly.

Patch Changes

• #2980 5c81e5f Thanks @​eliornl! - A node served from the node cache now saves its writes like a node that ran. They were applied but never saved, so a DeltaChannel read back without them once the state was loaded again, under every durability. A cache hit is still streamed as cached and isn't written back to the cache.

... (truncated)

Commits

• 0c12d5c chore: version packages (#3059)
• a17f609 feat(sdk): read tool-call-delta and ignore unknown delta variants (#3022)
• cc29753 chore: version packages (#3018)
• 53cc627 fix(langgraph): let other tasks finish when a resume value is invalid (#2986)
• ec1854a fix(langgraph): stop a subgraph from rerunning its finished tasks on resume (...
• af7e421 fix(langgraph): save a resumed exit run's pending writes once (#3027)
• bc01023 fix(langgraph): store message chunks as messages in the messages reducers (#3...
• b762123 chore(langgraph): consolidate DeltaChannel regression suites (#3019)
• 61b0049 fix(langgraph): keep a DeltaChannel's value when its field changes to a plain...
• 4510774 chore: version packages (#2977)
• Additional commits viewable in compare view

Updates @sap-ai-sdk/langchain from 2.16.0 to 2.17.0

Release notes

Sourced from @​sap-ai-sdk/langchain's releases.

v2.17.0

Compatibility Notes

• [orchestration, prompt-registry] LlamaGuard38B and LlamaGuard38BFilterConfig are now marked as deprecated and will be removed on 2027-09-20. Migrate usages before that date.

New Features

• [context-registry] Add tabular context registry package.
• [core] Add gemini-3.8-flash to the model types.
• [core] Add gpt-6-astra, gpt-6-luna, gpt-6.1-sol, gemini-3.6-flash, and nvidia--llama-nemotron-embed-1b to the model types.
• [langchain] Add reasoning content support to the Orchestration LangChain client. Reasoning blocks from model responses are surfaced in AIMessage.additional_kwargs.reasoning_content(preserving the signature for encrypted thinking), for both non-streaming and streaming responses. On subsequent turns, reasoning is forwarded back to the Orchestration service in reasoning_content from additional_kwargs.reasoning_content. Reasoning placed directly in message content is treated as output only and is not forwarded.
• [orchestration] Update generated client to latest spec (rel-1.13.2). Add optional provider_specific_fields property to LlmChoice and LlmChoiceStreaming response types.
• [orchestration] Update the generated client to spec v1.23. Adds the IBM Granite Guardian 4.1 content filter. Adds input filter target selectors. Adds the drop_unsupported_params option on LLMModelDetails.
• [prompt-registry] Update generated client to latest spec. Add deleteOrchestrationConfigsByScenario, deleteOrchestrationConfigHistory, deletePromptTemplatesByScenario, and deletePromptTemplateHistory API functions.
• [tabular-orchestration] Add experimental Tabular AI Orchestration package with TabularOrchestrationClient for predictions with Tabular Foundation Models.

Fixed Issues

• [langchain] Additional additional_kwargs properties (including tool_calls, intermediate_results, and reasoning_content) are now correctly placed on the AIMessage instead of the generation object.
• [langchain] Avoid retrying irrecoverable orchestration request and unsupported-model errors.
• [orchestration] Preserve prompt_tokens_details and completion_tokens_details in streamed token usage to fix missing cache and reasoning token details in OrchestrationStreamResponse.getTokenUsage() and getIntermediateResults().
• [orchestration] Fix missing prompt template or messages not being detected when no prompt is configured.

Improvements

• [core] Remove deprecated models cohere--command-a-reasoning (retirement date: 2026-10-30), mistralai--mistral-medium-instruct (retirement date: 2026-11-30) — use mistralai--mistral-medium instead, gemini-2.5-flash-lite (retirement date: 2026-10-20) — use gemini-3.5-flash-lite instead, sap-rpt-1.5 (retirement date: 2026-12-31) — use sap-rpt-1.6 instead, sap-rpt-1.5-large (retirement date: 2026-12-31) — use sap-rpt-1.6-large instead, sap-rpt-1-small (retirement date: 2026-12-31) — use sap-rpt-1.6 instead and sap-rpt-1-large (retirement date: 2026-12-31) — use sap-rpt-1.6-large instead.

Changelog

Sourced from @​sap-ai-sdk/langchain's changelog.

2.17.0

Minor Changes

• [feat] Add reasoning content support to the Orchestration LangChain client. Reasoning blocks from model responses are surfaced in AIMessage.additional_kwargs.reasoning_content(preserving the signature for encrypted thinking), for both non-streaming and streaming responses. On subsequent turns, reasoning is forwarded back to the Orchestration service in reasoning_content from additional_kwargs.reasoning_content. Reasoning placed directly in message content is treated as output only and is not forwarded.

Patch Changes

• [fix] Additional additional_kwargs properties (including tool_calls, intermediate_results, and reasoning_content) are now correctly placed on the AIMessage instead of the generation object.

• [fix] Avoid retrying irrecoverable orchestration request and unsupported-model errors.

Commits

• b225ff6 v2.17.0
• 0779863 feat: Support reasoning content in Orchestration langchain client (#2305)
• 18a3219 fix(langchain): Avoid retrying irrecoverable orchestration requests (#2465)
• 731757a chore: enable oxlint correctness-category lint warnings (#2380)
• 6e2857d chore: Use pnpm 11 catalogs to force certain dependencies on common v (#2129)
• f56c0f3 chore(deps): update langchain (#2371)
• bec9ecc chore(deps): update vitest monorepo to v5.0.2 (#2366)
• 74feac4 chore(deps): update dependency @​langchain/langgraph to v1.4.18 (#2365)
• cc67d12 chore(deps): update langchain (#2353)
• ed2375d chore(deps): update dependency @​langchain/core to v1.2.12 (#2348)
• Additional commits viewable in compare view

Updates @sap-ai-sdk/orchestration from 2.16.0 to 2.17.0

Release notes

Sourced from @​sap-ai-sdk/orchestration's releases.

v2.17.0

Compatibility Notes

• [orchestration, prompt-registry] LlamaGuard38B and LlamaGuard38BFilterConfig are now marked as deprecated and will be removed on 2027-09-20. Migrate usages before that date.

New Features

• [context-registry] Add tabular context registry package.
• [core] Add gemini-3.8-flash to the model types.
• [core] Add gpt-6-astra, gpt-6-luna, gpt-6.1-sol, gemini-3.6-flash, and nvidia--llama-nemotron-embed-1b to the model types.
• [langchain] Add reasoning content support to the Orchestration LangChain client. Reasoning blocks from model responses are surfaced in AIMessage.additional_kwargs.reasoning_content(preserving the signature for encrypted thinking), for both non-streaming and streaming responses. On subsequent turns, reasoning is forwarded back to the Orchestration service in reasoning_content from additional_kwargs.reasoning_content. Reasoning placed directly in message content is treated as output only and is not forwarded.
• [orchestration] Update generated client to latest spec (rel-1.13.2). Add optional provider_specific_fields property to LlmChoice and LlmChoiceStreaming response types.
• [orchestration] Update the generated client to spec v1.23. Adds the IBM Granite Guardian 4.1 content filter. Adds input filter target selectors. Adds the drop_unsupported_params option on LLMModelDetails.
• [prompt-registry] Update generated client to latest spec. Add deleteOrchestrationConfigsByScenario, deleteOrchestrationConfigHistory, deletePromptTemplatesByScenario, and deletePromptTemplateHistory API functions.
• [tabular-orchestration] Add experimental Tabular AI Orchestration package with TabularOrchestrationClient for predictions with Tabular Foundation Models.

Fixed Issues

• [langchain] Additional additional_kwargs properties (including tool_calls, intermediate_results, and reasoning_content) are now correctly placed on the AIMessage instead of the generation object.
• [langchain] Avoid retrying irrecoverable orchestration request and unsupported-model errors.
• [orchestration] Preserve prompt_tokens_details and completion_tokens_details in streamed token usage to fix missing cache and reasoning token details in OrchestrationStreamResponse.getTokenUsage() and getIntermediateResults().
• [orchestration] Fix missing prompt template or messages not being detected when no prompt is configured.

Improvements

• [core] Remove deprecated models cohere--command-a-reasoning (retirement date: 2026-10-30), mistralai--mistral-medium-instruct (retirement date: 2026-11-30) — use mistralai--mistral-medium instead, gemini-2.5-flash-lite (retirement date: 2026-10-20) — use gemini-3.5-flash-lite instead, sap-rpt-1.5 (retirement date: 2026-12-31) — use sap-rpt-1.6 instead, sap-rpt-1.5-large (retirement date: 2026-12-31) — use sap-rpt-1.6-large instead, sap-rpt-1-small (retirement date: 2026-12-31) — use sap-rpt-1.6 instead and sap-rpt-1-large (retirement date: 2026-12-31) — use sap-rpt-1.6-large instead.

Changelog

Sourced from @​sap-ai-sdk/orchestration's changelog.

2.17.0

Minor Changes

• [feat] Update generated client to latest spec (rel-1.13.2). Add optional provider_specific_fields property to LlmChoice and LlmChoiceStreaming response types.

• [compat] LlamaGuard38B and LlamaGuard38BFilterConfig are now marked as deprecated and will be removed on 2027-09-20. Migrate usages before that date.

• [feat] Update the generated client to spec v1.23. Adds the IBM Granite Guardian 4.1 content filter. Adds input filter target selectors. Adds the drop_unsupported_params option on LLMModelDetails.

Patch Changes

• [fix] Preserve prompt_tokens_details and completion_tokens_details in streamed token usage to fix missing cache and reasoning token details in OrchestrationStreamResponse.getTokenUsage() and getIntermediateResults().

• [fix] Fix missing prompt template or messages not being detected when no prompt is configured.

Commits

• b225ff6 v2.17.0
• 60fd777 chore(deps): update vitest monorepo to v5.0.5 (#2470)
• e041eaa feat: Update orchestration specification (#2446)
• c08bf98 chore(deps): update vitest monorepo to v5.0.4 (#2447)
• 731757a chore: enable oxlint correctness-category lint warnings (#2380)
• 0ac1b01 chore(deps): update vitest monorepo to v5.0.3 (#2391)
• 6e2857d chore: Use pnpm 11 catalogs to force certain dependencies on common v (#2129)
• 57111ad chore: Also mark Llama Guard filter types and function as deprecated (#2376)
• 7ffb0b9 feat: Update orchestration specification (#2375)
• f3a3449 fix: extract prompt merging into dedicated function (#2310)
• Additional commits viewable in compare view

Updates deepagents from 1.14.1 to 1.15.1

Release notes

Sourced from deepagents's releases.

deepagents@1.15.1

Patch Changes

• #966 0e4d3ee Thanks @​casparb! - ContextHubBackend.glob now only returns files under its path argument.

deepagents@1.15.0

Minor Changes

• #932 389e8da Thanks @​richardscarrott! - feat(deepagents): pin skills by name via pinnedSkills

  Pass skill names in pinnedSkills to guarantee the model gets those skills' instructions, for example when a user names a skill with a /skill: command or a picker. Before the next model call, the skills middleware appends each named skill's SKILL.md, without its frontmatter, as its own HumanMessage marked with additional_kwargs.lc_source: "pinned_skill" and the skill's name, path and description. The middleware clears the key once it uses it, and skips an unknown or unreadable skill. A tool can pin skills by returning Command({ update: { pinnedSkills } }), and a pinned skill discloses its metadata.include_tools the same way reading it does. A custom middleware can declare the key on its state schema with the exported pinnedSkillsValue.

• #929 9333679 Thanks @​richardscarrott! - feat(deepagents): load a skill's tools only when the skill is read

  Skills can list tools in metadata.include_tools. Pass those tools to createSkillsMiddleware as tools (an array or a resolver), and pass that middleware to createDeepAgent. The model can't see or call a skill tool until it reads a skill that lists it. On models that accept tool definitions mid-conversation this keeps the prompt cache intact, and needs @langchain/anthropic >= 1.5.12 or @langchain/openai >= 1.6.2.

  The skills middleware now sits just before prompt caching, after custom and harness-profile middleware.

Patch Changes

• #919 b448f22 Thanks @​thushanth-bengre-langchain! - Fix CompositeBackend matching routes by bare string prefix: a route /foo no longer captures /foobar.txt. Route prefixes are now normalized to a trailing slash, so routePrefixes returns /foo/, and /foo and /foo/ register as the same route.

• #928 5f42f2d Thanks @​thushanth-bengre-langchain! - Check filesystem permissions against the path an operation actually touches: rule matching drops . segments, and with FilesystemBackend (directly or behind a CompositeBackend route) rules are also checked on the symlink-resolved target. A deny on either path blocks the call, so allow rules must cover a symlink's target (including OS-level links like macOS /tmp → /private/tmp in non-virtual mode).

• #929 9333679 Thanks @​richardscarrott! - Mark failed tool results with status: "error". read_file returns a backend failure, such as a missing file, as an error ToolMessage instead of plain Error: … text. The patch tool calls middleware also marks the result it inserts for a tool call that never ran as an error.

deepagents@1.14.2

Patch Changes

• #877 9189214 Thanks @​thushanth-bengre-langchain! - feat(deepagents): add offloadBinaryContent option to FilesystemMiddleware to keep binary read_file content and inline HumanMessage media out of checkpointed message history. When enabled, payloads are written to the backend under /blobs and replaced in state with a small reference; model requests are rehydrated from a per-run cache or the backend.

• #881 188c2b5 Thanks @​thushanth-bengre-langchain! - Add UnsupportedContentMiddleware, which replaces input content the active model can't accept with a text placeholder before each model call, so a provider rejection doesn't fail every later turn too. createDeepAgent and subagents add it automatically, last in the middleware stack. File support is read from model.profile.fileMimeTypes (requires @langchain/core@^1.2.14, and @langchain/openai@^1.6.1 for OpenAI/Azure models).

  FilesystemMiddleware also now recovers from a rejected read_file result that filtering missed: it retries once with the content replaced, and persists the swap to state.

Commits

• 5864b06 chore: version packages (#967)
• 0e4d3ee fix(deepagents): make ContextHubBackend.glob honor path (#966)
• 9f3bfd5 chore: version packages (#926)
• 289ebba fix(quickjs): classify subagent concurrency changeset as patch (#958)
• 8b1c1de feat(quickjs): make subagent concurrency configurable (#957)
• 5f42f2d fix(deepagents): check filesystem permissions against the symlink-resolved pa...
• 02ff0f5 fix: patch axios and grpc security vulnerabilities (#952)
• 389e8da feat(deepagents): pin skills by name via pinnedSkills (#932)
• 9333679 feat(deepagents): load a skill's tools only when the skill is read (#929)
• b448f22 fix(deepagents): match CompositeBackend routes on path-segment boundaries (#919)
• Additional commits viewable in compare view

Updates marked from 18.0.14 to 18.1.1

Release notes

Sourced from marked's releases.

v18.1.1

18.1.1 (2026-10-12)

Bug Fixes

• cli: honor an explicitly empty string input (#4113) (ec3a4e5)
• end blockquote before a lazy line after a heading, hr or html block (#4115) (9e64a6a)
• keep a backslash in a code span from hiding emphasis (#4117) (6a57018)
• keep a decoded character reference from forming a new one (#4111) (6bbac0c), closes #38 #112
• let indented and type 3-5 html blocks interrupt a paragraph (#4124) (eaa1fff)
• make walkTokens linear by appending values in place (#4121) (627d9ae), closes #4120
• match pre, script, style and textarea as whole tag names (#4114) (0db74cb)
• preserve enclosing link state after parsing images (#4123) (cc07bb9)
• respect tab stops in list continuation indentation (#4119) (4bd2528)

Performance Improvements

• avoid repeated raw string copies in nested lists (#4082) (7e8754d)

v18.1.0

18.1.0 (2026-10-05)

Bug Fixes

• avoid cubic backtracking on unicode whitespace in a link destination (#4106) (0d20220)
• load CLI configs with top-level await (#4100) (c61543e)
• reject unbalanced parentheses in link destinations (#4107) (3363c99)

Features

• add linkParenPossible to lexer state to fail fast for link token generation (#4070) (4ee44f7)

Commits

• cdbdc01 chore(release): 18.1.1 [skip ci]
• 2134976 docs: add AGENTS.md (#4125)
• 699a969 chore(deps-dev): bump handlebars from 4.7.9 to 4.7.10 (#4133)
• 6a57018 fix: keep a backslash in a code span from hiding emphasis (#4117)
• eaa1fff fix: let indented and type 3-5 html blocks interrupt a paragraph (#4124)
• 7e8754d perf: avoid repeated raw string copies in nested lists (#4082)
• cc07bb9 fix: preserve enclosing link state after parsing images (#4123)
• 627d9ae fix: make walkTokens linear by appending values in place (#4121)
• 4bd2528 fix: respect tab stops in list continuation indentation (#4119)
• 9e64a6a fix: end blockquote before a lazy line after a heading, hr or html block (#4115)
• Additional commits viewable in compare view

Updates zod from 4.6.5 to 4.6.7

Release notes

Sourced from zod's releases.

v4.6.7

Commits:

• b500a2ee95134f81d74da7911d7378dd2853d042 4.6.7

v4.6.6

Six malicious commits pushed to main on September 21 with a compromised git credential were reverted the same day in #6620 and have since been removed from the history of main. None were ever published to npm. The npm provenance for 4.6.6 names 284243fb, the pre-rewrite copy of 9c219833 with identical contents.

Commits:

• 0cd3af5e9f50ad11bbcd4407a46a21b8848ba574 docs: load the en locale directly in Zod Mini
• 9d5e60f162997b764b40d45aa02481888db50dea docs: import locales from zod/locales, not z.locales
• 34a626f307871ecbb0123947559328dc52ed5e3e docs: tighten the z.locales tree-shaking note
• 74304ce5b5796b5edb9a88ceb8e170b6ea3e82df release: stage npm versions instead of publishing them (#6622)
• b2ef983e0711ff6517d6020b021071e2e5653efa docs: add Code for Japan to bronze sponsors
• 8fb2faf0584828a6a027d8279db36f6f52784462 docs: crop the Code for Japan logo and list it first in bronze
• bacad518c41610d654658172d325874bf0ee6290 docs: square the Code for Japan logo to match the other bronze tiles
• c42596bd1c87e941d7241d10e012c5c196adcc10 release: hold the run on the release-approval environment until the staged versions are approved (#6638)
• a2ba4a67986a3e6b2b131be30084a2f460d9ea67 fix(types): declare record output and input as members, not base type arguments (#6645)
• a0ebf970f2d752e5f58527cd2942ff48be4f46bd fix(v4): check array, set and map size bounds before parsing elements (#6652)
• 529f78dd9431b07f49647fee3aa0819b3f8b66b7 4.6.6
• 9c219833c3a3764fa1f1ea9c5ecec8254a63c0b1 ci: stage at verbose so a failed OIDC token exchange says why

Commits

• b500a2e 4.6.7
• 9c21983 ci: stage at verbose so a failed OIDC token exchange says why
• 529f78d 4.6.6
• a0ebf97 fix(v4): check array, set and map size bounds before parsing elements (#6652)
• a2ba4a6 fix(types): declare record output and input as members, not base type argumen...
• c42596b release: hold the run on the release-approval environment until the staged ve...
• bacad51 docs: square the Code for Japan logo to match the other bronze tiles
• 8fb2faf docs: crop the Code for Japan logo and list it first in bronze
• b2ef983 docs: add Code for Japan to bronze sponsors
• 74304ce release: stage npm versions instead of publishing them (#6622)
• Additional commits viewable in compare view

Upda...

Description has been truncated

[Akatuoro]

@dependabot ignore @opentelemetry/exporter-trace-otlp-proto minor version

[Akatuoro]

@opentelemetry/exporter-trace-otlp-proto 0.300.0 is based on opentelemetry 3, see cap-js/telemetry#556

Also: https://github.com/open-telemetry/opentelemetry-js/blob/main/doc/3.x/migration-guide.md