build(deps): Bump the npm-dependencies group across 1 directory with 5 updates

[dependabot]

Bumps the npm-dependencies group with 5 updates in the /actions/release/summarize-changelog directory:

Package	From	To
@langchain/anthropic	1.5.5	1.5.8
@langchain/google-genai	2.2.0	2.3.0
@langchain/openai	1.5.7	1.5.10
html-to-text	10.0.0	10.0.1
langchain	1.5.8	1.5.10

Updates @langchain/anthropic from 1.5.5 to 1.5.8

Release notes

Sourced from @​langchain/anthropic's releases.

@​langchain/anthropic@​1.5.8

Patch Changes

• #11421 041a755 Thanks @​priprocess! - Preserve the generic tool_search_tool_result block type in streaming output and message payload conversion. The Anthropic API emits this un-suffixed type for both tool search variants; previously only the variant-suffixed names were allowlisted, so multi-turn conversations using tool search still failed with INVALID_TOOL_RESULTS after a client-tool round-trip.

@​langchain/anthropic@​1.5.7

Patch Changes

• #11407 3ceef4b Thanks @​chenzimin! - Round-trip the tool search server-tool result blocks (tool_search_tool_bm25_tool_result / tool_search_tool_regex_tool_result) through message translation, mirroring web_search_tool_result. Without this, a multi-turn conversation using the tool search tool fails on the next request with 400 invalid_request_error … tool use … found without a corresponding … tool_result block, because the result block was dropped from the assistant message (streaming capture) and on re-send (toolTypes passthrough).

@​langchain/anthropic@​1.5.6

Patch Changes

• #11342 3b0e4c4 Thanks @​thushanth-bengre-langchain! - feat(anthropic): mark Anthropic provider errors as retryable or not

  Builds on stampRetryable in @langchain/core so the retry middleware can tell a transient failure from a deterministic one. Rate limits are marked retryable; context overflow, invalid tool results, bad credentials, and unknown models non-retryable. Anything else stays unmarked and retries as before.

  Also forwards a per-call maxRetries to the retry loop, so a surrounding retry loop such as modelRetryMiddleware can take over instead of the two multiplying against each other.

  Errors keep their original class, so instanceof against the @anthropic-ai/sdk error types is unaffected.

Commits

• d5264a1 chore: version packages (#11427)
• c26c87e fix(openai): send content null (not []) for tool-call-only v1 assistant messa...
• 041a755 fix(anthropic): preserve generic tool_search_tool_result blocks (#11421)
• e493ed6 chore: version packages (#11393)
• 8384848 fix(google-common): release endpoint routing fix as patch (#11413)
• 8cfff4d feat(google): add gateway support for genai (#11405)
• 7df258c chore(langchain): update langgraph deps (#11412)
• 3ceef4b fix(anthropic): round-trip tool search server-tool result blocks (#11407)
• fe8eec1 fix(openai): drop Gemini functionCall content blocks in Chat Completions mess...
• 0e7c765 fix(google-genai): throw ContentBlockedError when Gemini candidate has no con...
• Additional commits viewable in compare view

Updates @langchain/google-genai from 2.2.0 to 2.3.0

Release notes

Sourced from @​langchain/google-genai's releases.

@​langchain/google-genai@​2.3.0

Minor Changes

• #11400 0e7c765 Thanks @​thushanth-bengre-langchain! - fix(google-genai): throw a typed, catchable EmptyContentError instead of silently returning empty content / an empty generations array when a non-streaming call (.invoke(), .generate(), .batch()) gets back a Gemini candidate with no content — whether from an explicit block (safety/recitation filter, rejected prompt) or the model just not producing usable output (a malformed function call, a thinking model exhausting its token budget). The streaming path (.stream()) is unaffected by this behavior change; it now safely skips a contentless chunk instead of throwing a raw TypeError.

Patch Changes

• #11405 8cfff4d Thanks @​hntrl! - Add LangSmith gateway support for Google Gemini (Developer API) models. When LANGSMITH_GATEWAY is set, ChatGoogleGenerativeAI and ChatGoogle (and initChatModel("google-genai:...")) route requests through the gateway's Gemini path, using the gateway key (falling back to LANGSMITH_API_KEY). An explicit baseUrl/endpoint, an explicit apiKey, or a Vertex AI configuration suppress gateway routing. Also adds GEMINI_API_KEY as a fallback env var for ChatGoogleGenerativeAI.

Commits

• e493ed6 chore: version packages (#11393)
• 8384848 fix(google-common): release endpoint routing fix as patch (#11413)
• 8cfff4d feat(google): add gateway support for genai (#11405)
• 7df258c chore(langchain): update langgraph deps (#11412)
• 3ceef4b fix(anthropic): round-trip tool search server-tool result blocks (#11407)
• fe8eec1 fix(openai): drop Gemini functionCall content blocks in Chat Completions mess...
• 0e7c765 fix(google-genai): throw ContentBlockedError when Gemini candidate has no con...
• 5c9fdf2 fix(openai): retain cache_write_tokens, update to v7 sdk (#11399)
• 5ff9179 fix(google-genai): guard streaming chunks when candidate has no content (#10742)
• 43e4396 fix(core): include tool_call blocks and skip empty text blocks in ChatVertexA...
• Additional commits viewable in compare view

Updates @langchain/openai from 1.5.7 to 1.5.10

Release notes

Sourced from @​langchain/openai's releases.

@​langchain/openai@​1.5.10

Patch Changes

• #11419 c26c87e Thanks @​chiliec! - fix(openai): send content null (not []) for tool-call-only v1 assistant messages

@​langchain/openai@​1.5.9

Patch Changes

• #11399 5c9fdf2 Thanks @​gethin-langchain! - update to v7 openai sdk

• #11403 fe8eec1 Thanks @​thushanth-bengre-langchain! - Drop Gemini-native functionCall content blocks (already carried in tool_calls) when converting messages to Chat Completions API params, fixing requests that fail when a ChatGoogleGenerativeAI message is passed to ChatOpenAI (e.g. a cross-provider handoff in LangGraph).

• #11399 5c9fdf2 Thanks @​gethin-langchain! - Map OpenAI's cache_write_tokens to cache_creation in usage_metadata.input_token_details, mirroring the existing cached_tokens -> cache_read mapping across the Chat Completions and Responses APIs. Previously, prompt cache-write token counts were silently dropped.

@​langchain/openai@​1.5.8

Patch Changes

• #11342 3b0e4c4 Thanks @​thushanth-bengre-langchain! - feat(openai): mark OpenAI provider errors as retryable or not

  Builds on stampRetryable in @langchain/core so the retry middleware can tell a transient failure from a deterministic one. Timeouts and rate limits are marked retryable; aborts, context overflow, invalid tool results, bad credentials, and unknown models non-retryable. Anything else stays unmarked and retries as before.

  Also forwards a per-call maxRetries to the retry loop, so a surrounding retry loop such as modelRetryMiddleware can take over instead of the two multiplying against each other.

  Errors keep their original class, so instanceof against the openai SDK error types is unaffected.

Commits

• d5264a1 chore: version packages (#11427)
• c26c87e fix(openai): send content null (not []) for tool-call-only v1 assistant messa...
• 041a755 fix(anthropic): preserve generic tool_search_tool_result blocks (#11421)
• e493ed6 chore: version packages (#11393)
• 8384848 fix(google-common): release endpoint routing fix as patch (#11413)
• 8cfff4d feat(google): add gateway support for genai (#11405)
• 7df258c chore(langchain): update langgraph deps (#11412)
• 3ceef4b fix(anthropic): round-trip tool search server-tool result blocks (#11407)
• fe8eec1 fix(openai): drop Gemini functionCall content blocks in Chat Completions mess...
• 0e7c765 fix(google-genai): throw ContentBlockedError when Gemini candidate has no con...
• Additional commits viewable in compare view

Updates html-to-text from 10.0.0 to 10.0.1

Commits

• 1c39d98 html-to-text v10.0.1 cli v0.6.1
• bcc3d81 bump deepmerge-ts and some dev dependencies
• 59d1d3e replaced write-file-action with Node script
• 51adfec split GitHub workflows for push and PR to keep badges consistent
• See full diff in compare view

Updates langchain from 1.5.8 to 1.5.10

Release notes

Sourced from langchain's releases.

langchain@1.5.10

Patch Changes

• #11412 7df258c Thanks @​hntrl! - chore(langgraph): update langgraph deps to track serialization fix

langchain@1.5.9

Patch Changes

• #11369 d6ad973 Thanks @​hntrl! - fix(langchain): use unified endpoint for gateway

• #11342 3b0e4c4 Thanks @​thushanth-bengre-langchain! - feat(core): mark errors as retryable or not, and stop retrying the ones that aren't

  Retry middleware retried every failure up to maxRetries, including deterministic ones like a bad API key or an unknown model. Retries also nest, so a single such failure could cost dozens of API calls.

  @langchain/core/errors adds stampRetryable(error, retryable) and getRetryable(error). Marking an error leaves its class and shape untouched, so a provider SDK error can be classified without breaking instanceof. getRetryable returns undefined for errors nobody classified, and both are exported so tool authors can mark their own failures.

  modelRetryMiddleware and toolRetryMiddleware now respect the mark by default, and retries stop as soon as one is found rather than each layer spending its own budget. Aborted calls, context overflow, and oversized payloads are marked non-retryable out of the box. Models accept a per-call maxRetries so a surrounding retry loop can take over.

  Behavior change: errors marked non-retryable now fail on the first attempt. Unclassified errors — including any from third-party integrations or custom tools — retry exactly as before. Pass retryOn: () => true to restore the old default. A custom onFailedAttempt replaces the built-in handler and opts out of marking.

Commits

• e493ed6 chore: version packages (#11393)
• 8384848 fix(google-common): release endpoint routing fix as patch (#11413)
• 8cfff4d feat(google): add gateway support for genai (#11405)
• 7df258c chore(langchain): update langgraph deps (#11412)
• 3ceef4b fix(anthropic): round-trip tool search server-tool result blocks (#11407)
• fe8eec1 fix(openai): drop Gemini functionCall content blocks in Chat Completions mess...
• 0e7c765 fix(google-genai): throw ContentBlockedError when Gemini candidate has no con...
• 5c9fdf2 fix(openai): retain cache_write_tokens, update to v7 sdk (#11399)
• 5ff9179 fix(google-genai): guard streaming chunks when candidate has no content (#10742)
• 43e4396 fix(core): include tool_call blocks and skip empty text blocks in ChatVertexA...
• Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore <dependency name> major version will close this group update PR and stop Dependabot creating any more for the specific dependency's major version (unless you unignore this specific dependency's major version or upgrade to it yourself)
• @dependabot ignore <dependency name> minor version will close this group update PR and stop Dependabot creating any more for the specific dependency's minor version (unless you unignore this specific dependency's minor version or upgrade to it yourself)
• @dependabot ignore <dependency name> will close this group update PR and stop Dependabot creating any more for the specific dependency (unless you unignore this specific dependency or upgrade to it yourself)
• @dependabot unignore <dependency name> will remove all of the ignore conditions of the specified dependency
• @dependabot unignore <dependency name> <ignore condition> will remove the ignore condition of the specified dependency and ignore conditions

[github-actions]

Hi, thank you for creating your PR, we will check it out very soon

[github-actions]

Super-linter summary

Language	Validation result
BIOME_FORMAT	Pass ✅
BIOME_LINT	Pass ✅
CHECKOV	Pass ✅
GITLEAKS	Pass ✅
GIT_MERGE_CONFLICT_MARKERS	Pass ✅
JSCPD	Pass ✅
PRE_COMMIT	Pass ✅
SPELL_CODESPELL	Pass ✅
TRIVY	Pass ✅

All files and directories linted successfully

For more information, see the GitHub Actions workflow run

Powered by Super-linter