What were the notable AI Agents updates?

langchain-openai==1.4.3, chore(model-profiles): refresh model profile data (#39430), and fix(automations): surface non-MCP-installable integrations instead of silently dropping them.

AI Agents

LangChain, OpenHands, Cline, and LlamaIndex - agent frameworks and coding agents

Pick a date

Topics: AI Agents AI / ML Full archive →

The Wire · Showcase

LANGCHAIN PATCHES TOOL CALLS, OPENHANDS SURFACES HIDDEN INTEGRATIONS

By RepoJournal · Filed 07:02 UTC on August 11, 2026 · About AI Agents

11 people shipped this

mkondratek @mkondratek 2 cited

github-actions[bot] @github-actions[bot] 1 cited

langchain-oss-model-profiles[bot] @langchain-oss-model-profiles[bot] 1 cited

lokesh75-kank @lokesh75-kank 1 cited

Lothnic @Lothnic 1 cited

georgeglarson @georgeglarson 1 cited

jsimone @jsimone 1 cited

langchain-oss-model-profiles @langchain-oss-model-profiles

ccurme @ccurme

tofarr @tofarr

dependabot @dependabot

LangChain's OpenAI wrapper shipped a fix for invalid tool calls that were silently dropped, while OpenHands surfaced integrations the automation layer was burying.

LangChain released langchain-openai 1.4.3 [1] with a critical fix that filters invalid tool calls from content, addressing a silent failure mode where malformed tool responses would disappear without warning. The release also updates guidance for responses API handling across OpenAI-compatible providers [1]. Model profile data refreshed across OpenRouter integrations [2], with Google's Gemma 4 26B now supporting 262,144 max output tokens, a 16x increase from prior limits. OpenHands landed a fix [3] that surfaces non-MCP-installable integrations instead of silently dropping them in the automation marketplace, a regression that had blocked Jira and other external tools from appearing in templates. The team also refactored conversation creation parameters into options objects [4], pushing type safety deeper into the SDK. Context window usage tracking landed in OpenHands [5] with a manual compaction drawer, tested live against agent-server 1.39.1. Cline added Fable 5 support to Vertex and enabled custom model IDs with user-editable metadata [6], while fixing a stale feature tip that directed users to a non-existent Double-Check Completion setting [7]. Auto-compaction now respects user max output tokens in summarizer requests [8], fixing failures on local reasoning models like Qwen 3.x.

Action items

→ Upgrade langchain-openai to 1.4.3 if using tool calling langchain-ai/langchain [immediate]
→ Review OpenHands automations if integrations disappeared from marketplace OpenHands/OpenHands [plan]
→ Update Cline to latest for Fable 5 and compaction fixes cline/cline [plan]

References

[1] langchain-openai==1.4.3 ↗ langchain-ai/langchain
[2] chore(model-profiles): refresh model profile data (#39430) langchain-ai/langchain
[3] fix(automations): surface non-MCP-installable integrations instead of silently dropping them ↗ OpenHands/OpenHands
[4] refactor: replace positional createConversation params with an options object (#1587) ↗ OpenHands/OpenHands
[5] feat: context window usage meter, usage drawer, and manual compaction ↗ OpenHands/OpenHands
[6] feat(vscode): add Fable 5 + custom model IDs to Vertex; drop global-region picker filter ↗ cline/cline
[7] fix: remove stale Double-Check Completion feature tip ↗ cline/cline
[8] fix: respect user max output tokens in compaction summarizer requests ↗ cline/cline

Quick answers

What shipped in AI Agents on August 11, 2026?: LangChain's OpenAI wrapper shipped a fix for invalid tool calls that were silently dropped, while OpenHands surfaced integrations the automation layer was burying. In total, 17 commits, 25 pull requests, and 2 releases landed.
Who contributed to AI Agents on August 11, 2026?: 11 developers shipped this update, including github-actions[bot], langchain-oss-model-profiles[bot], langchain-oss-model-profiles, ccurme, lokesh75-kank, Lothnic, georgeglarson, and tofarr, and 3 more.
What were the notable AI Agents updates?: langchain-openai==1.4.3, chore(model-profiles): refresh model profile data (#39430), and fix(automations): surface non-MCP-installable integrations instead of silently dropping them.

LANGCHAIN SHIPS THREE CRITICAL RELEASES IN 12 HOURS, CLOSES SETUPTOOLS VULNERABILITY

LangChain pushed langchain-core 1.5.4, langchain 1.3.15, and langchain-anthropic 1.5.5 overnight while patching a setuptools supply-chain risk that gated the Hugging Face integration.

ai-agents 59 shipped

AI Agents Aug 10

LANGCHAIN PATCHES PYDANTIC 2.14 BREAKS, CLINE KILLS YOLO MODE

LangChain shipped critical compatibility fixes for Pydantic 2.14 serialization bugs while Cline removed a cosmetic toggle that masked real approval confusion.

ai-agents 10 shipped

AI Agents Aug 9

CLINE FIXES ABORT-WINDOW PROMPT LOSS, CLIPBOARD IMAGES LAND ON DESKTOP

Cline's core agent now preserves queued prompts when you hit stop, and desktop users can finally paste screenshots directly into the composer.

ai-agents 40 shipped

AI Agents Aug 8

LANGCHAIN PATCHES CLAUDE FILE TOOLS AND OPENAI CONTEXT ERRORS

LangChain shipped langchain-openai 1.4.2 with fixes for context window overflows and broken Claude rename operations, while OpenHands v1.11.0 adds per-run LLM cost tracking.

ai-agents 99 shipped

Elsewhere on the wire

Local LLMs Aug 12

OLLAMA CUTS REPEAT PENALTY BLOAT, LLAMA.CPP SHIPS POCKET-TTS SUPPORT

Ollama kills a decade of default penalty cruft that was drowning every model, while llama.cpp gains multimodal audio chops and vLLM fixes ROCm speculative decoding corruption.

+11

llms 258 shipped

@CachyOS Aug 12

CACHYOS KERNEL BUMPS TO 7.1.7 AS MIRROR NETWORK EXPANDS

CachyOS shipped kernel 7.1.7 overnight while extending its mirror footprint across Europe and adding Simplified Chinese to the wiki.

infra 60 shipped

Elixir & Phoenix Aug 12

ELIXIR TIGHTENS GUARD SYNTAX, PHOENIX FIXES MIX DETECTION

José Valim shut down a common footgun in Elixir guards while the team fixed a startup bug that breaks when mix is available but not yet started.

elixir 6 shipped

@dotnet Aug 12

BROWSER WASM BLOAT FIXED, RELOCATION SHRINKING SHIPS

dotnet/runtime eliminated unnecessary cryptography dependencies from browser builds while implementing code section relocation optimizations that shrink Wasm binaries.