Who contributed to AI Agents on August 7, 2026?

9 developers shipped this update, including langchain-oss-model-profiles, imnishitha, langchain-oss-model-profiles[bot], FraterCCCLXIII, hieptl, DevinVinson, github-actions[bot], and abeatrix, and 1 more.

What were the notable AI Agents updates?

feat(langchain): add `state_schema` param to `wrap_tool_call`, chore(model-profiles): refresh model profile data (#39299), and fix(langchain): restrict narrowed `ToolStrategy` in bound tools.

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 TIGHTENS TOOL MIDDLEWARE, CLINE SHIPS REASONING UNIFICATION

By RepoJournal · Filed 07:01 UTC on August 7, 2026 · About AI Agents

9 people shipped this

imnishitha @imnishitha 2 cited

hieptl @hieptl 2 cited

github-actions[bot] @github-actions[bot] 2 cited

abeatrix @abeatrix 2 cited

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

FraterCCCLXIII @FraterCCCLXIII 1 cited

saoudrizwan @saoudrizwan 1 cited

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

DevinVinson @DevinVinson

LangChain adds missing state_schema to wrap_tool_call for consistency across all middleware decorators, while Cline v3.0.51 resolves reasoning settings portably across AI SDK providers instead of per-provider thinking overrides.

LangChain closed a gap in its middleware decorator API by adding the `state_schema` parameter to `wrap_tool_call` [1], bringing it into parity with `before_model`, `after_model`, `wrap_model_call`, `before_agent`, and `after_agent`. In parallel, a critical fix restricts narrowed `ToolStrategy` in bound tools [3], ensuring that when middleware overrides narrow a union of structured-output schemas, the model can only choose from the effective subset, not the original full set. Model profile data refreshed across three providers [2], removing the dated `claude-opus-4-1-20250805` variant from Anthropic's roster. On the OpenHands front, the desktop app gained per-run LLM cost tracking in the Activity Log and exports [6], shipped multi-size icons for Windows and macOS [5], and rebranded to OpenHands Agent Canvas with UI polish [4]. Cline v3.0.51 and SDK v0.0.71 unified reasoning effort handling so that settings "now resolve portably across AI SDK providers" [7], eliminating fragmented per-provider thinking overrides and ensuring explicit disable requests take priority everywhere, including Ollama. The desktop app also landed a suite of polish fixes addressing silent turn failures, message duplication, and stuck composer, plus first-run setup guidance [10]. MCP gains an uninstall command [9] and OAuth support for pre-registered remote clients [11].

Action items

→ Review wrap_tool_call usage in agents relying on state_schema for middleware consistency [ref:2] langchain-ai/langchain [plan]
→ Upgrade Cline CLI to v3.0.51 and SDK to v0.0.71 to unify reasoning settings across all providers [ref:11, ref:12] cline/cline [monitor]
→ Test OpenHands desktop app first-run flow and cost tracking in Activity Log [ref:9, ref:14] OpenHands/OpenHands [monitor]

References

[1] feat(langchain): add `state_schema` param to `wrap_tool_call` ↗ langchain-ai/langchain
[2] chore(model-profiles): refresh model profile data (#39299) langchain-ai/langchain
[3] fix(langchain): restrict narrowed `ToolStrategy` in bound tools ↗ langchain-ai/langchain
[4] feat(settings): polish Agent Canvas version update UI ↗ OpenHands/OpenHands
[5] fix: ship multi-size app icons for Windows and macOS ↗ OpenHands/OpenHands
[6] feat: show per-run LLM cost in the Activity Log and exports ↗ OpenHands/OpenHands
[7] CLI v3.0.51 ↗ cline/cline
[8] SDK v0.0.71 ↗ cline/cline
[9] feat(cli): add mcp uninstall command ↗ cline/cline
[10] desktop: fix silent turn failures, message duplication, and stuck composer; add first-run setup guidance ↗ cline/cline
[11] feat(core): support pre-registered OAuth clients for remote MCP ↗ cline/cline

Quick answers

What shipped in AI Agents on August 7, 2026?: LangChain adds missing state_schema to wrap_tool_call for consistency across all middleware decorators, while Cline v3.0.51 resolves reasoning settings portably across AI SDK providers instead of per-provider thinking overrides. In total, 20 commits, 16 pull requests, and 3 releases landed.
Who contributed to AI Agents on August 7, 2026?: 9 developers shipped this update, including langchain-oss-model-profiles, imnishitha, langchain-oss-model-profiles[bot], FraterCCCLXIII, hieptl, DevinVinson, github-actions[bot], and abeatrix, and 1 more.
What were the notable AI Agents updates?: feat(langchain): add `state_schema` param to `wrap_tool_call`, chore(model-profiles): refresh model profile data (#39299), and fix(langchain): restrict narrowed `ToolStrategy` in bound tools.

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

AI Agents Aug 6

LANGCHAIN HARDENS ANTHROPIC TOOL CALLING AND GIT CREDENTIALS AS OPENHANDS SHIPS 1.10.0

LangChain fixed a critical credential leak in evaluation runs while Anthropic tool schemas get proper fallback handling, as three major agent frameworks push releases across the ecosystem.

ai-agents 54 shipped

AI Agents Aug 5

HITL APPROVAL GATES STOP SILENTLY FAILING OPEN, AIOHTTP FIXES REDIRECT LEAKS ACROSS LANGCHAIN STACK

LangChain plugged a critical security hole in human-in-the-loop middleware where invalid configurations silently disabled approval gates, while simultaneously patching a cascade of aiohttp vulnerabilities that leak headers across origin boundaries.

ai-agents 121 shipped

AI Agents Aug 4

LANGCHAIN PLUGS RESOURCE LEAK, CLINE ADDS CONTEXT-WINDOW RECOVERY

LangChain fixed a critical event-loop leak in streaming tracers that was breaking Python 3.14+ tests, while Cline shipped context-window overflow detection and recovery across its SDK.

+10

ai-agents 73 shipped

Elsewhere on the wire

Local LLMs Aug 8

OLLAMA SHIPS BLOCK-DIFFUSION SPECULATIVE DECODING, SGLANG LAUNCHES KIMI K3 DAY-0

Draft models just got faster: Ollama merged DFlash block-diffusion speculative decoding that proposes entire token sequences in one forward pass, while SGLang shipped production-ready support for Kimi K3, a 2.8T-parameter multimodal LatentMoE that routes through a latent space with 896 experts.

+12

llms 263 shipped

@CachyOS Aug 8

CACHYOS AUTOMATION PUSHES BROWSER UPDATES ACROSS THE STACK

Five browser and utility packages shipped overnight with automated version bumps, while the Welcome interface gains Traditional Chinese support and translation cleanup hits the queue.

infra 51 shipped

Elixir & Phoenix Aug 8

ELIXIR PARSER GETS LEANER AS PHOENIX AUDITS DEAD CODE

Daniel Kukula shipped three consecutive optimizations to Elixir's date parsing pipeline, removing accumulators and intermediate allocations in a focused performance push [ref:1] [ref:2] [ref:3].

elixir 14 shipped

@dotnet Aug 8

MAUI LOCKS IN NET11.0, RUNTIME UNIFIES STUB MEMORY

MAUI pivots its production CI to .NET 11.0 while Runtime consolidates stub allocation with JIT infrastructure, eliminating a parallel memory path that's leaked 15 years of technical debt.