The AI Toolchain — issue -222, January 7, 2026

Agno v2.3.23 adds async tool function support to Toolkit, automatically selected in async agent contexts.

$ git clone --branch v2.3.23 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v2.3.23

• ›Toolkit now supports async tool functions, automatically selected when the agent runs in an async context.

AutoGPT Platform v0.6.41 adds OAuth/SSO, Google Docs blocks, HITL redesign, and 256 MB file uploads.

$ git clone --branch autogpt-platform-beta-v0.6.41 https://github.com/Significant-Gravitas/AutoGPT.git
# already have the repo? check out this version:
$ git checkout autogpt-platform-beta-v0.6.41

• ›Adds GPT-5.2 as an available model and updates default model selections.
• ›Adds OAuth API and Single Sign-On (SSO) support for the platform.
• ›Adds Google Docs integration blocks for reading and writing documents.
• ›Adds auto-save Draft Recovery feature backed by IndexedDB persistence in the flow editor.
• ›Adds agent mode support to SmartDecisionMakerBlock with autonomous tool execution loops.

LangChain 1.2.1 adds Google GenAI embeddings support and enhanced init_chat_model validation.

$ git clone --branch langchain==1.2.1 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain==1.2.1

• ›Adds google_genai provider support to init_embeddings, enabling Google Generative AI embedding models to be initialized through the standard embeddings factory.
• ›Enhances init_chat_model with improved validation to catch misconfigured model parameters earlier.

PydanticAI v1.40.0 adds human-readable Temporal activity summaries and configurable retries for nonexistent tool calls.

$ git clone --branch v1.40.0 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v1.40.0

• ›Agents now retry calls to nonexistent tools up to the Agent retries limit, matching the existing retry behavior for real tools.
• ›Sets human-readable activity summaries for Temporal activities, improving observability in Temporal-based agent workflows.

Glamourous agentic coding for all

Crush v0.31.0 adds cross-tool skills loading from ~/.config/agents/skills and removes Claude Code support.

$ git clone --branch v0.31.0 https://github.com/charmbracelet/crush.git
# already have the repo? check out this version:
$ git checkout v0.31.0

• ›Supports loading skills from a shared ~/.config/agents/skills directory, enabling reuse across multiple agent tools.

GitHub Copilot CLI v0.0.375 adds Ctrl+T reasoning toggle and --share/--share-gist flags for session sharing

$ git clone --branch v0.0.375 https://github.com/github/copilot-cli.git
# already have the repo? check out this version:
$ git checkout v0.0.375

$ gh copilot suggest --share-gist 'Write a bash script to rotate AWS credentials'

• ›Adds --share and --share-gist flags for session sharing in non-interactive mode.
• ›Adds Ctrl+T keyboard shortcut to toggle reasoning summaries for supported models.
• ›SVG files are now treated as text files instead of binary images, enabling them to be read and edited as source.

Lightweight coding agent that runs in your terminal

Codex CLI gains multi-conversation agent control, thread rollback, cached web search, and an analytics config toggle.

$ git clone --branch rust-v0.79.0 https://github.com/openai/codex.git
# already have the repo? check out this version:
$ git checkout rust-v0.79.0

[analytics]
enabled = false

• ›Adds multi-conversation 'agent control' so a session can programmatically spawn or message other conversations.
• ›Adds thread/rollback app-server API so IDE clients can undo the last N turns without rewriting history.
• ›Adds web_search_cached flag to return cached-only Web Search results as a safer alternative to live requests.
• ›Allows global exec flags (model, json, sandbox toggles, etc.) to be passed after codex exec resume.
• ›Shows time/version-targeted announcement tips in the TUI, driven by a TOML file in the repo.

The open source coding agent.

OpenCode v1.1.6 adds theme support, subagent visibility controls, system CA flag, and tool truncation.

$ git clone --branch v1.1.6 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v1.1.6

• ›Adds --use-system-ca flag to support system certificate authorities in exec args.
• ›Adds truncation for all tools to prevent runaway output.
• ›Adds theme functionality to the TUI, with live reload when configuration changes.
• ›Adds ability to hide subagents from the primary agent's system prompt.
• ›Makes the General SubAgent visible in the agent list.

Qwen Code v0.6.1 graduates --acp to stable, adds German locale, Java SDK support, and direct /approval-mode arguments.

$ git clone --branch v0.6.1 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.6.1

$ qwen --acp

$ /approval-mode auto

• ›Graduates --experimental-acp to the stable --acp flag for Agent Communication Protocol support.
• ›Adds direct argument support for the /approval-mode command, allowing approval mode to be set inline.
• ›Adds German language support to the i18n system.
• ›Auto-detects LLM output language from the system locale.
• ›Adds Java SDK support (Feat/javasdk).

$ git clone --branch v0.6.1-nightly.20260107.f6771c08 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.6.1-nightly.20260107.f6771c08

$ qwen --acp

$ /approval-mode auto

$ git clone --branch v0.6.1-nightly.20260107.2c285394 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.6.1-nightly.20260107.2c285394

$ qwen --acp

$ /approval-mode auto

$ git clone --branch v0.6.0-preview.3 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.6.0-preview.3

$ qwen --acp

$ git clone --branch v0.6.0-nightly.20260107.b95d9a8d https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.6.0-nightly.20260107.b95d9a8d

$ qwen-code --acp

Zed v0.218.5 adds Dev Container support, worktree trust, Git tree view, and new agent tools.

$ git clone --branch v0.218.5 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.218.5

$ zed --wait ~/my-project

• ›Adds save_file and restore_file_from_disk agent tools to handle dirty buffers when editing files in the agent thread.
• ›Adds optional timeout_ms parameter to the agent terminal tool, letting the agent bound command runtime and prevent indefinite hangs.
• ›Adds auto_discover setting for Ollama to optionally limit visible models to only those manually configured in available_models.
• ›Introduces worktree trust mechanism, configurable via { "session": { "trust_all_worktrees": true } } in settings.
• ›Adds ability to open a project in a Dev Container when a .devcontainer/devcontainer.json file is present.

An open-source AI agent that brings the power of Gemini directly into your terminal.

gemini-cli v0.23.0 adds Gemini 3 Flash, HTTP/SSE MCP servers, /auth logout, Windows clipboard image paste, and more

$ git clone --branch v0.23.0 https://github.com/google-gemini/gemini-cli.git
# already have the repo? check out this version:
$ git checkout v0.23.0

$ gemini /auth logout

• ›Adds support for Gemini 3 Flash as an available model in the CLI
• ›Supports HTTP/SSE based MCP servers via the official ACP SDK
• ›New /auth logout command clears credentials and auth state
• ›Adds Windows clipboard image support with Alt+V paste workaround
• ›Introduces remote agent infrastructure for orchestrating remote agents

llama.cpp b7655 adds llama-debug, a new CLI utility for inspecting and validating converted GGUF models.

$ git clone --branch b7655 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b7655

$ ./build/bin/llama-debug -m models/Qwen2.5-0.5B-Instruct.gguf --prompt "Hello, my name is" --save-logits

$ ./scripts/utils/compare_tokens.py pytorch-gemma-3-270m-it llamacpp-gemma-3-270m-it-bf16 -v

• ›Adds llama-debug binary with --save-logits, --prompt / -p, --tensor-filter, --verbose, and --embedding flags for saving and inspecting logits, embeddings, and tensor outputs from a converted model.
• ›Adds compare_tokens.py script (in scripts/utils/) with an optional -v flag to diff tokenizer output — token IDs and prompts — between an original and a converted model.
• ›Integrates token-ID comparison into existing compare-logits.py and semantic_check.py verification scripts so tokenizer mismatches are caught before logit/embedding checks run.
• ›Uses llama_model_n_embd_out in the debug example to correctly handle late-interaction retriever models (e.g. LFM2-ColBert-350M) whose output embeddings are down-projected to a lower dimension.

$ git clone --branch b7652 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b7652

Langfuse v3.145.0 adds inline trace comments on partial IO data and selective metadata expansion on observations-v2 endpoints.

$ git clone --branch v3.145.0 https://github.com/langfuse/langfuse.git
# already have the repo? check out this version:
$ git checkout v3.145.0

• ›Adds selective expansion of metadata on observations-v2 API endpoints, allowing callers to request metadata only when needed.
• ›Adds inline comments on fractions of trace input/output data, enabling annotation at the sub-message level in the trace UI.