The daily firehose — everything the toolchain shipped today, already filtered.
// HOW THIS ISSUE IS MADE
We read every release from the 174 tools on our watchlist at the source — GitHub and GitLab release notes, vendor release pages and changelogs, project blogs and feeds, vendor press releases, and the source code behind the tag. Bug-fix-only releases and non-product newsroom noise are dropped; what's left is summarized down to the new capability, how to try it, and any screenshots or videos the release itself published. Every entry links to the sources it was built from.
›Adds max_tool_calls_from_history parameter to load a fixed number of tool calls from agent history, reducing token consumption and managing context size.
›Adds PPTXReader class to support ingesting Microsoft PowerPoint (.pptx) files.
Lightweight coding agent that runs in your terminal
Codex CLI gains --add-dir, MCP tool filtering, OAuth scopes, enterprise login controls, and auto-compaction at 90% context.
└──▷ GET THIS VERSION
$ git clone --branch rust-v0.48.0 https://github.com/openai/codex.git
# already have the repo? check out this version:$ git checkout rust-v0.48.0
└──▷ TRY IT
Give the agent write access to a second directory (e.g. a shared library) alongside your current project.
$ codex --add-dir ../shared-lib 'Refactor the auth module to use the shared validation helpers'
Restrict an MCP server to only the tools your workflow needs, reducing attack surface.
toml
# In your Codex config (e.g. ~/.codex/config.toml):
[mcp_servers.my_server]
enabled_tools = ["read_file", "search_code"]
Lock a team deployment to a specific login method and ChatGPT workspace via managed config.
toml
# In managed config:
forced_login_method = "chatgpt"
forced_chatgpt_workspace_id = "ws-abc123"
›Adds --add-dir flag to specify additional writable working directories beyond the default.
›MCP stdio servers now use the official Rust MCP SDK (rmcp) as their client backend.
›MCP stdio servers can specify a cwd to set their working directory.
›MCP servers support enabled_tools and disabled_tools config options to selectively allow or block individual tools.
›Streamable HTTP MCP servers can specify OAuth scopes during codex mcp login.
+10 moreshow less
›Adds forced_login_method and forced_chatgpt_workspace_id configuration options for enterprise managed-config control.
›Adds experimental codex stdio-to-uds subcommand.
›Adds experimental_bearer_token option to model provider definitions for custom auth.
›Auto-compacts conversation context when it reaches ~90% of the model's limit.
›Plan tool is now enabled by default in exec mode.
›Local tokenizer used for accurate context truncation in unified exec.
›Text cleared with Ctrl+C is preserved in input history and recoverable with the up arrow.
›MCP tool names are now prefixed with mcp__ for clearer namespacing.
›Dedicated error message for GitHub MCP servers missing a personal access token.
›When MCP auth expires, the user is prompted to log in again interactively.
└──▷ BREAKING ON UPGRADE
!MCP tool names are now prefixed with mcp__ — any automation or configuration referencing MCP tool names by their previous unprefixed names will break.
An open-source AI agent that brings the power of Gemini directly into your terminal.
gemini-cli v0.12.0-nightly adds an executeCommand endpoint with /extensions list support and extension install-time settings requests.
└──▷ GET THIS VERSION
$ git clone --branch v0.12.0-nightly.20251023.c4c0c0d1 https://github.com/google-gemini/gemini-cli.git
# already have the repo? check out this version:$ git checkout v0.12.0-nightly.20251023.c4c0c0d1
›Adds executeCommand endpoint supporting /extensions list to query installed extensions programmatically.
›Extensions can now declare and request user settings at install time.
›Surfaces internal errors through a unified event system for more consistent error visibility.
gemini-cli v0.12.0-nightly adds MCP prompt commands in non-interactive mode, loop-detection override, extensions alias, and a YOLO-mode disable setting.
└──▷ GET THIS VERSION
$ git clone --branch v0.12.0-nightly.20251023.a7faa208 https://github.com/google-gemini/gemini-cli.git
# already have the repo? check out this version:$ git checkout v0.12.0-nightly.20251023.a7faa208
›Enables running MCP prompt commands in non-interactive (scripted/piped) mode.
›Continues a request after the user disables loop detection mid-session.
›Adds extension as a short alias for the extensions command.
›Adds a setting to disable YOLO mode (auto-approve all tool calls).
oobabooga text-generation-webui v3.16 adds symlink support for portable builds and updates llama.cpp with Ling-mini-2.0 and Ring-mini-2.0 model support.
└──▷ GET THIS VERSION
$ git clone --branch v3.16 https://github.com/oobabooga/textgen.git
# already have the repo? check out this version:$ git checkout v3.16
›Enables running a portable Web UI build via a symlink, allowing flexible install layouts without copying the full package.
›Updates llama.cpp backend to add support for Ling-mini-2.0 and Ring-mini-2.0 model architectures.
Weave v0.52.11 adds CRUD endpoints for Ops, Datasets, Scorers, and Evaluations, plus incremental score aggregation and a fire-and-forget EvaluationLogger API.
└──▷ GET THIS VERSION
$ git clone --branch v0.52.11 https://github.com/wandb/weave.git
# already have the repo? check out this version:$ git checkout v0.52.11
›Adds new CRUD endpoints for interacting with Ops, enabling programmatic create/read/update/delete of Op objects via the trace server API.
›Adds new CRUD endpoints for interacting with Datasets, enabling programmatic create/read/update/delete of Dataset objects via the trace server API.
›Adds new CRUD endpoints for interacting with Scorers, enabling programmatic create/read/update/delete of Scorer objects via the trace server API.
›Adds new CRUD endpoints for interacting with Evaluations, enabling programmatic create/read/update/delete of Evaluation objects via the trace server API.
›Adds a fire-and-forget API for EvaluationLogger with promise chain coordination in the TypeScript SDK, enabling non-blocking evaluation logging.
+3 moreshow less
›Implements incremental score aggregation and update mechanisms in EvaluationLogger.
›Supports context propagation in EvaluationLogger, allowing trace context to flow through evaluation workflows.
›Introduces ImperativeEval in the TypeScript SDK (Part 1), enabling imperative-style evaluation authoring in TS.
THE AI TOOLCHAIN