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.
PydanticAI v1.97.0 adds MCPToolset, OnlineEvaluator error opt-in, streaming state tracking, and splits GoogleProvider into two classes.
└──▷ GET THIS VERSION
$ git clone --branch v1.97.0 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:$ git checkout v1.97.0
└──▷ USE IT
Evaluate agent calls that raised errors, not just successful completions, to catch failure-mode regressions.
python
evaluator = OnlineEvaluator(run_on_errors=True)
›Adds OnlineEvaluator.run_on_errors flag to opt into running evaluations on failed (errored) agent calls, not just successful ones.
›Adds MCPToolset (backed by fastmcp-slim[client]) as the new MCP integration class, replacing the deprecated MCPServer* and FastMCPToolset.
›Splits GoogleProvider(vertexai=True|False) into two separate classes: GoogleProvider (formerly google-gla:, now provider ID google:) and GoogleCloudProvider (formerly google-vertex:, now provider ID google-cloud:).
›Sets ModelResponse.state to incomplete while a response is still streaming, enabling callers to distinguish in-progress from finished responses.
›Promotes pydantic_graph.beta API out of beta into the stable namespace.
+2 moreshow less
›Adds stream_response() (singular) as the replacement for stream_responses(); the new method yields ModelResponse directly instead of a (ModelResponse, is_last) tuple.
›Replaces the bundled fasta2a A2A integration with an external fasta2a.pydantic_ai adapter (requires fasta2a v0.6.1+), following DataLayer's adoption of the project.
└──▷ BREAKING ON UPGRADE
!The google-gla: provider ID is renamed to google: and google-vertex: is renamed to google-cloud:; old names are deprecated and will be removed in v2.
!stream_responses() is deprecated in favor of stream_response(); the new singular form yields ModelResponse instead of (ModelResponse, is_last), so any code unpacking the tuple will break when migrated.
!Agent.to_a2a() and the bundled fasta2a integration are deprecated; users must switch to fasta2a.pydantic_ai (requires fasta2a v0.6.1) from the external package.
!The pydantic_graph.beta module is deprecated; import paths that relied on the .beta namespace must be updated to the stable API.
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
Claude Code v2.1.143 adds plugin dependency enforcement, projected context costs, and richer claude agents session control.
└──▷ GET THIS VERSION
$ git clone --branch v2.1.143 https://github.com/anthropics/claude-code.git
# already have the repo? check out this version:$ git checkout v2.1.143
└──▷ TRY IT
Launch the agents dashboard scoped to a specific model, effort level, and MCP config so every dispatched session inherits those defaults.
$ claude agents --model claude-opus-4-5 --effort high --mcp-config ./mcp.json
Allow background sessions to edit the working copy directly when git worktrees are impractical in your repo.
json
# In settings.json
{ "worktree": { "bgIsolation": "none" } }
›Adds plugin dependency enforcement: claude plugin disable refuses when another enabled plugin depends on the target and shows a copy-pasteable disable-chain hint; claude plugin enable force-enables transitive dependencies.
›Adds projected context cost (per-turn and per-invocation token estimates) to the /plugin marketplace browse pane.
›New worktree.bgIsolation: "none" setting lets background sessions edit the working copy directly without EnterWorktree, for repos where worktrees are impractical.
›PowerShell tool now passes -ExecutionPolicy Bypass by default; opt out with CLAUDE_CODE_POWERSHELL_RESPECT_EXECUTION_POLICY=1.
›PowerShell tool is now enabled by default on Windows for Bedrock, Vertex, and Foundry users; opt out with CLAUDE_CODE_USE_POWERSHELL_TOOL=0.
+8 moreshow less
›claude agents now accepts --add-dir, --settings, --mcp-config, --plugin-dir, --permission-mode, --model, --effort, and --dangerously-skip-permissions to configure the dashboard and sessions dispatched from it.
›claude --bg --dangerously-skip-permissions now persists across retire→wake cycles.
›/bg now preserves --mcp-config, --settings, --add-dir, --plugin-dir, and --strict-mcp-config, keeping MCP servers and settings across respawn.
›/bg and ←-detach now preserve --fallback-model, so backgrounded workers degrade gracefully on overload.
›/bg and ←-detach now preserve --allow-dangerously-skip-permissions, keeping bypass-permissions available in the Shift+Tab cycle.
›Background sessions now preserve the model and effort level set after waking from idle.
›Shift+Tab in attached agent sessions now includes auto mode in the permission cycle.
›Background sessions launched from claude agents now honor permissions.defaultMode from settings.json.
└──▷ BREAKING ON UPGRADE
!Worktree cleanup no longer falls back to rm -rf when git worktree remove fails, preventing accidental deletion of gitignored or in-progress files — workflows that relied on that fallback cleanup behavior will need an alternative removal step.
›Integrates KV offloading subsystem with the Hybrid Memory Allocator (HMA), including scheduler-side sliding window group support and multi-connector HMA.
›Adds bi-directional KV cache transfers between prefill (P) and decode (D) nodes in disaggregated serving.
›Adds Mooncake KVConnectorStats for KV transfer observability in disaggregated serving.
›Enables independent drafter attention backend selection for speculative decoding.
›Adds EAGLE speculative decoding support for Mistral, Gemma4 MTP, MiMo-V2.5 MTP, and Cohere Eagle.
›Adds tool calling support via XGrammar 0.2.0 with structural tags for strict tool calling combined with reasoning.
›Adds Cohere reasoning and tool parsers, plus LFM2/2.5 tool parser.
›Adds Responses API streaming tool/function calling with required and named tool/function choice, plus resubmitting output items with missing fields.
›Adds rendered prompt text in chat completion response.
›Adds Fastokens support in the tokenizer.
›Adds RayExecutorV2 enabled by default for Ray-based execution.
›Adds IndexCache support for DSA models.
›Enables FlashInfer top-k/top-p sampler by default.
›Adds OOM prevention via max_split_size_mb during model loading.
›Adds two-phase pause mechanism to prevent scheduler deadlock.
›Adds NIXL connector bumped to 1.x for disaggregated serving.
›Adds new model architecture support: MiMo-V2.5, Laguna XS.2, Moondream3, Qianfan-OCR, Cohere MoE, and Cohere Eagle.
›Adds AMD ROCm support for DeepSeek V4, including TP4 AITER MLA and pipeline parallelism.
›Adds AMD Dynamic Batch Optimization (DBO) on ROCm.
›Adds CPU FP8 attention for AMX/AVX-512, FP8 W8A16 linear, FP8 W8A16 MoE, and DNNL AVX2 W8A8 Int8.
›Adds Intel XPU top-k/top-p sample kernel and LoRA support.
›Adds IBM Power VSX attention backend.
›Adds NVFP4 KV cache support and Triton dequant/QDQ emulation kernels for Hopper and AMD.
›Adds Humming MXFP4 MoE backend.
›Adds TurboQuant hybrid model and uniform quantization support.
›Adds configurable model weights loading tracking.
›Adds Triton JIT compilation monitor.
›Adds LoRA initial expert parallel (EP) support.
›Adds ViT CUDA graph support for Qwen2.5-VL.
›Adds Qwen3.5/Mamba hybrid model support in Model Runner V2.
›Adds Gated DeltaNet Attention for Qwen 3.5/3.6 on CPU.
›Adds RISC-V OMP thread auto-binding for CPU backend.
›Reduces Docker image size by approximately 2.5 GB via deferred FlashInfer cubin download.
›Embeds container image provenance metadata.
└──▷ BREAKING ON UPGRADE
!C++20-compatible compiler is now required to build vLLM (for PyTorch compatibility); builds with older compilers will fail.
!transformers v4 is formally deprecated; users must migrate to transformers v5.
Phoenix v15.10.0 adds trace feedback on session turns, OTel GenAI semconv conversion, and agent session retention caps.
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v15.10.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-v15.10.0
›Converts OTel GenAI semantic convention attributes to OpenInference format automatically during trace ingestion.
›Adds a trace feedback toolbar to session turns, enabling annotation directly within multi-turn agent sessions.
›Introduces agent session retention capping to bound how many sessions are stored.
›Curates the agent model menu for a streamlined model selection experience in the agents UI.
Phoenix client v2.7.0 adds agent session summaries, ATIF v1.7 trajectory upload, token counts in REST payloads, and TS trace annotations.
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-client-v2.7.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-client-v2.7.0
└──▷ TRY IT
Retrieve a summary of a specific agent session to quickly assess what happened without reading the full trace.
$ curl -X GET 'https://<phoenix-host>/agents/<agent_id>/sessions/<session_id>/summary' -H 'Authorization: Bearer <token>'
›Adds GET /agents/{agent_id}/sessions/{session_id}/summary endpoint to retrieve summarized views of individual agent sessions.
›Includes token counts in span, trace, and session payloads returned by the REST API.
›Supports ATIF v1.7 trajectory upload via the client library.
›Adds TypeScript trace annotations to the phoenix-client, with clarified note semantics and a skills audit.
›Adds session-tagged identifiers to support open and axial coding workflows.
+3 moreshow less
›Supports vendor passthrough tools, enabling pass-through of provider-native tool definitions.
›Adds Playground manipulation tools with confirmation step for agent workflows.
›Types frontend REST calls against the OpenAPI schema for stronger client-side safety.
└──▷ BREAKING ON UPGRADE
!The v1 /chat route and its associated code have been removed.
THE AI TOOLCHAIN