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.
Validate JWT tokens against a specific audience claim in an AgentOS deployment.
python
from agno.middleware.jwt import JWTMiddleware
middleware = JWTMiddleware(
secret="<your-secret>",
audience="https://api.myapp.example.com"
)
›Introduces the Skills class, enabling agents to be extended with capabilities defined by Anthropic's Agent Skill specification.
›Adds header_provider function parameter to MCPTools instances, allowing dynamic header generation (e.g. rotating auth tokens or per-user IDs) on each MCP tool call.
›Adds audience parameter to the JWTMiddleware constructor to set the expected audience when validating JWT tokens.
›MCPTools now defaults to StreamableHttp as the transport when a URL is present for connecting to external MCP servers.
›Adds A2AClient and support for the a2a protocol when using remote agents, enabling Google ADK agents to run as remote agents via AgentOS (beta).
+1 moreshow less
›Extends native reasoning support to OpenAI GPT-5.1 and 5.2, new Gemini 3, 3.5, and deepthink models, and new DeepSeek r1 and reasoner models.
HolmesGPT 0.18.2 adds TCP connectivity checks, GCP & AKS MCP integrations, and an improved Elasticsearch/OpenSearch toolset.
└──▷ GET THIS VERSION
$ git clone --branch 0.18.2 https://github.com/HolmesGPT/holmesgpt.git
# already have the repo? check out this version:$ git checkout 0.18.2
›Adds a connectivity check toolset supporting TCP checks for network reachability diagnostics.
›Adds GCP MCP integration for Google Cloud Platform context in investigations.
›Adds Azure (AKS) MCP integration for Azure Kubernetes Service context.
›Updates the Prometheus toolset configuration.
└──▷ BREAKING ON UPGRADE
!The Elasticsearch/OpenSearch toolset has breaking configuration changes — consult the updated docs at https://holmesgpt.dev/data-sources/builtin-toolsets/elasticsearch/#elasticsearch-opensearch before upgrading.
$ git clone --branch rust-v0.78.0 https://github.com/openai/codex.git
# already have the repo? check out this version:$ git checkout rust-v0.78.0
└──▷ TRY IT
Edit a long or complex prompt in your preferred editor mid-session instead of typing it inline in the TUI.
$ # While in the Codex TUI, press Ctrl+G to open the current prompt in $VISUAL or $EDITOR, edit, save, and quit — edits are synced back automatically.
›Adds Ctrl+G to open the current prompt in your configured external editor ($VISUAL/$EDITOR) and sync edits back into the TUI.
›Supports project-aware config layering: loads repo-local .codex/config.toml, honors configurable project_root_markers, and merges with system config like /etc/codex/config.toml.
›Supports enterprise-managed configuration requirements on macOS via an MDM-provided TOML payload.
›Exec policy rules can now include human-readable justifications via a justification arg to prefix_rule() in *.rules files, with policy loading following the unified config-layer stack.
›Improves tui2 transcript navigation with multi-click selection, a copy shortcut/affordance, and a draggable auto-hiding scrollbar.
+1 moreshow less
›Starts Windows PowerShell sessions in UTF-8 mode to reduce encoding-related prompt/output issues.
Pi v0.37.5 adds runtime model/thinking controls, exported truncation utilities, and a full UI component library for extensions.
└──▷ GET THIS VERSION
$ git clone --branch v0.37.5 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:$ git checkout v0.37.5
└──▷ USE IT
Cap tool output to a safe byte budget before returning it, preventing context blowout on large command output.
typescript
import { truncateTail, DEFAULT_MAX_BYTES } from "pi/tools";
const raw = await runShellCommand(cmd);
return truncateTail(raw, { maxBytes: DEFAULT_MAX_BYTES });
›Adds setModel(), getThinkingLevel(), and setThinkingLevel() ExtensionAPI methods so extensions can switch model and thinking level at runtime.
›Exports truncation utilities (truncateHead, truncateTail, truncateLine, formatSize, DEFAULT_MAX_BYTES, DEFAULT_MAX_LINES, TruncationOptions, TruncationResult) for use in custom tools.
›Exports the full suite of built-in UI components (ArminComponent, AssistantMessageComponent, BashExecutionComponent, BorderedLoader, and 20+ more) plus utilities renderDiff and truncateToVisualLines for extension developers.
›New truncated-tool.ts example demonstrating output truncation with custom rendering for extensions.
›New preset.ts example demonstrating preset configurations with model, thinking, and tools switching.
+1 moreshow less
›New Common Patterns and Key Rules sections in docs/tui.md with copy-paste code for SelectList, BorderedLoader, SettingsList, setStatus, setWidget, and setFooter.
Qwen Code v0.6.0-nightly adds German locale support and auto-detects LLM output language from system locale.
└──▷ GET THIS VERSION
$ git clone --branch v0.6.0-nightly.20260106.b19bb6cb https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:$ git checkout v0.6.0-nightly.20260106.b19bb6cb
›Adds German language support for the CLI interface.
›Auto-detects LLM output language from the system locale, so responses match the user's OS language without manual configuration.
›Supports merging ChatCompletionContentPart entries and filters empty messages in API request handling.
llama.cpp server adds thinking/reasoning content blocks to the Anthropic Messages API for reasoning models
└──▷ GET THIS VERSION
$ git clone --branch b7644 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:$ git checkout b7644
›Adds thinking content blocks to the Anthropic Messages API when using --reasoning-format deepseek with the thinking parameter enabled — both non-streaming (thinking block before text in content array) and streaming (thinking_delta events with correct block indices) are supported.
›Tracks reasoning state across streaming chunks via the anthropic_has_reasoning member variable to correctly handle partial streaming of thinking content.
Phoenix v12.28.0 adds JSONL dataset uploads and a built-in tool selection correctness eval metric.
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v12.28.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-v12.28.0
›Adds support for JSONL format in dataset uploads.
›Adds a built-in tool selection correctness metric to the evals framework.
›Adds a skeleton loader for paragraphs in the playground UI while content loads.
THE AI TOOLCHAIN