Heads up This site is currently under heavy development.
Subscribe Get it delivered — the daily firehose, filtered to the tools you run, plus the documentation changes vendors never announce. Compare plans →

The AI Toolchain — issue -166, March 5, 2026

THE AI TOOLCHAIN NO. -166
Tail
THE DAILY RELEASE FIREHOSE
PUBLISHED MARCH 5, 2026 · EVERY WEEKDAY
EDITIONS tail grep head diff uniq

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.

VIEW
ISSUE VIEW full issue
Do you prefer this view?
$ tct list   # 13 tools matched
AI & LLM Tooling
◆  AI Coding Agents

Anthropic Claude Code

Sources Release notes → v2.1.69 NOTES

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 v2.1.69 adds Voice STT in 20 languages, new hook events, plugin hot-reload, and a suite of new settings and CLI options.

└──▷ GET THIS VERSION
$ git clone --branch v2.1.69 https://github.com/anthropics/claude-code.git
# already have the repo? check out this version:
$ git checkout v2.1.69
└──▷ TRY IT
Strip Claude's default git commit/PR instructions from the system prompt when your team enforces its own commit conventions via a CLAUDE.md rule.
$ CLAUDE_CODE_DISABLE_GIT_INSTRUCTIONS=1 claude
Hot-reload plugins mid-session after editing a SKILL.md or plugin manifest, without restarting Claude Code.
$ /reload-plugins
  • Adds /claude-api skill for building applications with the Claude API and Anthropic SDK
  • Adds Ctrl+U on an empty bash prompt (!) to exit bash mode, matching Escape and Backspace
  • Adds numeric keypad support for selecting options in Claude's interview questions
  • Adds optional --name argument to /remote-control and claude remote-control to set a custom session title visible in claude.ai/code
  • Expands Voice STT support to 20 total languages, adding Russian, Polish, Turkish, Dutch, Ukrainian, Greek, Czech, Danish, Swedish, and Norwegian
+16 moreshow less
  • Adds effort level display (e.g., "with low effort") to the logo and spinner so the active effort setting is always visible
  • Adds agent name display in the terminal title when using claude --agent
  • Adds sandbox.enableWeakerNetworkIsolation setting (macOS only) to allow Go programs like gh, gcloud, and terraform to verify TLS certificates when using a custom MITM proxy with httpProxyPort
  • Adds includeGitInstructions setting and CLAUDE_CODE_DISABLE_GIT_INSTRUCTIONS env var to remove built-in commit and PR workflow instructions from Claude's system prompt
  • Adds /reload-plugins command to activate pending plugin changes without restarting
  • Adds ${CLAUDE_SKILL_DIR} variable so skills can reference their own directory in SKILL.md content
  • Adds InstructionsLoaded hook event that fires when CLAUDE.md or .claude/rules/*.md files are loaded into context
  • Adds agent_id and agent_type fields to hook events for subagent and --agent sessions
  • Adds worktree field to status line hook commands with name, path, branch, and original repo directory in --worktree sessions
  • Adds pluginTrustMessage in managed settings to append organization-specific context to the plugin trust warning
  • Adds pathPattern to strictKnownMarketplaces for regex-matching file/directory marketplace sources alongside hostPattern restrictions
  • Adds plugin source type git-subdir to point to a subdirectory within a git repo
  • Adds oauth.authServerMetadataUrl config option for MCP servers to specify a custom OAuth metadata discovery URL when standard discovery fails
  • Extends policy limit fetching (e.g., remote control restrictions) to Team plan OAuth users, not just Enterprise
  • [VSCode] Adds compaction display as a collapsible "Compacted chat" card with the summary inside
  • [VSCode] Permission mode picker now respects permissions.disableBypassPermissionsMode from effective Claude Code settings
Was this useful?

Charm Crush

Sources Release notes → v0.47.2 NOTES

Glamourous agentic coding for all

Crush v0.47.2 adds OpenAI GPT-5.4 and GPT-5.4 Pro model support plus Zhipu and QiniuCloud providers.

└──▷ GET THIS VERSION
$ git clone --branch v0.47.2 https://github.com/charmbracelet/crush.git
# already have the repo? check out this version:
$ git checkout v0.47.2
  • Adds support for OpenAI GPT-5.4 and GPT-5.4 Pro models (required an explicit Crush update to enable).
  • Adds Zhipu and Zhipu Coding (bigmodel.cn) as new AI providers.
  • Adds QiniuCloud as a new AI provider.
Was this useful?

GitHub Copilot CLI

Sources Release notes → v0.0.422 NOTES

GitHub Copilot CLI v0.0.422 adds GPT-5.4, JSON output mode, startup hooks, plugin SSH installs, and session metrics logging.

└──▷ GET THIS VERSION
$ git clone --branch v0.0.422 https://github.com/github/copilot-cli.git
# already have the repo? check out this version:
$ git checkout v0.0.422
└──▷ TRY IT
Pipe Copilot CLI output into jq for programmatic processing of responses in CI or scripts.
$ gh copilot prompt --output-format json 'Summarize open TODOs in this repo' | jq '.content'
Auto-copy selected text to clipboard whenever you highlight it in alt-screen mode — useful for quickly grabbing code snippets.
yaml
copy_on_select: true
Ensure a standard set of plugins is installed automatically for every session by declaring them in config.
yaml
enabledPlugins:
  - my-org/copilot-plugin-terraform
  - ssh://[email protected]/infra/copilot-plugin-k8s
  • Adds --output-format json flag to emit JSONL in prompt mode for programmatic integrations.
  • Adds copy_on_select config option to auto-copy selected text to clipboard in alt-screen mode.
  • Adds enabledPlugins config key for automatic plugin installation at startup.
  • Adds exitPlanMode.request protocol method for SDK plan approval support.
  • Renames launch_messages config setting to companyAnnouncements — existing configs using launch_messages must be updated.
+12 moreshow less
  • Renames repository config file from .github/copilot/config.json to settings.json — existing repo-level config files must be moved.
  • Session usage metrics (requests, tokens, code changes) are now persisted to events.jsonl after each session ends.
  • Loads personal hooks from ~/.copilot/hooks in addition to repo-level .github/hooks.
  • Adds /copy command to copy the last response to clipboard.
  • Adds startup prompt hooks to auto-submit prompts or slash commands when a session starts.
  • Supports installing plugins from ssh:// URLs.
  • Adds GPT-5.4 model support.
  • Plugin-contributed LSP servers are now loaded, started, and shown in /lsp show.
  • Adds Ctrl+R reverse incremental command history search (matching Bash behavior).
  • Adds mouse scroll support to /diff in alt-screen mode.
  • Automatic notifications when background shell commands and agents complete.
  • /delegate now prompts for a target remote in multi-remote repositories.
└──▷ BREAKING ON UPGRADE
  • !The launch_messages config setting is renamed to companyAnnouncements; existing configs using launch_messages will need to be updated.
  • !The repository config file is renamed from .github/copilot/config.json to settings.json; existing repo-level config files must be moved to the new path.
Was this useful?

Block Goose

Sources Release notes → v1.27.0 NOTES

an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM

Goose v1.27.0 adds structured shell output, tree-sitter AST analysis, Anthropic adaptive thinking, and self-signed HTTPS for goosed.

└──▷ GET THIS VERSION
$ git clone --branch v1.27.0 https://github.com/block/goose.git
# already have the repo? check out this version:
$ git checkout v1.27.0
└──▷ TRY IT
Run a shell command and consume the structured stdout/stderr output separately for automated analysis or CI pipelines.
$ # When Goose executes shell commands, the tool now returns:
# { "stdout": "...", "stderr": "..." }
# Reference stdout or stderr independently in follow-up instructions or recipes.
  • Adds structured {stdout, stderr} return from the shell tool with an explicit output schema, making it easier to parse command results programmatically.
  • New platform extension with tree-sitter AST parsing enables deeper code analysis within Goose sessions.
  • Supports Anthropic adaptive (extended) thinking mode for compatible Claude models.
  • Adds fullscreen and Picture-in-Picture (PiP) display modes for MCP Apps.
  • Introduces self-signed HTTPS support for the goosed server, enabling secure local deployments.
+3 moreshow less
  • Adds code-mode feature flag to include pctx/Code Mode as an optional dependency.
  • Adds base_path field to custom provider config for more flexible provider routing.
  • Adds GPT 5-3-Codex model support in the Databricks provider.
Was this useful?

OpenAI Codex CLI

Sources Release notes → rust-v0.111.0 2 RELEASES · 2026-03-05 NOTES STABLE

Lightweight coding agent that runs in your terminal

Codex CLI rust-v0.111.0 enables Fast mode by default, adds local ESM imports in js_repl, and exposes MCP elicitation as structured flows.

└──▷ GET THIS VERSION
$ git clone --branch rust-v0.111.0 https://github.com/openai/codex.git
# already have the repo? check out this version:
$ git checkout rust-v0.111.0
└──▷ TRY IT
Import a local workspace utility script inside a js_repl session to reuse shared helpers without copying code.
$ const utils = await import('./scripts/utils.mjs');
utils.analyzeResults(data);
  • Enables Fast mode by default for all sessions; TUI session header now displays whether the session is running in Fast or Standard mode.
  • Adds dynamic import of local .js and .mjs files in js_repl, allowing workspace scripts to be reused directly from the REPL.
  • Codex now reports enabled plugins (MCPs, apps, skills) to the model at session start, improving discovery of installed integrations.
  • App-server v2 exposes MCP elicitation as a structured request/response flow instead of raw events, simplifying client integrations.
  • Expands image workflow support for clients with client-side handling of image-generation events and model metadata for image-capable web search.
1 more release in this issue · 2026-03-05
rust-v0.110.0 NOTES STABLE

Codex CLI gains a plugin system, multi-agent TUI upgrades, a /fast toggle, workspace-scoped memories, and a Windows installer.

└──▷ GET THIS VERSION
$ git clone --branch rust-v0.110.0 https://github.com/openai/codex.git
# already have the repo? check out this version:
$ git checkout rust-v0.110.0
└──▷ TRY IT
Switch to fast mode mid-session to reduce latency when iterating quickly on a task.
$ /fast
Spin up a multi-agent workflow in the TUI and approve sub-agent actions interactively.
$ /agent
  • Adds a plugin system that loads skills, MCP entries, and app connectors from config or a local marketplace, with an install endpoint for enabling plugins from the app server.
  • Expands TUI multi-agent flow with approval prompts, /agent-based enablement, ordinal nicknames, role-labeled handoff context, and clearer prompts.
  • Adds a persisted /fast toggle in the TUI plus app-server support for fast and flex service tiers.
  • Improves memories with workspace-scoped writes, renamed memory settings, and guardrails against saving stale or polluted facts.
  • Adds a direct Windows installer script to published release artifacts.
Was this useful?

SST OpenCode

Sources Release notes → v1.2.18 NOTES

The open source coding agent.

OpenCode v1.2.18 adds non-interactive auth login options, desktop deep links for new sessions, and clickable inline tool/task components.

└──▷ GET THIS VERSION
$ git clone --branch v1.2.18 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v1.2.18
  • Adds options to auth login command to skip interactive questions, enabling non-interactive/scripted authentication flows.
  • Adds desktop deep link support for creating new sessions directly from external links.
  • Adds onClick handler to InlineTool and Task components in the TUI.
  • Opens search with Mod+F in the desktop app even when the editor is not focused.
  • Forks Ghostty for a web-based terminal implementation in the desktop app.
Was this useful?

Earendil Works Pi

Sources Release notes → v0.56.2 NOTES

AI agent toolkit: unified LLM API, agent loop, TUI, coding agent CLI

Pi v0.56.2 adds GPT-5.4 support, a new treeFilterMode setting, and native Mistral conversations integration.

└──▷ GET THIS VERSION
$ git clone --branch v0.56.2 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:
$ git checkout v0.56.2
  • Supports GPT-5.4 across openai, openai-codex, azure-openai-responses, and opencode providers, with gpt-5.4 now the default for openai and openai-codex.
  • Adds gpt-5.3-codex as a fallback model for github-copilot until upstream model catalogs include it.
  • New treeFilterMode setting lets you choose the default /tree filter mode from default, no-tools, user-only, labeled-only, or all.
  • Adds native Mistral conversations integration via the SDK-backed provider, preserving Mistral-specific thinking and replay semantics.
└──▷ BREAKING ON UPGRADE
  • !The default model for the openai and openai-codex providers is now gpt-5.4; any workflow pinned to the previous default will silently switch models on upgrade.
Was this useful?
◆  AI Model & Data Infrastructure

Microsoft ONNX Runtime

Sources Release notes → v1.24.3 NOTES

ONNX Runtime v1.24.3 adds DQ→MatMulNBits fusion for NvTensorRtRtx, OrtEnv.DisableDllImportResolver for NuGet, and DynamicQGemm ukernel in MLAS.

└──▷ GET THIS VERSION
$ git clone --branch v1.24.3 https://github.com/microsoft/onnxruntime.git
# already have the repo? check out this version:
$ git checkout v1.24.3
  • Adds OrtEnv.DisableDllImportResolver API to prevent fatal errors when a DLL import resolver conflict is detected in NuGet environments.
  • Adds 'library_path' metadata entry to OrtEpDevice instances for plugin and provider bridge execution providers.
  • Adds DynamicQGemm function pointers and ukernel interface to MLAS.
  • Adds DQ→MatMulNBits fusion transformer to the NvTensorRtRtx execution provider.
  • Increases kMaxValueLength to 8192 in config.
+5 moreshow less
  • Builds Windows ARM64X binaries as part of the packaging pipeline.
  • Adds Python 3.14 CI support with upgraded dependencies.
  • Delivers up to 4x QMoE CPU performance improvement on 4-bit workloads.
  • Supports using the embedded WASM module in Blob URL workers when wasmBinary is provided in WebGPU EP.
  • Allows a new memory info name for WebGPU EP.
Was this useful?

Ollama

Sources Release notes → v0.17.7 NOTES

Get up and running with Kimi-K2.6, GLM-5.2, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

Ollama v0.17.7 adds thinking-level support in the API and context-length compaction for ollama launch.

└──▷ GET THIS VERSION
$ git clone --branch v0.17.7 https://github.com/ollama/ollama.git
# already have the repo? check out this version:
$ git checkout v0.17.7
└──▷ TRY IT
Control reasoning depth on a thinking model by passing a named thinking level through the API instead of a raw token budget.
$ curl http://localhost:11434/api/chat -d '{"model": "qwq", "options": {"thinking": "medium"}, "messages": [{"role": "user", "content": "Explain RSA encryption."}]}'
  • Supports named thinking levels (e.g., "medium") for thinking models via Ollama's API.
  • Adds context-length parameter to enable compaction when using ollama launch.
Was this useful?
◆  AI Agent Frameworks

OpenAI Agents SDK

Sources Release notes → v0.10.5 NOTES

OpenAI Agents SDK v0.10.5 adds explicit prefix mode control to MultiProvider.

└──▷ GET THIS VERSION
$ git clone --branch v0.10.5 https://github.com/openai/openai-agents-python.git
# already have the repo? check out this version:
$ git checkout v0.10.5
  • Adds explicit prefix modes to MultiProvider, giving developers direct control over how model-name prefixes are applied when routing across multiple model providers.
Was this useful?

PydanticAI

Sources Release notes → v1.66.0 NOTES

PydanticAI v1.66.0 adds native structured output for Qwen 3.5 models and support for the Gemini image-preview model.

└──▷ GET THIS VERSION
$ git clone --branch v1.66.0 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v1.66.0
  • Enables native structured output for Qwen 3.5 models.
  • Adds support for gemini-3.1-flash-image-preview (Nano Banana 2) as a supported model.
Was this useful?
◆  Local LLM Runtimes

llama.cpp

Sources Release notes → b8210 NOTES

CUDA backend gains async CPU-to-GPU tensor copies with relaxed sync requirements, improving multi-token throughput.

└──▷ GET THIS VERSION
$ git clone --branch b8210 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b8210
  • Adds CPU-to-CUDA async copy capability to ggml_backend_cuda_cpy_tensor_async(), replacing synchronous copies to reduce GPU pipeline stalls.
  • Introduces ggml_backend_sync_mode opt-in mechanism to relax explicit synchronization between host-to-device copies and graph execution, extensible to backends beyond CUDA (e.g., Vulkan).
Was this useful?
Other / Uncategorized
◆  AI OBSERVABILITY

Langfuse

Sources Release notes → v3.156.0 NOTES

Langfuse v3.156.0 adds ClickHouse Cloud auth, LiveKit OTEL span parsing, SSO tokenEndpointAuthMethod, thinking-parts support in playground/evals, and typed ObservationsV2Response.

└──▷ GET THIS VERSION
$ git clone --branch v3.156.0 https://github.com/langfuse/langfuse.git
# already have the repo? check out this version:
$ git checkout v3.156.0
  • Adds tokenEndpointAuthMethod support to multi-tenant SSO configurations.
  • Adds extraHeaders support for the Anthropic adapter in LLM connections.
  • Types the data field of ObservationsV2Response instead of map<string, unknown>, giving callers a concrete schema.
  • Adds ClickHouse Cloud authentication provider for self-hosted deployments.
  • Parses LiveKit attributes on OpenTelemetry spans, surfacing LiveKit trace data natively.
+9 moreshow less
  • Supports 'thinking parts' in the playground and evals, enabling reasoning-model workflows.
  • Includes triggering user info in webhook and GitHub dispatch payloads.
  • Maps input/output to the Pydantic root span in the trace UI.
  • Enables ClickHouse query condition caching for analytics queries, reducing latency at scale.
  • Adds dashboard widget definition version tracking.
  • Validates that high-cardinality dimensions require a top-N shape on the v2 dashboard path, preventing misconfigured widgets.
  • Shows full trace data on hover at small row height in the trace table.
  • Defaults to hiding Langfuse environments in the filters UI to reduce noise.
  • Adds the analytics_events_core ClickHouse view.
Was this useful?
my-toolchain — 0 tools
paste an install list to detect your tools

A brew list, a Brewfile, requirements.txt, a Dockerfile — or just the product names, free-form. Nothing leaves your browser.

    browse all tools →