Backfill issue -068 — The AI Toolchain

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   # 12 tools matched

AI & LLM Tooling

◆ AI Agent Frameworks

Agno (formerly Phidata)

v2.6.14 NOTES

Agno v2.6.14 adds CRUD endpoints for learnings on AgentOS.

└──▷ GET THIS VERSION

$ git clone --branch v2.6.14 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v2.6.14

›Adds create, read, update, and delete endpoints for learnings on AgentOS.

Was this useful?

LangChain

langchain-anthropic==1.4.6 NOTES

LangChain Anthropic 1.4.6 adds package version tracking to tracing metadata and streaming tool call chunk validation.

└──▷ GET THIS VERSION

$ git clone --branch langchain-anthropic==1.4.6 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-anthropic==1.4.6

›Adds package version tracking to tracing metadata, enabling richer observability when debugging LangChain Anthropic pipelines.
›Validates tool call chunks during streaming in standard tests, surfacing malformed partial tool calls earlier in the development cycle.

Was this useful?

OpenClaw

v2026.6.6 NOTES

Your own personal AI assistant. Any OS.

OpenClaw v2026.6.6 adds OpenRouter OAuth, Claude Fable 5, QQBot group mentions, trusted diagnostics tool I/O capture, and ClawHub reusable package publishing.

└──▷ GET THIS VERSION

$ git clone --branch v2026.6.6 https://github.com/openclaw/openclaw.git
# already have the repo? check out this version:
$ git checkout v2026.6.6

└──▷ TRY IT

Check for managed plugin version drift across installed ClawHub packages before a deployment.

$ openclaw doctor

›Adds OpenRouter OAuth support as a new provider authentication path.
›Adds Claude Fable 5 model support.
›Adds QQBot group mention toggle for channel integrations.
›Emits Claude CLI commentary progress events and bridges inter-tool commentary into channel progress.
›Enables trusted diagnostics channels to capture tool input/output content for observability.

+9 moreshow less

›Adds first-assistant-event tracing and slow-initial-reply warnings to the Control UI.
›Supports reusable ClawHub package publishing with dry-run skip-approval option.
›Moves the local llama.cpp runtime into its own provider plugin and batches embeddings across files.
›Persists the agent model catalog cache to speed up startup.
›Exposes the active connection host in the TUI footer.
›Improves iPad and iPhone control surfaces.
›Prewarms TUI runtime plugins for faster startup.
›Supports declared installed trusted hooks for plugins.
›Reports managed plugin version drift via the doctor command.

Was this useful?

browser-use

0.13.2 NOTES

browser-use 0.13.2 adds provider-prefixed model strings to ChatBrowserUse

└──▷ GET THIS VERSION

$ git clone --branch 0.13.2 https://github.com/browser-use/browser-use.git
# already have the repo? check out this version:
$ git checkout 0.13.2

›Adds support for provider-prefixed model strings (e.g. openai/gpt-5.5) in ChatBrowserUse, letting callers specify the provider inline without separate configuration.

Was this useful?

◆ AI Coding Agents

Anthropic Claude Code

v2.1.176 3 RELEASES · 2026-06-12 NOTES STABLE

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 v2.1.176 adds language-aware session titles, footer link badges, and smarter Bedrock credential caching.

└──▷ GET THIS VERSION

$ git clone --branch v2.1.176 https://github.com/anthropics/claude-code.git
# already have the repo? check out this version:
$ git checkout v2.1.176

›Session titles are now generated in the language of the conversation; pin a specific language with the language setting.
›New footerLinksRegexes setting lets you add regex-matched link badges to the footer row via user or managed settings.

2 more releases in this issue · 2026-06-12

v2.1.175 NOTES STABLE

Claude Code adds enforceAvailableModels managed setting to lock model selection within admin-defined allowlists.

└──▷ GET THIS VERSION

$ git clone --branch v2.1.175 https://github.com/anthropics/claude-code.git
# already have the repo? check out this version:
$ git checkout v2.1.175

└──▷ USE IT

Lock an enterprise deployment so users can never select a model outside the admin-approved list, even via project or user settings.

json

{
  "enforceAvailableModels": true,
  "availableModels": ["claude-3-5-sonnet-20241022", "claude-3-haiku-20240307"]
}

›Adds enforceAvailableModels managed setting that forces the Default model to fall back to the first allowed model when it would otherwise resolve to a disallowed model, and prevents users or projects from widening a managed availableModels list.

v2.1.174 NOTES STABLE

Claude Code v2.1.174 adds wheel-scroll acceleration control and detailed VS Code usage attribution breakdowns.

└──▷ GET THIS VERSION

$ git clone --branch v2.1.174 https://github.com/anthropics/claude-code.git
# already have the repo? check out this version:
$ git checkout v2.1.174

└──▷ HOW TO FIND IT

Review where your token spend is going — broken down by subagents, MCP tools, skills, and context size — without leaving VS Code.

📍In VS Code, open the Claude Code panel, then run /usage to open the Account & usage dialog showing the last 24h or 7d breakdown.

›Adds wheelScrollAccelerationEnabled setting to disable mouse-wheel scroll acceleration in fullscreen mode.
›Adds usage attribution panel in the VS Code Account & usage dialog (/usage) showing cache misses, long context, subagents, and per-skill/agent/plugin/MCP breakdowns over the last 24h or 7d.

Was this useful?

SST OpenCode

v1.17.4 NOTES

The open source coding agent.

OpenCode v1.17.4 adds MCP server cwd support, connector-based auth, and new v2 session API endpoints.

└──▷ GET THIS VERSION

$ git clone --branch v1.17.4 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v1.17.4

└──▷ USE IT

Pin a local MCP server to a specific subdirectory so it resolves paths relative to your project root rather than the shell's working directory.

json

# In your opencode config, set cwd on a local MCP server entry:
{
  "mcpServers": {
    "my-server": {
      "command": "node",
      "args": ["server.js"],
      "cwd": "./tools/my-server"
    }
  }
}

›Adds cwd support for local MCP servers, allowing them to start from a workspace-relative directory.
›Adds connector-based authentication flows with support for stored provider credentials.
›Adds v2 API endpoints to create and fetch sessions, list session questions, and resolve the active location.
›MCP server log notifications are now surfaced in opencode logs.

Was this useful?

Earendil Works Pi

v0.79.2 NOTES

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

Pi v0.79.2 adds an experimental first-time setup flow with theme selection and opt-in analytics behind PI_EXPERIMENTAL=1.

└──▷ GET THIS VERSION

$ git clone --branch v0.79.2 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:
$ git checkout v0.79.2

└──▷ TRY IT

Trigger the first-time setup flow to configure theme and analytics on a fresh install or new agent directory.

$ PI_EXPERIMENTAL=1 pi

›Adds experimental first-time setup flow (enabled via PI_EXPERIMENTAL=1) that prompts for dark/light theme choice and opt-in analytics, storing a trackingId in settings.json.

Was this useful?

Alibaba Qwen Code

v0.18.0 NOTES

Qwen Code v0.18.0 adds Agent Team parallelism, a Workflow tool sandbox, user-level auto-memory, /fork and /stats commands, and a desktop ACP app.

└──▷ GET THIS VERSION

$ git clone --branch v0.18.0 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.18.0

└──▷ TRY IT

Copy the second-to-last AI response to the clipboard — useful when you want an earlier answer without scrolling.

$ /copy 2

Spawn a background sub-agent to work on a parallel task without blocking the current session.

$ /fork

Review cross-session token and request statistics directly in the terminal.

$ /stats

›Adds /fork background-agent command to spawn isolated sub-agents from the CLI.
›Adds /stats interactive dashboard command with cross-session usage tracking.
›Adds /skills picker dialog — browse, search, toggle, and pick skills interactively.
›Adds /copy N command to copy the Nth-last AI message to the clipboard.
›Adds user-level auto-memory stored at ~/.qwen/memories/ so memories persist across projects.

+23 moreshow less

›Adds description field to ExtensionConfig for richer extension metadata.
›Adds declarative agent frontmatter v1 — wires permissionMode, maxTurns, and a color allowlist into agent skill files.
›Enforces auto-skill- directory prefix for all auto-generated skills.
›Adds Workflow tool (Phase 1) — minimal node:vm sandbox with sequential agent() calls for scripted multi-step workflows.
›Adds experimental Agent Team feature for parallel sub-agent coordination.
›Enables loop and cron tools by default so scheduled and looping agents need no manual opt-in.
›Adds InstructionsLoaded hook that fires when instruction files are loaded, enabling prompt-expansion extensions.
›Adds user prompt expansion hooks in core for pre-processing user input before it reaches the model.
›Injects TRACEPARENT environment variable into shell child processes for distributed-tracing continuity.
›Adds terminal sequence notifications support in hooks.
›Adds layered tool-output truncation with per-message budget and per-tool limits to prevent context bloat.
›Adds standalone auto-update support so the CLI self-updates without an external package manager.
›Adds desktop app package with Qwen ACP SDK integration.
›Adds ACP background notification surfacing in the VS Code extension.
›Adds support for desktop Qwen integration via ACP.
›Adds daemon-mode feature batch: background daemon process management.
›Adds retry-visibility telemetry (qwen-code.llm_request span, Phase 4b) for observing LLM request retries.
›Adds sub-agent span telemetry (qwen-code.subagent) with concurrent isolation (Phase 3).
›Adds release-asset verification to the installer and switches public docs to a standalone entrypoint.
›Adds auto-generated CHANGELOG.md synced from GitHub releases.
›Adds @qwen /triage workflow for automated issue and PR triage in CI.
›Adds PR review workflow using the bundled /review skill in CI.
›Honors skill.allowedTools by auto-approving tools declared in a skill's manifest.

Was this useful?

◆ Local LLM Runtimes

KoboldCpp

v1.115.2 NOTES

KoboldCpp v1.115.2 adds Anthropic /v1/messages API, --sdvramlimit for image/video VRAM capping, and MTP/Gemma4 model support.

└──▷ GET THIS VERSION

$ git clone --branch v1.115.2 https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:
$ git checkout v1.115.2

└──▷ TRY IT

Cap GPU VRAM used by image/video generation so large models load on low-VRAM systems by splitting the compute graph.

$ koboldcpp --model mymodel.gguf --sdmodel sdxl.gguf --sdvramlimit 6

Enable MTP speculative drafting with a Gemma4 assistant drafter to boost token throughput and verify hitrate.

$ koboldcpp --model gemma-4-26B-A4B-it-UD-Q4_K_M.gguf --draftmodel gemma-4-26b-A4B-it-assistant-Q4_0.gguf --debugmode

›Adds full support for the Anthropic /v1/messages API, including multimodal and tool calling.
›New --sdvramlimit flag limits VRAM usage for image and video generation by splitting the compute graph, enabling any image gen model to load on constrained hardware.
›New --usemtp flag (and GUI toggle) enables Multi-Token Prediction drafting when built into the model; alternatively, pass a MTP drafter via --draftmodel.
›New --reasoningeffort flag (and GUI toggle) sets the default reasoning effort, using --gendefaults under the hood to apply API defaults.
›Use --debugmode to inspect MTP drafting hitrate during inference.

+6 moreshow less

›Adds support for Gemma4 UV (12B) models, their vision/audio mmprojs, and Gemma assistant models.
›Music generation now falls back to a loaded text LLM when no --musicllm is specified.
›Video generation in SDUI overhauled: supports selecting a specific end frame in Img2Img, specifying both start and end frames via two reference images in txt2img, up to 32 FPS with extended limits enabled, and a lower default VAE tiling threshold of 640.
›Adds new chat templates for LFM2.5 and Cohere2Moe (hotfix 1.115.2).
›Launcher context size default increased to 16k; max prompt length for image generation increased.
›Accepts additional SSL file extensions for HTTPS configuration.

Was this useful?

llama.cpp

b9606 2 RELEASES · 2026-06-12 NOTES STABLE

llama.cpp b9606 adds EAGLE3 speculative decoding support, including Gemma4 EAGLE3 models from RedHatAI.

└──▷ GET THIS VERSION

$ git clone --branch b9606 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b9606

›Adds EAGLE3 speculative decoding support, enabling faster inference via a third-generation draft model architecture, with support for Gemma4 EAGLE3 models from RedHatAI.

└──▷ BREAKING ON UPGRADE

!The output_layer_inp cparams field is renamed to embeddings_layer_inp.

1 more release in this issue · 2026-06-12

b9603 NOTES STABLE

llama.cpp b9603 adds q5_0 and q5_1 GEMM/GEMV OpenCL kernels for Qualcomm Adreno GPUs.

└──▷ GET THIS VERSION

$ git clone --branch b9603 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b9603

›Adds q5_0 and q5_1 GEMM and GEMV OpenCL kernels for Qualcomm Adreno GPUs, enabling accelerated inference with 5-bit quantized models on Adreno hardware.

Was this useful?

Other / Uncategorized

◆ AI OBSERVABILITY

Arize Phoenix

arize-phoenix-v17.5.0 NOTES

Phoenix 17.5.0 adds a subagents toggle in assistant settings and a calendar-based time range picker in the UI.

└──▷ GET THIS VERSION

$ git clone --branch arize-phoenix-v17.5.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v17.5.0

›Adds a subagents toggle to assistant settings, enabling control over whether subagents are used during agent interactions.
›Adds a calendar-based date picker to the time range selector in the UI for more precise trace and data filtering.

Was this useful?

Langfuse

v3.185.0 NOTES

Langfuse v3.185.0 adds conversation starters for agents, monitor bucket live preview, shift-selection in traces, and automatic ClickHouse compatibility detection.

└──▷ GET THIS VERSION

$ git clone --branch v3.185.0 https://github.com/langfuse/langfuse.git
# already have the repo? check out this version:
$ git checkout v3.185.0

›Automatically detects ClickHouse versions on startup and sets compatibility flags without manual configuration.
›Adds conversation starters to the Langfuse agent interface.
›Adds bucket live preview scoped by evaluation window in monitors.
›Raises the organization limit for monitors from the previous cap to 20.
›Enriched blob export (isEnrichedBlobExportAvailable) is now gated on V4 environment variables for self-hosted deployments.

+4 moreshow less

›Batch export now reads sessions from the events table for V4 users.
›Checkboxes in traces and observations now support shift-selection for bulk actions and are more performant.
›Removes the beta label from the corrections feature, marking it generally available.
›Adds a new experimental feature modal to the web UI.

Was this useful?

// END OF TRANSMISSION

The AI Toolchain · curated daily · June 12, 2026
Releases & features from the open-source and commercial AI tools that practitioners actually run.

The AI Toolchain — issue -068, June 12, 2026