Backfill issue -060 — 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   # 2 tools matched

AI & LLM Tooling

◆ AI Coding Agents

Earendil Works Pi

v0.79.9 NOTES

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

Pi v0.79.9 adds chat-template thinking support for vLLM/Hugging Face models like DeepSeek via OpenAI-compatible providers.

└──▷ GET THIS VERSION

$ git clone --branch v0.79.9 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:
$ git checkout v0.79.9

›Supports chat_template_kwargs-based thinking controls for OpenAI-compatible custom providers, enabling vLLM/Hugging Face chat-template models (e.g. DeepSeek) to use provider-native thinking levels.

Was this useful?

Alibaba Qwen Code

v0.18.4 NOTES

Qwen Code v0.18.4 adds QQ Bot channel, token-usage stats, configurable ACP timeouts, daemon idle detection, and a second-resolution session wakeup engine.

└──▷ GET THIS VERSION

$ git clone --branch v0.18.4 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.18.4

└──▷ TRY IT

Check whether the Qwen Code daemon is idle before scheduling a new task in CI or an orchestration script.

$ curl -s 'http://localhost:<port>/health?deep=true'

›Adds GET /health?deep=true endpoint with daemon idle detection for health-check integrations that need to know when the daemon is truly inactive.
›Makes ACP permission timeout configurable in the serve layer, allowing operators to tune how long the agent waits for permission grants before timing out.
›Passes the original API call ID (toolCallId) to the hook system, giving hook authors a stable handle to correlate tool invocations across the pipeline.
›Exposes token usage statistics for cost visibility, surfacing per-run input/output token counts in the UI.
›Adds a second-resolution session wakeup engine for the loop, enabling tighter scheduling of self-paced prompt-only /loop wakeups.

+6 moreshow less

›Adds a QQ Bot (QQ机器人) channel adapter, extending Qwen Code's messaging-platform integrations.
›Adds workflow tool token budget tracking with per-run UI surfacing, making token consumption visible at the workflow level.
›Auto-reveals the exit_plan_mode tool when the agent enters plan mode, reducing manual steps to leave plan mode.
›Adds a desktop-pet skill for generating pixel-art companion applications.
›Adds i18n support for extension displayName and description, enabling localized extension metadata.
›Localizes tool display names in TUI and web-shell badges via i18n.

Was this useful?

// END OF TRANSMISSION

The AI Toolchain · curated daily · June 20, 2026
Releases & features from the open-source and commercial AI tools that practitioners actually run.

The AI Toolchain — issue -060, June 20, 2026

Earendil Works Pi

Alibaba Qwen Code