RepoJournal
Agentic Coding Agentic Coding
110 wires and counting

$ follow Agentic Coding

Keep up with Agentic Coding in about 3 minutes: what actually shipped — the commits, pull requests, releases, and security advisories that matter.

or

fair warning: these emails are deeply technical. diffs, version numbers, CVEs, benchmark deltas. if that's not your idea of a good read, this isn't your newsletter.

Folds into your digest — weekly by default, monthly if you prefer. Unsubscribe in one click.

$ status

wire 2026-11-01
stories 83

© 2026 RepoJournal Home Showcase How it works Privacy

$ the-wire · showcase

Azure Claude routing in opencode, Canvas lag fixes in OpenHands

By RepoJournal · Filed · About Agentic Coding · Composed from the cited sources · methodology

opencode can now talk to Claude on Azure AI Foundry over Anthropic Messages, and OpenHands spent the day cutting the wait operators feel in Canvas: upload failures, reply rendering, and a startup poll that fired too slowly for short boots.

feat(ai): add Azure Messages for Foundry Claude models anomalyco/opencode

by rekram1-node

The Azure provider previously offered only Responses and Chat, so the 15 Claude models on Azure AI Foundry in models.dev were pointed at an anthropic-compatible route that knew nothing about your Azure resource URL or auth. Azure.configure(...).messages(id), Azure.messagesModel and the @opencode/ai/providers/azure/messages entry point now give those models a native Messages path, so if you prov...

fix: upload Cloud attachments to the runtime workspace OpenHands/OpenHands

by ak684

OHE uploads could land in the default directory because Cloud app metadata omits the conversation workspace, so the runtime rejected the file and the first message went out without it. Matched browser recordings show the upload moving from HTTP 422 to HTTP 200 with startup-check.txt included in the sent message, which means attaching a file to a Cloud conversation should no longer silently drop...

fix: display received replies without a reading-speed delay OpenHands/OpenHands

by ak684

Canvas kept revealing received text long after it arrived, including after the final reply. In a controlled 201-character burst measured in the running app, display delay fell from 1,457 ms to 129 ms (91% less, 1,328 ms saved), a rendering figure and not a change to model generation or end-to-end latency.

fix: recover transient conversation failures without repeated errors OpenHands/OpenHands

by ak684

Startup readiness polling now runs at 500 ms for the first ten seconds before reverting to 3,000 ms, cutting the early wait by up to 2.5 seconds; the author is explicit that this is a scheduling bound, not a measured sandbox-speed gain. In the live OHE fault test the session went from stuck on the startup page after 15 seconds to an assistant reply visible in 7.21 seconds.

fix(tui): keep keymap command objects while their data is unchanged anomalyco/opencode

by simonklee

The TUI rebuilt every keymap layer on each keystroke because createLayer handed back fresh command objects, so layers were re-registered constantly while data was unchanged; identical objects are now reused and replaced only when a field changes or disappears, keeping labels and slash entries current. The rest of the day is long tail: two dialog and menu geometry naming passes and schema-copy f...

Quick answers

What shipped in Agentic Coding on November 1, 2026?
opencode can now talk to Claude on Azure AI Foundry over Anthropic Messages, and OpenHands spent the day cutting the wait operators feel in Canvas: upload failures, reply rendering, and a startup poll that fired too slowly for short boots. In total, 42 commits, 40 pull requests, and 1 releases landed.
Who contributed to Agentic Coding on November 1, 2026?
7 developers shipped this update, including Dax Raad, simonklee, rekram1-node, kitlangton, nexxeln, ak684, and enyst.
What were the notable Agentic Coding updates?
feat(ai): add Azure Messages for Foundry Claude models, fix: upload Cloud attachments to the runtime workspace, and fix: display received replies without a reading-speed delay.