vibehacker
News
Factory Blog ·

Factory ships Droid 1.0, making scripted tool calls the default so one request can run hundreds of tools

Released this week, Droid 1.0 lets the model write a sandboxed JavaScript Script that runs many tool calls in one request (Factory says it cuts cost 15% on Opus 5.5), trims the base prompt 37%, and ships a @factory/droid-sdk for building on the same harness. On Factory's own Terminal-Bench 4 runs it beat Codex, Claude Code, Cursor, OpenCode and Devin on the same models at lower cost, but those numbers are self-reported.

More news

View all

DeepSeek Harness alpha adds an experimental tool that hands tasks to your local Codex or Claude Code

Released Oct 28 as pre release dsh v0.2.2 alpha.2, the experimental subagent oar plugin delegates a self contained task to an installed Codex or Claude Code CLI, which works in its own context and returns one final answer with no follow ups or resume. The alpha also lets you update third party plugins with rollback if a hot update fails, and adds six UI languages…

LangChain's Managed Deep Agents v0.10 adds browser use and async subagents that run in the background

Released Oct 29 in public beta on LangSmith Cloud (US), agents can now work on sites with no API and hand you the browser window at sign in so passwords stay out of traces, and a subagent marked async runs long jobs while the main agent keeps chatting. It also replies in the originating Slack thread with a Stop button, and drops connectors.mcp() in favor of declaring MCP servers in tools/mcp.py or tools/mcp.ts…

LangChain Blog

Kimi Code 2.4.0 imports your recent Claude Code and Codex sessions so you can keep working in Kimi

Released Oct 28, kimi import sessions (or /import sessions ) converts up to 100 Claude Code and Codex sessions from the last 15 or 30 days into regular Kimi sessions in their original project folders, locally and without calling a model or touching the originals. The release also makes Plan mode user only, so the agent can't enter it or auto approve plans on its own (changelog)…

Spotted something we missed? Start a thread.