vibehacker
News
GitHub (huggingface/safeagents) ·

Hugging Face's safeagents is a local proxy that blocks risky coding-agent tool calls before they run

Released this week under Apache 2.0, it sits between your agent and its model server on 127.0.0.1:8787, blocking calls to undeclared tools or to hosts nobody mentioned in the conversation and running about 110 atif-scan detectors for things like reading /proc/1/environ or curl --noproxy, with Claude Code, OpenCode, Pi and three other harnesses tested in CI. Start with --log-only, and note it can't see URLs a script builds at runtime, so keep default-deny egress in your sandbox (Tensor Journal).

Read at github.com
Claude Code
An AI coding agent for terminal, IDE, web, and Slack

More news

View all

Gemini API's Computer Use tool is now generally available for browser, mobile and desktop agents

Since Oct 30, the computer use tool is GA in both the Interactions and GenerateContent APIs: your loop sends a screenshot, gets back a UI action with an intent explaining it, executes it (e.g. with Playwright) and repeats. Responses can carry a safety decision of require confirmation or blocked, and Google still says to run it in a sandboxed VM or container (docs)…

Gemini API Release Notes

Ramen open-sources Aura Autopilot, an agent that playtests Unreal, Unity and Godot games in real time

Launched Oct 30 under Apache 2.0, it runs as an editor plugin for Unreal Engine 5, Unity 6 and Godot 4 on Windows and macOS, plays your game with real inputs from prompts like "verify the tree shakes when you hit it with an axe", writes a report, and learns each game as it goes. It takes your own API key, or autopilot connect claude / autopilot connect codex lets Claude Code or Codex drive it on your existing subscription (GitHub)…

Business Wire

Microsoft's open-source TokenEconomizer plugin cuts Claude Code and Codex costs by learning from past sessions

The Apache 2.0 plugin from Microsoft Research and UC Berkeley works with Claude Code, Codex, Copilot CLI and Pi: it prunes tools a repo never uses, redirects turns stuck in failure loops, and can optionally inject learned repo facts, cutting mean SWE Bench cost 18.4–30.9% in the paper's runs. It needs Python 3.10+ and no API key of its own, and everything it learns stays in /.tokeneconomizer (MarkTechPost)…

GitHub (microsoft/TokenEconomizer)

Radar 1.19 lets you ask a question to start a Kubernetes AI investigation with your own Claude Code or Codex

Released Oct 29 alongside its Product Hunt launch, the open source Kubernetes UI now starts read only root cause investigations from a typed question or the Home cluster card, driving your local Claude Code, Codex, Cursor or OpenCode CLI with no API key or account ( radar diagnose ). Its built in MCP server also works with any agent, and radar mcp read only hides the write tools (docs)…

GitHub (skyhook-io/radar)

Spotted something we missed? Start a thread.