First screen, captured 3 days ago.

OpenShell – A safe, private runtime for fleets of autonomous AI agents

What HN thinks

Everyone agrees the agents must be caged; nobody agrees on the cage, and a few would like the cage to stop phoning home.

Exfiltration is the unsolved problem
Network allowlists and credential swapping help, but any outbound channel can leak data, so commenters discuss splitting capabilities across agents (Lethal Trifecta) or going fully offline.
Just use VMs or containers
Many already isolate agents in VMs, microVMs or containers and question whether OpenShell adds much beyond policy, MITM credential injection and orchestration.
Default telemetry is a red flag
Opt-out 'anonymous' telemetry drew criticism, with some arguing anonymity is impossible, while others noted it is documented and easy to disable.

Consensus: Agents should run sandboxed with restricted network and no direct access to real secrets, but no current tool fully prevents data exfiltration.

Recommended in the comments

iron proxy (tool): Solved domain restrictions and credential swapping in one go, including SigV4 re-signing.
matchlock ↗ (tool): Combines real VM isolation with secret placeholders so agents never see keys.
Lima ↗ (tool): Makes running agents in a VM painless.
smolvm ↗ (tool): Fast boot, egress control and VM forking (recommended by its author).
OrbStack (product): Suggested as the best option for macOS users.
sandbox-runtime ↗ (tool): Simple config with credential masking strikes a good security/convenience balance.

Summarised by Claude from the comment thread 3 days ago.

Points: 483
Comments: 209
Best rank: #2
Posted: 2026-10-22 by niemandhier
Category: Ai ml

nvidia ai-agents sandboxing rust agent-runtime open-source

All captures:

Full page

Captured 3 days ago. Click for the whole screenshot.

Looks similar

OpenAI CLI (github.com)

203 points by armcat 4 days ago | 110 comments

GitHub repository for openai-cli, OpenAI's official command-line interface for the OpenAI API, written mainly in Go and licensed Apache-2.0.

openai cli openai-api go

Screenshot of github.com: Show HN: Making a smart TV dumb and then smart

Show HN: Making a smart TV dumb and then smart (github.com)

69 points by animalhot 3 days ago | 38 comments

GitHub repo for r2d2-tv, a local home screen, private remote browser, and push-to-talk voice assistant that replaces vendor software on rooted LG webOS TVs.

webos smart-tv lg voice-assistant

Screenshot of github.com: PyTorch-native TPU back end for vLLM

PyTorch-native TPU back end for vLLM (github.com)

4 points by zniturah 6 days ago | 0 comments

GitHub repository for vllm-torchtpu, a PyTorch-native TPU backend for the vLLM LLM inference engine.

vllm tpu pytorch llm-inference

Screenshot of github.com: Show HN: Omnara – Open-source alternative to Claude Managed Agents

Show HN: Omnara – Open-source alternative to Claude Managed Agents (github.com)

85 points by kmansm27 4 days ago | 34 comments

GitHub repository for Omnara, an open-source, self-hostable managed agent platform for running durable AI agents with sandboxes, alternative to Claude Managed Agents.

ai-agents omnara durable-execution claude-managed-agents

Screenshot of github.com: One Must Fall 2097 Remastered

One Must Fall 2097 Remastered (github.com)

38 points by hmm37 2 days ago | 21 comments

GitHub repository for a TypeScript/WebGL2 remaster of the 1994 robot fighting game One Must Fall 2097, with HD art, mods, LAN play and desktop/browser builds.

one-must-fall-2097 fighting-game remaster typescript

Screenshot of github.com: Show HN: Farol – a risk map of a pull request for the person who approves it

Show HN: Farol – a risk map of a pull request for the person who approves it (github.com)

49 points by ericholiveira 1 day ago | 28 comments

GitHub repository for Farol, an open-source CLI and GitHub Action that scores pull request merge risk, highlighting risky files, reach, diff coverage and contract breaks.

code-review pull-requests github-actions static-analysis