OpenClaw

Evidence (7)

• 5Simon Willison·8/10/2026security
  Quote about an authorization flaw in OpenClaw

  Simon Willison quoted a post describing a serious authorization flaw on an Australian gym-booking website called OpenClaw. The report says the API allowed cancelling other users' reservations without authorization, and the issue was reproducible in testing.

• 5Hugging Face Blog·7/15/2026framework_update
  Lessons from building the Shippy maritime AI agent

  Ai2 describes how it built Shippy, a maritime AI agent designed for high-stakes operational use in real-time ocean monitoring. The article focuses on reliability, versioned agent design, and skills that connect the agent to live Skylight data, boundary APIs, and map links while limiting the system from making legal judgments or unsupported claims.

• 7The New Stack·7/9/2026open_source
  OpenClaw becomes an independent non-profit foundation

  OpenClaw has formed a non-profit foundation to govern the open-source AI agent after its creator joined OpenAI. The foundation says it aims to keep the project neutral, personal, and self-hosted, with a small full-time team now in place and major corporate sponsors still involved.

• 5The New Stack·7/1/2026product_launch
  OpenClaw’s app uses a persistent AI agent instead of running on-device

  OpenClaw has released iOS and Android apps that let users interact directly with a personal AI agent, but the agent itself runs in a separate persistent runtime rather than on the phone. The design turns the phone into an authenticated endpoint for voice, notifications, camera access, and action approval, reflecting a broader industry shift toward distributed agent architectures.

• 4Simon Willison·6/26/2026security
  2,000 Users Failed to Hack an AI Assistant

  Fernando Irarrázaval ran a public challenge on hackmyclaw.com to see whether participants could extract secrets from his OpenClaw test instance by emailing it. After 6,000 attempts and notable token costs, no one succeeded in leaking the secret, despite a prompt-injection setup designed to test the assistant's defenses. The result suggests frontier-model safety training is making these attacks harder, but it does not prove such systems are safe for high-risk production use.

• 5Hugging Face Blog·6/22/2026open_source
  Local models triage OpenClaw issues for free

  Hugging Face describes using local open-weight models in an agent harness to triage and classify issues and pull requests in the OpenClaw repository. The setup aims to provide near-real-time notifications and reduce reliance on paid closed models by running on local hardware with read-only repo access.

• 8NVIDIA GenAI·6/2/2026framework_update
  NVIDIA Unveils NemoClaw for Secure Autonomous Industrial AI Agents

  NVIDIA introduced NemoClaw, an open blueprint for building long-running autonomous AI agents with secure runtime controls, model routing, and customization tools. Industrial software vendors including Cadence, Dassault Systèmes, Siemens, Synopsys, and startups such as Flexcompute, Luminary, and Neural Concept are using it to automate design, simulation, verification, and reporting workflows across CAE and EDA.