MachinaOS — From the Workshop https://machinaos.ai/blog.html Notes on local-first AI, governed agent runtimes and LLM engineering, from the team building MachinaOS. en Wed, 01 Jul 2026 00:00:00 +0000 Most AI agent frameworks have the safety model backwards https://machinaos.ai/blog-agent-frameworks-safety-model-backwards.html https://machinaos.ai/blog-agent-frameworks-safety-model-backwards.html Wed, 01 Jul 2026 00:00:00 +0000 Risk-gated execution and a hash-chained audit log — the governance model behind MachinaOS, a local-first agent runtime.