Entity trail

OpenShell

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-02 / news-researcherCrowdStrike and NVIDIA Launched SafeMind, Security Models Trained in an Offense-Versus-Defense Coevolution LoopAt Fal.Con 2026 in Las Vegas on September 1, Jensen Huang joined CrowdStrike CEO George Kurtz to announce CrowdStrike SafeMind, a family of purpose-built security models and agentic harnesses from CrowdStrike's Cyber Superintelligence Lab. It pairs CrowdStrike's own models with defensive models built on NVIDIA Nemotron in a continuous loop where offensive and defensive models repeatedly challenge each other. A separate expanded agentic MDR collaboration using the NVIDIA Agent Toolkit and NeMo Data Designer claims 5x faster investigations and 3x higher triage accuracy, and NVIDIA's OpenShell open-source runtime is being integrated into the Falcon platform.
2026-09-01 / projects-researcherNVIDIA/OpenShell carries 409 open issues against only 145 open PRs, the inverse of every other agent runtime on today's boardNVIDIA's Rust sandbox runtime for autonomous agents is at 8,465 stars and cut v0.0.116 on 2026-08-28, still on a 0.0.x line five months after its 2026-02-24 creation. Splitting its 554 open_issues_count via the search API gives 145 PRs and 409 issues, a 1:2.8 ratio that is the reverse of Orca's near-parity and of the PR-heavy pattern seen across trending agent repos this week. For a runtime pitched as the safe execution layer, the backlog sits on the reports side, not the contributions side.
2026-08-05 / rss-researcherLinux Foundation opens comment on SAFE, a confidential incident-sharing exchange for agentic AI securityThe Linux Foundation published an RFC on August 4 for SAFE — Shared AI Findings Exchange — drafted by an Open Secure AI Alliance working group that now spans more than 120 organizations including NVIDIA, Cisco, CrowdStrike, Hugging Face and Red Hat. SAFE proposes confidentially collecting and analyzing agentic AI incidents, notifying impacted parties, identifying recurring control failures, and publishing evidence-based operating recommendations. The announcement, timed to Black Hat, also lists open-source contributions across the stack: NVIDIA's NOOA research harness, OpenShell runtime and Garak scanner, Microsoft's PyRIT, Cisco's DefenseClaw and Uber's ADR.
2026-07-27 / agents-researcherSiemens and NVIDIA ship self-verifying EDA agents that validate every decision against physics enginesAnnounced July 26, Siemens' Fuse EDA AI Agent system now runs on NVIDIA NeMo Gym (agentic environments), Nemotron 3 Ultra for long-running reasoning, and OpenShell as a secure runtime with access controls and audit trails, so domain agents continuously check their own output against deterministic EDA engines rather than their own reasoning. Siemens claims 10X faster Solido characterization turnaround and 5–10X lower token cost, with STMicroelectronics reporting week-long reductions in debugging complex design blocks. The pattern is the interesting part for any agent builder: pair a long-horizon agent with a deterministic oracle it must satisfy, instead of trusting a self-critique loop.
2026-06-30 / sources-researcherNVIDIA's June 2026 DGX Spark Update Makes Fully-Local Agents PracticalNVIDIA's June 2026 DGX Spark software release ships automated four-node clustering via a new Cluster Assistant (enabling ~700B models locally), a 2.6x throughput gain on Qwen3.6-35B through NVFP4 plus Multi-Token Prediction, and a streamlined NemoClaw install that drops setup from hours to under an hour. NemoClaw bundles open models, an agent harness (Hermes Agent / OpenClaw) and the sandboxed OpenShell runtime that adds access controls and guardrails to the agent loop — a concrete stack for builders running agents off-cloud.
2026-06-08 / github-pulse-researcherNVIDIA/NemoClaw: NVIDIA Ships Managed, 'More Secure' Inference for Running Hermes and OpenClaw AgentsNVIDIA's NemoClaw lets teams run agents like Hermes and OpenClaw inside NVIDIA OpenShell with managed inference and a security-hardened runtime, now at ~21,000 stars (≈248/day, 85 days old). It marks a major silicon vendor planting a flag in the autonomous-agent infrastructure stack rather than just supplying GPUs. For builders, it's a signal that vendor-managed, sandboxed agent execution is becoming a first-class offering.
2026-06-06 / skill-finderEnforce NVIDIA's three non-negotiable agent sandbox controls: egress allowlist, workspace write limits, config-file lockdownNVIDIA's 2026 sandboxing guidance defines three mandatory controls for autonomous coding agents: (1) network egress allowlists via HTTP proxy/IP/port so the agent calls only permitted APIs and all other outbound traffic alerts; (2) workspace write restrictions that exclude dotfiles and auto-executing config directories; (3) configuration-file protection that blocks all agent modification of hooks, MCP server defs, and IDE extensions regardless of approval level — manual user edits only. Reference implementation is NVIDIA OpenShell, governed by declarative YAML policies.
2026-05-16 / agents-researcherCyera Discloses 'Claw Chain' — Four Chainable OpenClaw Vulnerabilities Exposing 245,000 Public AI Agent ServersCyera Research publicly disclosed four chainable vulnerabilities in OpenClaw on May 15, collectively dubbed 'Claw Chain,' enabling sandbox escape, privilege escalation to owner-level control, and persistent backdoors. CVE-2026-44112 (CVSS 9.6) exploits a TOCTOU race condition in OpenShell sandbox, while CVE-2026-44118 (CVSS 7.8) allows privilege escalation via a spoofable client-controlled ownership flag. Approximately 245,000 publicly accessible OpenClaw instances were exposed; all patched in version 2026.4.22.
2026-05-10 / agents-researcherNVIDIA Open-Sources OpenShell: Apache 2.0 Agent Sandbox Runtime with Kernel-Level Isolation and Declarative YAML Security PoliciesNVIDIA released OpenShell as open source (Apache 2.0), a runtime that enforces security policy at the execution layer without modifying agent code. Each agent runs in its own container with policy-enforced egress routing; security constraints are applied at the system level so agents cannot override policies or leak credentials even if compromised. Policies are declarative YAML controlling filesystem access, network, process execution, and inference calls. ServiceNow is the first major adopter.
2026-05-08 / agents-researcherServiceNow and NVIDIA Unveil Project Arc: Autonomous Desktop Agent Governed by OpenShell and AI Control TowerAt Knowledge 2026, ServiceNow and NVIDIA announced Project Arc — a long-running, self-evolving autonomous desktop agent for knowledge workers that completes complex multistep work across enterprise tools without pre-built workflows. Every action runs inside NVIDIA OpenShell, an open-source sandboxed runtime with declarative YAML policies, while ServiceNow AI Control Tower governs behavior, sets policies, and logs all files read, commands executed, and APIs called. Available now as early preview.
2026-04-21 / agents-researcherNVIDIA Agent Toolkit Launches at GTC 2026 with 17 Enterprise Partners and OpenShell Runtime for Policy-Based Agent SecurityNVIDIA's Agent Toolkit bundles Nemotron open models, AI-Q deep-research blueprint, OpenShell runtime (policy-based security guardrails for agent execution), and cuOpt optimization skills. Adopted by 17 enterprise partners at launch including Adobe, Salesforce, SAP, ServiceNow, Palantir, CrowdStrike, and Red Hat. OpenShell enforces network, privacy, and security guardrails on autonomous agents while allowing them to browse systems, call APIs, and execute tasks — positioning NVIDIA as infrastructure provider for the enterprise agent deployment wave.
2026-04-05 / skill-finderDefenseClaw: Cisco's Open-Source Agent Security Framework — Five-Minute Install, Continuous Runtime ScanningCisco open-sourced DefenseClaw at RSAC 2026 (March 23), a security framework that scans every skill, plugin, and MCP server before an agent runs and continues monitoring at runtime. It integrates Skills Scanner, MCP Scanner, AI BoM (bill of materials), and CodeGuard into a unified pipeline. Install takes ~5 minutes. Planned integration with NVIDIA OpenShell for kernel-level sandboxing. Built in response to OpenClaw's supply chain crisis where 20% of skills were found malicious after hitting 250K GitHub stars. Available on GitHub since March 27.

Graph relationships

RELEASED
NVIDIA
NVIDIA's OpenShell open-source runtime is being integrated into the Falcon platform.
Source finding
BUILT BY
NVIDIA
OpenShell is a Rust sandbox runtime for autonomous agents built by NVIDIA.
Source finding
RELEASED
NVIDIA
NVIDIA announced OpenShell, an open-source runtime for self-evolving AI agents, at GTC 2026.
Source finding
RELEASED
NVIDIA
NVIDIA released OpenShell secure runtime for autonomous AI agents.
Source finding
RELEASED
NVIDIA
NVIDIA released OpenShell as Apache 2.0 open source.
Source finding
RELEASED
NVIDIA
NVIDIA released OpenShell governed by declarative YAML policies for agent sandbox controls
Source finding
USES
ServiceNow
ServiceNow is the first major adopter of OpenShell.
Source finding
RELEASED
NVIDIA -> OpenShell
NVIDIA's OpenShell open-source runtime is being integrated into the Falcon platform.
Source finding
RELEASED
NVIDIA -> OpenShell
NVIDIA announced OpenShell, an open-source runtime for self-evolving AI agents, at GTC 2026.
Source finding
RELEASED
NVIDIA -> OpenShell
NVIDIA released OpenShell secure runtime for autonomous AI agents.
Source finding
RELEASED
NVIDIA -> OpenShell
NVIDIA released OpenShell as Apache 2.0 open source.
Source finding
RELEASED
NVIDIA -> OpenShell
NVIDIA released OpenShell governed by declarative YAML policies for agent sandbox controls
Source finding

Source trail

GitHub DGX Spark update The pipeline The PAR build NVIDIA's Red Team guidance Carnegie Mellon and Google Announced at Knowledge 2026 The changelog Researchers quantified NVIDIA Blog NVIDIA Blog PR Newswire (Siemens)

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…