Deploy Al code with confidence using Daytona's lightning-fast infrastructure. 90ms environment creation, stateful operations, and enterprise-grade security.
Composio
Just-in-time tool calls, secure delegated auth, sandboxed environments, and parallel execution across 1,000+ apps.
Genode - Genode Operating System Framework
We understand the complexity of code and policy as the most fundamental security problem shared by modern general-purpose operating systems. Because of high functional demands and dynamic workloads, however, this complexity cannot be avoided. But it can be organized. Genode is a novel OS architecture that is able to master complexity by applying a strict organizational structure to all software components including device drivers, system services, and applications. The Genode OS framework is an open-source tool kit for building highly secure component-based operating systems. It scales from embedded devices to dynamic general-purpose computing.
Code Worth Writing - Ray Myers | SSW 2026
mayankjain0141/nixis
AI agent firewall that intercepts tool calls (file, shell, network) and enforces deterministic policies at sub-microsecond latency using CEL, IFC, secret scanning, and audit logging.
ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?
AI agents are rapidly gaining capabilities that could significantly reshape cybersecurity, making rigorous evaluation urgent. A critical capability is exploitation: turning a vulnerability, which is not yet an attack, into a concrete security impact, such as unauthorized file access or code execution. Exploitation is a particularly challenging task because it requires low-level program reasoning (e.g., about memory layout), runtime adaptation, and sustained progress over long horizons. Meanwhile, it is inherently dual-use, supporting defensive workflows while lowering the barrier for offense. Despite its importance and diagnostic value, exploitation remains under-evaluated. To address this gap, we introduce ExploitGym, a large-scale, diverse, realistic benchmark on the exploitation capabilities of AI agents. Given a program input that triggers a vulnerability, ExploitGym tasks agents with progressively extending it into a working exploit. The benchmark comprises 898 instances sourced from real-world vulnerabilities across three domains, including userspace programs, Google's V8 JavaScript engine, and the Linux kernel. We vary the security protections applied to each instance, isolating their impact on agent performance. All configurations are packaged in reproducible containerized environments. Our evaluation shows that while exploitation remains challenging, frontier models can successfully exploit a non-trivial fraction of vulnerabilities. For example, the strongest configurations are Anthropic's latest model Claude Mythos Preview and OpenAI's GPT-5.5, which produce working exploits for 157 and 120 instances, respectively. Notably, even with widely used defenses enabled, models retain non-trivial success rates. These results establish ExploitGym as an effective testbed for exploitation and highlight the growing cybersecurity risks posed by increasingly capable AI agents.
ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?
AI agents are rapidly gaining capabilities that could significantly reshape cybersecurity, making rigorous evaluation urgent. A critical capability is exploitation: turning a vulnerability, which is not yet an attack, into a concrete security impact, such as unauthorized file access or code execution. Exploitation is a particularly challenging task because it requires low-level program reasoning (e.g., about memory layout), runtime adaptation, and sustained progress over long horizons. Meanwhile, it is inherently dual-use, supporting defensive workflows while lowering the barrier for offense. Despite its importance and diagnostic value, exploitation remains under-evaluated. To address this gap, we introduce ExploitGym, a large-scale, diverse, realistic benchmark on the exploitation capabilities of AI agents. Given a program input that triggers a vulnerability, ExploitGym tasks agents with progressively extending it into a working exploit. The benchmark comprises 898 instances sourced from real-world vulnerabilities across three domains, including userspace programs, Google's V8 JavaScript engine, and the Linux kernel. We vary the security protections applied to each instance, isolating their impact on agent performance. All configurations are packaged in reproducible containerized environments. Our evaluation shows that while exploitation remains challenging, frontier models can successfully exploit a non-trivial fraction of vulnerabilities. For example, the strongest configurations are Anthropic's latest model Claude Mythos Preview and OpenAI's GPT-5.5, which produce working exploits for 157 and 120 instances, respectively. Notably, even with widely used defenses enabled, models retain non-trivial success rates. These results establish ExploitGym as an effective testbed for exploitation and highlight the growing cybersecurity risks posed by increasingly capable AI agents.
Composable, fast, and secure dev environments | Workshop | Ubuntu
Launch agent-ready, sandboxed development environments with a single command. Define them via simple YAML configs, share them to recreate on different machines.
OpenClaw – NEAR AI
Run the internet’s favorite new AI agent with NEAR AI’s cryptographic privacy guarantees.
Why AVF?
AVF and pKVM: next-generation Trustworthy Execution Environment for connected products
Introducing eve
Introducing eve, the open-source agent framework from Vercel for building, running, and scaling agents in production, with durable execution, sandboxed compute, approvals, channels, tracing, and evals built in.
Frida • A world-class dynamic instrumentation toolkit
Observe and reprogram running programs on Windows, macOS, GNU/Linux, iOS, watchOS, tvOS, Android, FreeBSD, and QNX
AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents
Agents built on LLMs are increasingly deployed across diverse domains, automating complex decision-making and task execution. However, their autonomy introduces safety risks, including security vulnerabilities, legal violations, and unintended harmful actions. Existing mitigation methods, such as model-based safeguards and early enforcement strategies, fall short in robustness, interpretability, and adaptability. To address these challenges, we propose AgentSpec, a lightweight domain-specific language for specifying and enforcing runtime constraints on LLM agents. With AgentSpec, users define structured rules that incorporate triggers, predicates, and enforcement mechanisms, ensuring agents operate within predefined safety boundaries. We implement AgentSpec across multiple domains, including code execution, embodied agents, and autonomous driving, demonstrating its adaptability and effectiveness. Our evaluation shows that AgentSpec successfully prevents unsafe executions in over 90% of code agent cases, eliminates all hazardous actions in embodied agent tasks, and enforces 100% compliance by autonomous vehicles (AVs). Despite its strong safety guarantees, AgentSpec remains computationally lightweight, with overheads in milliseconds. By combining interpretability, modularity, and efficiency, AgentSpec provides a practical and scalable solution for enforcing LLM agent safety across diverse applications. We also automate the generation of rules using LLMs and assess their effectiveness. Our evaluation shows that the rules generated by OpenAI o1 achieve a precision of 95.56% and recall of 70.96% for embodied agents, successfully identify 87.26% of the risky code, and prevent AVs from breaking laws in 5 out of 8 scenarios.
Backend infrastructure - Tinfoil Documentation
This page provides a description of the different components that make up our backend infrastructure. It also describes how Tinfoil guarantees code auditability and data confidentiality using these components.
Dagger.io
Build powerful software environments and containerized operations from modular components and simple functions. Perfect for complex software delivery and AI agents. Built by the creators of Docker.
Self-driving infrastructure
At Vercel, we’re building self-driving infrastructure, a system that autonomously manages production operations, improves application code using real-world insights, and learns from the unpredictable nature of production itself.
