ironcurtain

IronCurtain Un runtime seguro para agentes de IA autónomos, donde la política de seguridad se deriva de una constitución legible por humanos. Cuando alguien escribe "seguro", deberías ser escéptico de inmediato.¿Qué queremos decir con seguro? !WARNING Prototipo de investigación. IronCurtain es un...

polaris

Polaris 是 Kubernetes 的一个开源策略引擎 Polaris 是 Kubernetes 的一个开源策略引擎，用于验证和修复资源配置。它包含 30 多个内置配置策略，并支持使用 JSON Schema 构建自定义策略。当以命令行模式或作为突变 Webhook 运行时，Polaris 能够根据策略标准自动修复问题。 Polaris 可以在三种不同模式下运行： 作为仪表盘 - 根据策略即代码验证 Kubernetes 资源。 作为准入控制器 - 自动拒绝或修改不符合组织策略的工作负载。 作为命令行工具 - 将策略即代码集成到 CI/CD 流程中，用于测试本地 YAML 文件。...

gate0

gate0 Un pequeño motor de micropolíticas auditable, terminante, determinista. ¿Por qué Gate0? Construí Gate0 porque estaba harto de depurar políticas basadas en RegEx en producción. Quería algo que fuera aburrido, acotado e imposible de bloquear. Si necesitas un motor de políticas flexible y de...

clawshield-public

ClawShield Proxy de seguridad para agentes de IA. Se sitúa delante de OpenClaw y escanea cada mensaje en busca de inyección de instrucciones, fugas de PII y secretos, antes de que lleguen al modelo o salgan de la red. Incluye 5 agentes de IA especializados, un panel de control integrado y un moto...

clawdstrike

Clawdstrike 为群智时代打造的EDR。 默认拒绝，签名真相。 状态：pre-1.0 beta 。公开API已稳定；默认值在1.0之前可能进一步收紧。 Clawdstrike 是一个策略引擎、一个EDR，以及一个签名审计链，三者合为一体。AI 代理的 toolcall 与内核级的 fileaccess、processexec、networkflow、dylibload 或 launchpersistence 在同一事件分类体系中。一个策略引擎评估它们。一个 Ed25519 签名的因果图记录它们。默认行为是拒绝（fail closed）。 同一个引擎以 Rust...

opa

Open Policy Agent Open Policy Agent OPA es un motor de políticas de propósito general, de código abierto, que permite la aplicación unificada de políticas conscientes del contexto en toda la pila tecnológica. OPA se enorgullece de ser un proyecto graduado en el panorama de la Cloud Native Computi...

AgentGuard

AgentGuard Sandbox de confianza cero para agentes de IA autónomos. AgentGuard envuelve cualquier agente de IA LangChain, CrewAI, AutoGen, scripts personalizados con barreras de seguridad. Con un solo cambio en el comando: Antes peligroso — el agente tiene acceso completo al sistema python...

mcp-doorman

mcp-doorman La puerta de seguridad para servidores MCP: cada llamada a herramienta se verifica en la entrada. CI npm Licencia: Apache-2.0 PRs bienvenidos mcp-doorman es un proxy plug-and-play que se sitúa entre tu agente de IA Claude Desktop, Claude Code, Cursor, VS Code, cualquier cliente MCP y...

CVE-2026-12624

A flaw was found in Vault's Access Control List ACL policy engine. The engine did not consistently enforce wildcard deny rules against LIST requests with a trailing slash. This vulnerability could allow an authenticated attacker with specific permissions to enumerate the names of entries they wer...

Nexora

Nexora Nexora 是一个由 AI 驱动的企业端点治理与安全智能系统，可跨机器监控软件使用情况、实施基于策略的风险分类，并生成结构化的行为与安全报告。 它专为需要以下能力的组织而设计：掌握应用程序使用情况、执行软件治理规则，以及通过确定性策略规则结合 AI 辅助分析，检测风险或未经授权的工具行为。 当 AI 成为攻击者时，Nexora 就是防御者。 概述 Nexora 以轻量级端点代理结合策略引擎与 AI 推理层的方式运行。它将原始进程与应用程序活动转化为结构化安全智能，但不会赋予 AI 直接执行权限。 核心设计原则： 确定性策略执行（非 AI 驱动的风险决策） 透明遥测采集 仅...

Sapien: A Stateful Policy Engine for Autonomous AI Agents

Contextual security defenses prevent AI agents from taking rogue actions by synthesizing a task-specific policy and enforcing it on the agent's tool calls. In multi-step tasks, however, which actions are valid often depends on what the agent has already done and learned. We present Sapien, a poli...

opa v1.20.2

Open Policy Agent Open Policy Agent OPA is an open source, general-purpose policy engine that enables unified, context-aware policy enforcement across the entire stack. OPA is proud to be a graduated project in the Cloud Native Computing Foundation CNCF landscape. For details read the CNCF...

CVE-2026-54523

Kyverno is a policy engine designed for cloud native platform engineering teams. From 1.18.0 until 1.18.2, the NamespacedMutatingPolicy CEL compiler exposes the generator library to matchConditions, allowing a namespace-scoped policy to invoke generator.applynamespace, resources with an arbitrary...

GO-2025-3615 vulnerabilities

Vulnerabilities for packages: commercial-kyverno, kyverno-fips, kyverno...

GO-2025-3562 vulnerabilities

Vulnerabilities for packages: commercial-kyverno, kyverno-notation-aws, kyverno, kyverno-notation-aws-fips...

GO-2025-3652 vulnerabilities

Vulnerabilities for packages: kyverno-notation-aws...

GO-2026-5575 vulnerabilities

Vulnerabilities for packages: kyverno-notation-aws...

PT-2026-71092

RustFS is a distributed object storage system built in Rust. Prior to 1.0.0-beta.12, RustFS evaluates the ForAllValues: and ForAnyValue: set qualifiers with the negated string operators StringNotEquals, StringNotEqualsIgnoreCase, StringNotLike, ArnNotEquals, and ArnNotLike using each other's...

CVE-2026-12624

Vault’s ACL policy engine allowed LIST requests with a trailing slash on denied paths to bypass a narrow wildcard deny when a token had a broad allow. This could enable enumeration of names under a denied path. Affected products include Vault Community Edition, with fixes in 2.0.3, and Vault Ente...

CVE-2026-12624: Incorrect Authorization

Vault’s ACL policy engine did not consistently enforce a wildcard glob deny rule against LIST requests made with a trailing slash on the denied path. This may allow a token holding a broader allow rule alongside a narrower wildcard deny rule to enumerate the names of entries beneath a path it was...