CVE-2026-20701

An access issue was addressed with additional sandbox restrictions. This issue is fixed in macOS Sequoia 15.7.5, macOS Sonoma 14.8.5, macOS Tahoe 26.4. An app may be able to connect to a network share without user consent...

Apple macOS 安全漏洞

Apple macOS is a specialized operating system developed by Apple for Mac computers. Apple macOS has a security vulnerability that can be exploited by an attacker to cause an application to connect to a network share without the user's consent...

PT-2026-27552

Name of the Vulnerable Software and Affected Versions macOS Sequoia versions prior to 15.7.5 macOS Sonoma versions prior to 14.8.5 macOS Tahoe versions prior to 26.4 Description An issue involving insufficient sandbox restrictions existed, potentially allowing an application to connect to a netwo...

AgentRFC: Security Design Principles and Conformance Testing for Agent Protocols

AI agent protocols -- including MCP, A2A, ANP, and ACP -- enable autonomous agents to discover capabilities, delegate tasks, and compose services across trust boundaries. Despite massive deployment MCP alone has 97M+ monthly SDK downloads, no systematic security framework for these protocols...

When the Abyss Looks Back: Unveiling Evolving Dark Patterns in Cookie Consent Banners

To comply with data protection regulations such as the EU General Data Protection Regulation GDPR and the California Consumer Privacy Act CCPA, websites widely deploy cookie consent banners to collect users' privacy preferences. In practice, however, these interfaces often embed dark patterns tha...

Could your face change what you pay? NYC wants limits on biometric tracking

New York City lawmakers are pushing to ban private businesses from using biometric tools like voice and facial recognition software to track the public. While the desire to use surveillance technology in stores to fight shoplifting is understandable, lawmakers and privacy advocates are worried th...

PT-2026-26788

Name of the Vulnerable Software and Affected Versions Ory Hydra affected versions not specified Description The Admin APIs – listOAuth2Clients, listOAuth2ConsentSessions, and listTrustedOAuth2JwtGrantIssuers – in Ory Hydra are susceptible to SQL injection due to flaws in the pagination...

Claude Code has a Workspace Trust Dialog Bypass via Repo-Controlled Settings File

Claude Code resolved the permission mode from settings files, including the repo-controlled .claude/settings.json, before determining whether to display the workspace trust confirmation dialog. A malicious repository could set permissions.defaultMode to bypassPermissions in its committed...

Malicious code in hydra-node-consent-sdk (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 1e998f8c12c66f99b303fcb859d154aef24479204cf099b789871a4ee1a9943c The package hydra-node-consent-sdk was found to contain malicious code...

MAL-2026-1746 Malicious code in hydra-node-consent-sdk (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 1e998f8c12c66f99b303fcb859d154aef24479204cf099b789871a4ee1a9943c The package hydra-node-consent-sdk was found to contain malicious code...

Malicious code in hydra-consent-sdk (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector bb154df736a12683f5680d8355dde8927952f257d80b1355152bcf337754be96 The package hydra-consent-sdk was found to contain malicious code...

MAL-2026-1745 Malicious code in hydra-consent-sdk (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector bb154df736a12683f5680d8355dde8927952f257d80b1355152bcf337754be96 The package hydra-consent-sdk was found to contain malicious code...

CVE-2026-32235

An allowlist bypass flaw has been discovered in the npm @backstage/plugin-auth-backend package. Instances that have enabled experimental Dynamic Client Registration or Client ID Metadata Documents and configured allowedRedirectUriPatterns are affected. A specially crafted redirect URI can pass th...

GHSA-WQVH-63MV-9W92 @backstage/plugin-auth-backend: OAuth redirect URI allowlist bypass

Impact The experimental OIDC provider in @backstage/plugin-auth-backend is vulnerable to a redirect URI allowlist bypass. Instances that have enabled experimental Dynamic Client Registration or Client ID Metadata Documents and configured allowedRedirectUriPatterns are affected. A specially crafte...

Cross-site Scripting (XSS)

Overview ha-mcp is a Home Assistant MCP Server - Complete control of Home Assistant through MCP Affected versions of this package are vulnerable to Cross-site Scripting XSS in the OAuth consent form rendering process. An attacker can execute arbitrary JavaScript in the server operator's browser b...

GHSA-PF93-J98V-25PV ha-mcp has XSS via Unescaped HTML in OAuth Consent Form

Summary The ha-mcp OAuth consent form renders user-controlled parameters via Python f-strings with no HTML escaping. An attacker who can reach the OAuth endpoint and convince the server operator to follow a crafted authorization URL could execute JavaScript in the operator's browser. This affects...

ha-mcp has XSS via Unescaped HTML in OAuth Consent Form

Summary The ha-mcp OAuth consent form renders user-controlled parameters via Python f-strings with no HTML escaping. An attacker who can reach the OAuth endpoint and convince the server operator to follow a crafted authorization URL could execute JavaScript in the operator's browser. This affects...

EUVD-2026-11383

ha-mcp OAuth 2.1 DCR mode enables network reconnaissance via an error oracle...

Server-side Request Forgery (SSRF)

Overview ha-mcp is a Home Assistant MCP Server - Complete control of Home Assistant through MCP Affected versions of this package are vulnerable to Server-side Request Forgery SSRF via the haurl parameter in the OAuth consent form and forged tokens in REST and WebSocket tool calls. An attacker ca...

ha-mcp OAuth 2.1 DCR mode enables network reconnaissance via an error oracle

Summary The ha-mcp OAuth consent form beta feature accepts a user-supplied haurl and makes a server-side HTTP request to haurl/api/config with no URL validation. An unauthenticated attacker can submit arbitrary URLs to perform internal network reconnaissance via an error oracle. Two additional co...