Signal-Based Model Access Risk Analysis for AI System Operations Security

Artificial intelligence AI systems are now ubiquitous across domains such as security, finance, healthcare, consumer technology, and large-scale cloud services, where they process massive volumes of data and make consequential decisions daily. This widespread adoption has created a broad attack...

MAL-2026-10690 Malicious code in qwen-asr-pvt (PyPI)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 68e8b2f6db0443a648cdf03348c8dd9351568469f84b8d61022f1ed1da2a330e The package's pyproject.toml declares an unpinned runtime dependency on transformers4576, a lookalike of the widely used HuggingFace transformers...

patchtriage

PatchTriage !CIhttps://github.com/d01ki/patchtriage/actio...

arkindex-client (>=0.1.0, <=0.1.1), arlo (>=1.2.49, <=1.2.64) +18 more potentially affected by CVE-2025-50181 via urllib3 (1.24.0)

urllib3 PYPI version 1.24.0 is affected by a known vulnerability. The following packages have a transitive dependency on urllib3 and may be impacted: - arkindex-client =0.1.0, =1.2.49, =3.0.0, =0.6.2, =1.0.3, =0.1.2.15, =3.1.3, =0.3.9, =1.5.6, =0.10.0, =1.2.0, =1.2.4 and more Source cves:...

hf-model-provenance-scanner

hf-model-provenance-scanner !CIhttps://github.com/poojaki...

CVE-2026-45256 Missing permission check in thr_kill2(2)

When used to deliver a signal to a specific thread, thrkill22 called pcansignal to determine whether the operation was permitted but did not check the result before delivering the signal. The signal was sent even when the permission check failed. The system call returned the resulting error to th...

How Humans, Bots, and Agents Communicate about Vulnerabilities in Pull Requests

Developers may reference vulnerabilities in pull request discussions through both explicit identifiers, such as CVEs or GHSAs, and implicit security-related language e.g., "unauthorized access" or "SQL injection". Prior work has primarily focused on explicit identifiers, potentially overlooking...

Uncovering Hidden Attack Paths in Cloud Environments Using Runtime Signals

Wiz now layers runtime signals into the Security Graph, exposing hidden attack paths to give security teams a complete picture of risk...

Fake AI Agent Skill Passed Security Scans and Reportedly Reached 26,000 Agents

Security firm AIR built a fake AI agent skill, pushed it through a popular skill marketplace and an Instagram ad, and says it reached roughly 26,000 agents, including some on corporate accounts. Every skill security scanner the firm tested it against marked it safe. The payload was harmless by...

Scammers Use Fake GitHub Stars, VirusTotal Reviews to Spread Crypto Clipper

A multi-platform malware campaign abuses fake trust signals to infect Windows and Mac users with a crypto clipper packed with 15,500 attacker wallets...

AgentRiskBOM: A Risk-Scoping Security Bill of Materials for Agentic AI Systems

Agentic AI systems retrieve private context, invoke tools, write files, call external services, coordinate with other agents, and may act without human approval. Existing bill of materials artifacts improve transparency for dependencies, model metadata, and training provenance, but leave an agent...

Malicious code in @mastra/github-signals (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector e263abbd9ed9975f3f8aaecfd1a780616bc3aef00238ed9ba9075b5bf4e38f37 The package was found to contain malicious code or consuming dependency that contains malicious code Source: ghsa-malware...

Malicious code in backoffice-charges-module (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 047eb92a0e8bb401b2c205765616c9b4b715ee7cfd33d2e6ef9dc8d645b77f04 On every npm install, the preinstall lifecycle script node index.js /dev/null 2&1 silently HTTPS-POSTs a JSON payload to https://avamnrwqo7.rbmock.de...

Embedded Malicious Code

Overview Affected versions of this package are vulnerable to Embedded Malicious Code. An attacker using the npm account ehindero hijacked the @mastra npm organization and, in a roughly 30–45 minute burst starting around 01:12 UTC, republished the entire @mastra catalog. The Mastra source code was...

ARES: A Platform for Adaptive Role-Based Evaluation of Social Engineering Risks in Human--AI Games

This work introduces ARES, a platform and open pilot dataset for auditing adaptive social engineering risks in LLM-mediated social decision-making through controlled social games. ARES supports human--human, human--AI, and AI--AI settings, combining configurable game templates, role-conditioned L...

MAL-2026-5612 Malicious code in gpt-sdk (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 8b9bdc5e04979d5b4f73407bcedaecc9df24dbb03e0bfbc0edefe333023dc50c On npm install, postinstall.js runs unconditionally and collects a wide range of installer-side reconnaissance data: hostname and FQDN, contents of...

Reconstructing AI activity in investigations

AI systems are now part of everyday work. Investigators need a consistent way to reconstruct what happened within them. Security teams are already investigating activity involving Microsoft 365 Copilot and Azure AI services—from prompt injection attempts to unexpected data access. Those signals a...

FreeBSD -- sigqueue(2) missing capability mode restriction

Problem Description: sigqueue2 was marked as permitted in capability mode with the introduction of Capsicum in 2011, but the implementation of kernsigqueue did not include a capability mode check restricting signal delivery to the calling process's own PID. Impact: A process in capability mode ca...

RecurGuard: Runtime Monitoring for Reasoning-Token Consumption Attacks

Reasoning-capable large language models can be induced to spend their generation budget on injected decoy tasks rather than answering the user's question, causing denial of service when no final answer is produced and denial of wallet when excess output tokens are billed. Input-side safety...

From Operating Model to Product: How We Built the ROC for Detection-Speed Remediation

In the first article in this series, we made the case for a prevention-led operating model. This article is about what happened next: the decision to build something that did not exist, and what it took to make it real. Turning an operating model into a product sounds straightforward until you ar...