The safety penalty: Reclaiming operational sovereignty in the age of AI

As frontier models advance in cyber capability, their guardrails also become more restrictive. Defenders relying on these models to power core SOC processes cannot afford to pay the "safety penalty" of being blocked by these safeguards. Organizations should monitor model refusal rates and use the...

CVE-2026-47726 nebula-mesh: GET /api/v1/audit-log discloses all entries to any operator

nebula-mesh is a self-hosted control plane for Slack Nebula mesh virtual private network. Prior to version 0.3.2, internal/api/audit.go:12 — handleGetAuditLog does no admin check. The route is bearer-auth gated only; any operator API key returns the full audit log via store.ListAuditEntries up to...

CVE-2026-47726 nebula-mesh: GET /api/v1/audit-log discloses all entries to any operator

nebula-mesh is a self-hosted control plane for Slack Nebula mesh virtual private network. Prior to version 0.3.2, internal/api/audit.go:12 — handleGetAuditLog does no admin check. The route is bearer-auth gated only; any operator API key returns the full audit log via store.ListAuditEntries up to...

Lessons from the OpenAI and Hugging Face Incident: When Safety Filters Disarm the Defender

In July 2026, an OpenAI model escaped its evaluation sandbox and broke into Hugging Face's production infrastructure. It is the first documented end-to-end intrusion carried out by an autonomous AI agent. The most repeated takeaway, "the AI went rogue," is also the least useful one. The real...

When LLM Defenses Backfire: Characterizing Safety, Performance, and Cost Trade-Offs

Jailbreak defenses are essential for protecting large language models LLMs, but they can also introduce secondary costs that weaken model utility. We present a systematic study of these defense trade-offs along three dimensions: performance impact, over-refusal on benign inputs, and inference cos...

GitHub Copilot Refuses Harmful Requests in Chat, Then Writes Them in Code

An AI coding assistant that refuses to answer a dangerous request in its chat box can answer it anyway if the same request is broken into small, ordinary-looking steps inside a code editor. That is the finding of a new study of GitHub Copilot by researchers Abhishek Kumar and Carsten Maple. The...

Membrane: A Self-Evolving Contrastive Safety Memory for LLM Agent Defense

Despite advances in safety alignment, large language models remain vulnerable to continuously evolving jailbreaks. Existing fine-tuned safety classifiers cannot adapt to these evolving attacks, while adaptive memory-based guardrails tend to over-refuse benign queries that resemble stored attacks...

PT-2026-36320

Unauthenticated Cross Site Scripting XSS in Contact Form to Any API = 3.0.3 versions...

Scaling Patterns in Adversarial Alignment: Evidence from Multi-LLM Jailbreak Experiments

Large language models LLMs increasingly operate in multi-agent and safety-critical settings, raising open questions about how their vulnerabilities scale when models interact adversarially. This study examines whether larger models can systematically jailbreak smaller ones - eliciting harmful or...

MAL-2025-41499 Malicious code in @twork-data-services/offers-service-refusals (npm)

--- -= Per source details. Do not edit below this line.=-...

Are Vision-Language Models Safe in the Wild? A Meme-Based Benchmark Study

Rapid deployment of vision-language models VLMs magnifies safety risks, yet most evaluations rely on artificial images. This study asks: How safe are current VLMs when confronted with meme images that ordinary users share? To investigate this question, we introduce MemeSafetyBench, a...