Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

AI agents are moving into production fast. They read email, call tools, query databases, and trigger business processes, often chaining those steps in ways no single team designed. Building them has never been easier. Securing what happens after an agent is manipulated is the hard part. That is the gap Salt and AWS will tackle together on Tuesday, November 10, at 9:00 am PT (12:00 pm ET) in a live session: AWS + Salt: Build Fast, Stay Secure with Agentic AI. Save your seat →

Security teams have never been better at finding problems. Discovery is broader, detection is sharper, and the volume of findings keeps climbing. But a finding is not a fix. When teams identify a risky AI agent, an exposed MCP server, or an active attack against an API, what happens next usually looks the same: someone triages it, someone opens a ticket, and someone on another team eventually makes the change. Until then, the risk stays exactly where it was.

On October 21, 2026, two critical vulnerabilities, PHP2Shell (CVE-2026-109618 and CVE-2026-109619), were publicly disclosed. Both are rated 9.2 and both allow remote code execution on vulnerable servers. Government portals, public services, and commercial sites across multiple countries went offline as a precaution while teams rushed to patch. But the flaws were known on October 20, and the patch didn’t ship until the 21st.

APRA's information security standard already covers the APIs your agents run on. Most institutions cannot yet evidence it. All APRA-regulated institutions in Australia are bound to meet CPS 234. This is the Prudential Standard for Information Security. Whether that's a bank, insurance provider, or super fund, they all fall under CPS 234, as do their third-party providers. Together, these institutions hold trillions of dollars in assets.

Nearly half of the organizations we surveyed — 49.8% — confirmed or suspected that an AI agent in their environment took an action in the past year that nobody intended, expected, or authorized. That is not a prediction. It is what security and technology professionals told us when we asked them directly, across a survey of over 380 organizations, about their experience over the last 12 months. Agentic AI has moved fast. The data shows it has moved faster than most organizations can govern.

AI Detection and Response (AI-DR), also called AI runtime security, is quickly becoming a must-have. Prompt injection, jailbreaks, and data leakage are real, and the prompt is often where attacks begin. But the prompt is only the first hop. In an agentic enterprise, a single instruction can trigger activity well beyond the LLM, all the way to your business systems and data.

Salt Labs found that the agentic AI platform Manus could be hijacked with a single email. By hiding malicious instructions inside an ordinary message, researchers got Manus to execute malicious code and, from there, reach the email, cloud storage, and code repository accounts a user had connected to it. The full attack required nothing from the victim beyond asking Manus to check their inbox. No stolen password, no clicked link.

Salt Labs recently showed that a single email could hijack the AI agent platform Manus and reach a victim’s connected accounts. The full technical breakdown, with complete evidence, payloads, and screenshots, is in our research team’s write-up, and Dark Reading first reported the finding in an exclusive. This post is the shorter version: what the finding means and what it tells every organization now deploying AI agents. For the full technical detail, read the Salt Labs research blog.

As AI adoption accelerates inside enterprises, security and IT teams are increasingly asking a foundational question: what exactly is connected to our AI, and what can it access? For organizations running Claude Enterprise, Salt now has a direct answer. Salt Claude Connect links Salt to Claude’s Compliance API and provides continuous, read-only visibility into the MCP (Model Context Protocol) servers connected to your Claude Enterprise organization.

The problem is not a lack of controls. It is connecting them into one attack story. The more time I spend with enterprise AI deployments, the clearer one thing becomes: AI security is incredibly fragmented. There are LLM guardrails, AI gateways, MCP security tools, API security, endpoint controls, SASE, code scanning, and runtime detection. Each solves a real problem, but agentic systems do not experience them as separate layers, and neither do attackers.