Google Cloud is expanding Gemini Enterprise with a catalog of partner-built security agents and AI-workload protections, announced at Google Cloud Next. The framing: threat actors increasingly use AI to accelerate cyberattacks, while enterprise defense is fragmented across identity, network, endpoint, data, cloud, and application layers, often split across a dozen or more products, each with its own context. The new partners bring those contexts together so defenders can orchestrate multi-step security workflows directly inside Gemini Enterprise.

The offerings fall into two categories: security agents that teams invoke directly in Gemini Enterprise, and protections for AI and agentic workloads. On the defense side, Check Point’s AI Defense Plane, integrated with the Google Cloud Agent Gateway and Agent Registry, provides discovery of AI workloads, risk-posture monitoring, non-compliant behavior detection, and real-time guardrails against prompt injection, data exposure, and rogue agentic behavior. CrowdStrike Falcon Guardian similarly extends guardrails and runtime protection to agentic workloads in Gemini Enterprise, while a separate CrowdStrike agent brings security context into investigation and response. Cyberhaven’s Linea agent adds discovery and classification of sensitive data across endpoints, cloud apps, and agentic workflows, powered by its data lineage model, and extends across Antigravity. Cyera Agent Guardian delivers data security posture management and data loss prevention for agents running on Gemini, correlating machine identities, delegated permissions, and data classification to verify authorized behavior. Fortinet’s FortiAIGate provides LLM runtime protection for Google Cloud customers, and Thales’ AI Security Fabric offers visibility, runtime protection, and centralized governance across agentic AI interactions.

Several agents target identity and incident response. Britive’s Emergency Termination Agent lets teams contain a compromised human or non-human identity from a single natural-language request, confirming the identity, listing active privileged sessions, revoking sessions with human approval, disabling the identity, and gathering audit context — lowering mean time to containment while preserving least-privilege access. Acalvio’s ShadowPlex deception agent automates deployment of decoys and honeytokens, including network decoys, identity honey accounts, RAG decoys, and honey skills, with no manual configuration. Ping Identity’s PingID self-service agent lets employees resolve MFA resets and device recovery in natural language without a help-desk ticket. Obsidian Security’s Risk Analyzer and Breach Response agents discover AI agents across enterprise environments, assess risk, and flag governance gaps such as which agents hold escalated privileges or write access. Transmit Security’s Agent Intelligence identifies agentic activity interacting with customer-facing applications, distinguishing legitimate agents from malicious ones.

Other agents bring security context into existing workflows. Exabeam is introducing the next generation of Nova, a unified multi-agent AI system that coordinates specialized agents for behavior analysis, risk prioritization, investigation, and response with shared operational context. Fastly’s Autonomous Edge Defense Agent (AEDA) investigates edge and infrastructure incidents in plain language, pairing an organization’s own telemetry with anonymized intelligence from Fastly’s global customer base. Menlo Security’s HEAT Shield analyzes web content and blocks zero-day threats and prompt injection at runtime, alongside its Orchestrator, which turns SOC responses into natural-language workflows, and MARS for runtime security across human and agent interactions. Palo Alto Networks’ CLARA agent continuously scans cloud infrastructure and AI workloads to surface and prioritize vulnerabilities. Qualys ROCky lets users ask questions like how exposed they are to Log4Shell or whether they’re meeting CISA KEV deadlines, with answers ranked by TruRisk score, and can stage and deploy patches. Splunk’s Security AI agent converts telemetry streams into real-time intelligence across security and observability data. Snyk validates what AI agents generate in real time, catching vulnerabilities before they reach a repo. Endor Labs’ AURI agents bring AI-native application security, including a SAST triage agent that classifies and prioritizes code findings inside Google Antigravity. Zscaler’s Risk360 agent quantifies cyber risk, analyzes Zero Trust posture, and recommends mitigations through natural-language interaction.

The agents are available through the Google Cloud Marketplace today, working with the security tools organizations already use and supporting custom agentic workflows on top of them.