Cisco AI Defense

Cisco's enterprise AI security platform — the most comprehensive offering from a major network vendor. Provides three core capabilities: Discover (AI Cloud Visibility including AI BOM and MCP Catalog for supply chain governance), Detect (algorithmic red teaming with multi-turn, multi-language adversarial testing for models and agents), and Protect (runtime guardrails for prompts, responses, and agentic interactions including MCP traffic inspection). Integrates with NVIDIA NeMo Guardrails. Maps to NIST AI RMF, MITRE ATLAS, and OWASP LLM Top 10.

Covers the full AI security lifecycle from supply chain to runtime. The Feb 2026 expansion added critical agentic capabilities: real-time inspection of MCP traffic, detection of poisoned tools and unauthorized tool use, and AI BOM for tracking AI software assets including MCP servers and third-party dependencies. Network-layer enforcement means policies apply across all AI traffic traversing Cisco infrastructure.

Deploy when you need organisation-wide AI security enforcement integrated with existing Cisco infrastructure. The breadth of coverage (supply chain + red teaming + runtime) reduces the need for multiple point solutions. Enterprise procurement with network integration and policy management required. Strongest for organisations already in the Cisco ecosystem.

Risks addressed — mapped to both OWASP Top 10 standards. 6 in LLM, 4 in Agentic.

What Yuntona stores. Single source of truth — fork it on GitHub.

name: Cisco AI Defense
slug: cisco-ai-defense
type: Mixed
category: AI Guardrails & Firewalls
url: https://www.cisco.com/site/us/en/products/security/ai-defense/index.html

reviewed:   2026-04
added:      2026-04
updated:    2026-04

risks:
  llm:  [LLM01, LLM02, LLM03, LLM06, LLM07, LLM08]
  asi:  [ASI01, ASI02, ASI04, ASI07]

complexity:    Enterprise Only
pricing:       —
audience:      Blue Team
lifecycle:     [deploy]

tags: [Agentic, AI-SPM, Cisco, Enterprise, Firewall, MCP]