Adversa AI

Active, privately held.

Sources 1, 2, 3 · checked 2026-11-08

• AI red teaming for agentic AI, LLMs, generative AI, and MCP, run by an autonomous red teaming engine with an optional managed service of dedicated red-team experts
• Tests mapped to OWASP Top 10 and MITRE ATLAS; findings also mapped to NIST AI RMF, EU AI Act, and ISO/IEC 42001
• Agent and MCP testing: tool misuse, tool hijack, goal manipulation, inter-agent attacks, command injection, privilege escalation
• Reports: risk-ranked findings with reproducible attack artifacts, executive summary PDF, technical report, and compliance report
• Re-scan to confirm mitigations hold; campaigns can trigger on a new model version or prompt template change (continuous testing)
• Coding Agent Security Platform and AI Detection and Response (runtime)

Sources 1, 2, 3 · checked 2026-11-08

• Teams that want automated, continuous testing with an option to add human red-team experts
• Teams testing coding agents and MCP-connected agents

Sources 1, 2, 3 · checked 2026-11-08

Contact vendor (book a demo or pilot)

Sources 1, 2, 3 · checked 2026-11-08

• SOC 2 Type I
• SOC 2 Type II
• ISO 27001
• GDPR (stated as compliant, not a certification)

Listed as stated by the vendor. Scope can differ by product or subsidiary; request current reports.

Sources 1, 2, 3 · checked 2026-11-08

• No public pricing
• Coverage percentages and speed figures on the service page are vendor claims; ask for a sample report

Sources 1, 2, 3 · checked 2026-11-08