Find and fix code vulnerabilities faster with Codename MDASH | BRK319
John Nix and Henrique Oliveira present Codename MDASH through a pharmacy application demo, following a chained exploit from discovery through security review and developer remediation.
Multi-model vulnerability discovery
This intermediate-level Microsoft Ignite 2026 session introduces MDASH’s multi-model agentic scanning architecture. MDASH orchestrates specialized AI agents and security-focused models to identify, validate, and prioritize code vulnerabilities, helping teams understand risk and focus on the issues that matter most.
The session covers the changing software security landscape and AI-driven vulnerability discovery, including customer findings and examples of complex vulnerabilities discovered by MDASH. It also discusses increased CVE discovery and MDASH’s impact inside Microsoft.
Demonstration: from exploit to validated fix
The demonstration uses the Caldova pharmacy application to connect vulnerability discovery with security analyst and developer workflows:
- Application and agent workflows: An introduction to the application and the AI agents involved in its analysis.
- Attacker perspective: An exploit demonstration involving chained vulnerabilities and broken authorization.
- Security analyst review: A workflow using Defender, repository scans, findings, and remediation pull requests.
- Developer remediation: Rebuilding the application and validating fixes.
Together, these segments show how findings move from investigation into repository-based remediation, with validation included as part of the fix workflow.
Operational considerations
The closing segment covers operational guidance and best practices, Foundry integration, scaling, GitHub integration, and audience questions.
Session chapters
- 00:00 — Introduction to the session and Codename MDASH
- 01:11 — Changing software security landscape and AI-driven vulnerability discovery
- 03:37 — Multi-model agentic scanning architecture
- 06:40 — Customer findings and complex vulnerability examples
- 09:12 — Increased CVE discovery and impact inside Microsoft
- 10:09 — Caldova pharmacy application demo and AI agent workflows
- 12:10 — Attacker exploit demonstration: chained vulnerabilities and broken authorization
- 15:06 — Security analyst workflow: Defender, repository scans, findings, and remediation pull requests
- 20:15 — Developer remediation, application rebuild, and fix validation
- 22:43 — Operational guidance, best practices, integrations, scaling, and audience Q&A