MATERIAL_CHANGES

Daily brief — 2026-08-30

# Agentic Security Change Brief — 2026-08-30

## Material Changes

### Project Perception documents human sign-off for every high-impact action, with defender-set objectives and guardrails.
- Decision impact: AUTHORIZATION, OPERATIONS, RISK
- Assertions: ASRT-HR-PP-001

### Project Perception documents a red/blue/green multi-agent architecture coordinated by orchestrated workflows and an orchestration harness.
- Decision impact: ARCHITECTURE, EVALUATION, RISK
- Assertions: ASRT-HR-PP-002

### Project Perception assigns work execution to agents while reserving judgment to humans.
- Decision impact: AUTHORIZATION, EVALUATION, RISK
- Assertions: ASRT-HR-PP-003

### Project Perception is in preview, initially delivered through Microsoft Defender, with broader Microsoft Security expansion planned.
- Decision impact: DEPLOYMENT, EVALUATION, PROCUREMENT
- Assertions: ASRT-HR-PP-005

### Codename MDASH is a private-preview agentic code scanner that surfaces code-level weaknesses across connected repositories.
- Decision impact: ARCHITECTURE, EVALUATION, PROCUREMENT
- Assertions: ASRT-HR-MD-001

### MDASH uses AI confidence and severity for triage while directing uncertain high-severity findings to human investigation before action.
- Decision impact: AUTHORIZATION, EVALUATION, RISK
- Assertions: ASRT-HR-MD-002

### MDASH accepts Defender CLI and connector-triggered remote scans and labels their origin for provenance.
- Decision impact: ARCHITECTURE, EVALUATION, OPERATIONS
- Assertions: ASRT-HR-MD-003

### MDASH exports validated findings to CSV for existing remediation workflows.
- Decision impact: OPERATIONS, PROCUREMENT
- Assertions: ASRT-HR-MD-004

### MDASH is in private preview and its documentation/access path requires authorization.
- Decision impact: DEPLOYMENT, EVALUATION, PROCUREMENT
- Assertions: ASRT-HR-MD-005

### The MDASH portal view surfaces only the most recent 90 days of scan activity.
- Decision impact: EVALUATION, OPERATIONS, RISK
- Assertions: ASRT-HR-MD-006

### Microsoft Learn identifies Project Perception as Limited Public Preview.
- Decision impact: DEPLOYMENT, PROCUREMENT
- Assertions: ASRT-HR-MO-001

### Project Perception combines signals, context, models, and red/blue/green agents through coordinated playbooks.
- Decision impact: ARCHITECTURE, EVALUATION, RISK
- Assertions: ASRT-HR-PP-002

### Project Perception supports autonomous execution initiated through reusable playbooks.
- Decision impact: AUTHORIZATION, OPERATIONS, RISK
- Assertions: ASRT-HR-PP-003

### Operators can approve or reject requested agent actions, stop sessions, and provide alternative guidance.
- Decision impact: AUTHORIZATION, OPERATIONS, RISK
- Assertions: ASRT-HR-PP-001

### Microsoft AI Code Security documents a multistage prepare/scan/validate/deduplicate pipeline using risk ranking, taint analysis, LSP type resolution, and multi-model debate.
- Decision impact: ARCHITECTURE, EVALUATION, RISK
- Assertions: ASRT-HR-AC-001

### The scanning stage documents more than 100 specialized AI agents targeting vulnerability classes, including injection, memory-safety, and auth-bypass auditors.
- Decision impact: ARCHITECTURE, EVALUATION, RISK
- Assertions: ASRT-HR-AC-002

### The Defender CLI can generate and apply code fixes directly from agentic scan results.
- Decision impact: AUTHORIZATION, OPERATIONS, RISK
- Assertions: ASRT-HR-AC-003

### Agentic code-scanning findings are published to Microsoft Security Exposure Management for organization-wide triage.
- Decision impact: ARCHITECTURE, OPERATIONS, PROCUREMENT
- Assertions: ASRT-HR-AC-004

### Microsoft AI Code Security is in preview across listed Azure commercial regions, with UAE limited to MDASH CLI scans.
- Decision impact: DEPLOYMENT, PROCUREMENT, RISK
- Assertions: ASRT-HR-AC-005

### The scanner claims broad language support with specially tuned expertise for C, C++, Java, and C#.
- Decision impact: EVALUATION, PROCUREMENT
- Assertions: ASRT-HR-AC-006

### Open Security has released ISPM Enterprise SQL v1 as its first autonomous cyber-defense benchmark.
- Decision impact: EVALUATION, PROCUREMENT
- Assertions: ASRT-HR-OS-001

### ISPM Enterprise SQL v1 uses a frozen synthetic enterprise and common questions, reporting 95% compatibility intervals across 127 tasks.
- Decision impact: EVALUATION, RISK
- Assertions: ASRT-HR-OS-002

### The benchmark supports local runs with user-owned model keys before submission for an official run.
- Decision impact: AUTHORIZATION, EVALUATION, RISK
- Assertions: ASRT-HR-OS-003

### Open Security labels the benchmark service as beta.
- Decision impact: PROCUREMENT, RISK
- Assertions: ASRT-HR-OS-004

## Control/Autonomy Changes

## Availability & Lifecycle Changes

## Capability Changes

## Research & Benchmark Changes

## Evidence Conflicts

## Why It Matters

## Sources