# Microsoft AI runtime security

URL: https://airuntimesecuritytools.com/vendors/microsoft/
Updated: 2026-11-16

Microsoft documents controls at several layers, and several are still in preview. Prompt Shields in Azure AI Content Safety detect user prompt and document attacks, and in Foundry can block flagged inputs before the model processes them. Defender for Endpoint runtime protection inspects the agent loop and can block risky activity before it runs, but it is public preview, with listed limits for certificate pinning, HTTP/3 and Claude Code. Alerts flow into Defender XDR. Agent 365 is listed at 15 dollars per user per month. Gaps from 16 Nov 2026: a latency figure in milliseconds, fail-open or fail-closed behavior and generally available endpoint blocking are none found.

## Strength and weakness for AI Runtime Security Tools

**Strength:** Defender can block prompt injection and risky agent actions before they run on the endpoint, and Prompt Shields block attack inputs in Foundry (microsoft-19, microsoft-41).

**Weakness:** None found on 2026-11-16 for published latency in milliseconds or fail-open and fail-closed behavior; endpoint blocking is preview with a documented Claude Code limit (microsoft-46, microsoft-21, microsoft-23).

## Scores and evidence

Microsoft is 15th in the overall ranking at 45.0 of 100. Each score below lists the fact ids behind it; the facts follow further down with their sources and read dates.

**Microsoft: score per criterion**

| Criterion | Score | Weight | Why | Facts |
|---|---|---|---|---|
| Action types stopped before they complete | 4 of 10 | 30 | Blocking of prompt attacks in Foundry and policy evaluation of tool invocations are documented; endpoint runtime blocking is public preview and the reported runtime stop has no availability date, so both count only as mentions. Blocking of shell, file, package and network actions as generally available: none found. | [microsoft-41](#fact-microsoft-41), [microsoft-30](#fact-microsoft-30), [microsoft-19](#fact-microsoft-19), [microsoft-26](#fact-microsoft-26), [microsoft-63](#fact-microsoft-63), [microsoft-66](#fact-microsoft-66), [microsoft-10](#fact-microsoft-10) |
| Layers covered | 6 of 10 | 20 | Model, application (API, where the customer application decides) and agent tool-invocation layers are documented. Endpoint enforcement is public preview only, so generally available endpoint enforcement: none found. | [microsoft-41](#fact-microsoft-41), [microsoft-30](#fact-microsoft-30), [microsoft-45](#fact-microsoft-45), [microsoft-42](#fact-microsoft-42), [microsoft-19](#fact-microsoft-19) |
| Latency and failure mode | 0 of 10 | 15 | The vendor says latency is minimal, but a millisecond figure and fail-open or fail-closed behavior: none found. | [microsoft-21](#fact-microsoft-21), [microsoft-46](#fact-microsoft-46) |
| Deployment | 6 of 10 | 15 | Documented operating systems, Intune or PowerShell onboarding and API delivery. Published time to deploy: none found. | [microsoft-13](#fact-microsoft-13), [microsoft-25](#fact-microsoft-25), [microsoft-45](#fact-microsoft-45), [microsoft-48](#fact-microsoft-48) |
| Detection and response | 6 of 10 | 5 | Alerts flow into the vendor's own XDR incidents and agent data is queryable in advanced hunting. The block record and audit mode belong to the preview runtime protection, and export to an outside SIEM or SOAR: none found. | [microsoft-47](#fact-microsoft-47), [microsoft-17](#fact-microsoft-17), [microsoft-24](#fact-microsoft-24), [microsoft-22](#fact-microsoft-22) |
| Transparency | 6 of 10 | 10 | Documentation is public and some prices are published, and the MCP gateway source is observed on GitHub. Published research: none found, and third-party press reports no pricing for the new agent controls. | [microsoft-52](#fact-microsoft-52), [microsoft-49](#fact-microsoft-49), [microsoft-50](#fact-microsoft-50), [microsoft-53](#fact-microsoft-53), [microsoft-66](#fact-microsoft-66) |
| Proof | 6 of 10 | 5 | A large public company with third-party coverage. Named customers or analyst coverage for these AI runtime products: none found. | [microsoft-64](#fact-microsoft-64), [microsoft-58](#fact-microsoft-58) |

## Where it ranks

| View | Rank | Total |
|---|---|---|
| [Overall ranking](/ranking/) | 15th | 45.0 |
| [Developers run coding agents with shell, file and package access](/best-for/coding-agents/) | 15th | 47.0 |
| [A company shipping a customer-facing LLM application](/best-for/customer-llm-app/) | 15th | 40.0 |
| [Staff use agents inside SaaS and cloud workspaces that act on company data](/best-for/saas-agents/) | joint 14th | 51.0 |
| [A regulated company that must prove what was blocked and respond to incidents](/best-for/regulated-response/) | 14th | 49.0 |
| [A lean team that must have runtime control in place this quarter](/best-for/fast-rollout/) | 15th | 49.0 |

## Product, deployment, price and owner

- Product assessed: Defender for Endpoint AI agent runtime protection (preview), Azure AI Content Safety Prompt Shields, Defender for Cloud AI threat protection, Agent 365 real-time protection
- Deployment: Cloud-managed Defender for Endpoint client onboarding (Intune or PowerShell for runtime protection); API-based Content Safety; Purview browser extension; Entra and Defender portal controls
- Time to deploy: none found on 2026-11-16 (discovery starts automatically after device onboarding; Purview policies need at least 24 hours)
- Price: Agent 365 is listed at $15.00 per user per month, paid yearly; Content Safety Standard shows a dash; Defender for Endpoint and Cloud Apps prices none found on 2026-11-16 on readable pages.
- Owner: Microsoft Corporation
- Website: learn.microsoft.com

## Facts and sources

- **microsoft-10** (vendor): Microsoft documents Endpoint DLP detection of pasting content to restricted service domains in supported browsers, evaluated on the pasted content.Quote: "Detects when a user attempts to paste content to a restricted service domain. Evaluation is performed on the content that is being pasted."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/purview/endpoint-dlp-learn-about), read 2026-11-16.
- **microsoft-12** (vendor): Microsoft says local agent discovery needs no extra deployment beyond onboarding to Defender for Endpoint and starts automatically; Plan 2 is the minimum license.Quote: "You don't need additional deployment, configuration, or scripts beyond the device onboarding requirements. If the device meets all of the prerequisites, agent discovery begins automatically."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/discover-local-ai-agents), read 2026-11-16.
- **microsoft-13** (vendor): Microsoft lists Windows, macOS, Linux, Android and iOS as Defender for Endpoint operating systems; local agent discovery is documented for Windows and macOS.Quote: "Microsoft Defender for Endpoint supports the following operating systems: Windows, macOS, Linux, Android, and iOS."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/microsoft-defender-endpoint), read 2026-11-16.
- **microsoft-17** (vendor): Microsoft says local agent data can be queried in advanced hunting, including an AgentsInfo table that carries MCP servers and declared tools.Quote: "Use advanced hunting to proactively investigate local AI agent presence, understand how agents are configured, and identify the agents and users that carry the most risk."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/discover-local-ai-agents), read 2026-11-16.
- **microsoft-19** (vendor): Microsoft says Defender for Endpoint runtime protection inspects the agent loop on endpoints and can block risky activity before it executes; this is a public preview feature.Quote: "Microsoft Defender inspects the agent loop (user prompts, tool calls, and tool responses) and can block risky activity before it executes"Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/whats-new-in-microsoft-defender-endpoint), read 2026-11-16.
- **microsoft-20** (vendor): Microsoft documents agent-native event inspection for agents that expose vendor event interfaces (Claude Code, Codex CLI, GitHub Copilot CLI) and network inspection for others.Quote: "Agents such as Claude Code, Codex CLI, and GitHub Copilot CLI expose these interfaces, and Defender uses them to inspect agent activity"Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/ai-agent-runtime-protection-overview), read 2026-11-16.
- **microsoft-21** (vendor): Microsoft says each runtime scan is an inline check with minimal added latency, but publishes no millisecond figure on this page.Quote: "Each scan is a fast, inline check rather than continuous monitoring of the agent process, so the added latency is minimal."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/ai-agent-runtime-protection-overview), read 2026-11-16.
- **microsoft-22** (vendor): Microsoft documents block, audit and disabled modes, advises starting in audit, and protects the setting with tamper protection.Quote: "Microsoft recommends starting in audit mode to observe detections and validate accuracy before switching to block mode for active enforcement."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/ai-agent-runtime-protection-overview), read 2026-11-16.
- **microsoft-23** (vendor): Microsoft lists preview limits: network inspection skips agents using certificate pinning or HTTP/3, and Claude Code may keep processing a tool response in Block mode after a post-tool-response block.Quote: "Network inspection doesn't support agents that use certificate pinning or HTTP/3."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/ai-agent-runtime-protection-overview), read 2026-11-16.
- **microsoft-24** (vendor): Microsoft says a block is notified to the user in the agent UI and by Windows toast, recorded in protection history, and sent as a security alert correlated into incidents.Quote: "a security alert is sent to Defender, correlated into incidents for the SOC to investigate."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/ai-agent-runtime-protection-overview), read 2026-11-16.
- **microsoft-25** (vendor): Microsoft says runtime protection is configured in audit or block mode through local PowerShell or organization-wide Intune.Quote: "For local PowerShell and organization-wide Microsoft Intune configuration options"Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/ai-agent-runtime-protection-overview), read 2026-11-16.
- **microsoft-26** (vendor): The support matrix lists runtime protection per agent on Windows (preview), including Claude Code, GitHub Copilot CLI and app, Claude Desktop Code, VS Code Copilot and Claude Code extensions, plus OpenClaw, Ollama Desktop and ChatGPT Desktop via network inspection.Quote: "Network inspection covers agents that communicate with LLMs through well-known inference services. It doesn't cover agents that use certificate pinning or HTTP/3."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/ai-agent-support-matrix), read 2026-11-16.
- **microsoft-28** (vendor): Microsoft Defender integrates with Agent 365 so that real-time protection evaluates tool calls and responses for Work IQ MCP and customer MCP tools onboarded to Agent 365; the default rule audits and custom rules can block before execution.Quote: "custom rules can block matching actions before they execute"Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-cloud-apps/ai-agent-protection), read 2026-11-16.
- **microsoft-30** (vendor): Microsoft says real-time protection evaluates tool invocations against policy and near-real-time detections cover indirect prompt injection, secret leakage and suspicious IP access.Quote: "Near-real-time detections analyze Agent 365 observability data for threats such as indirect prompt injection, evasion techniques, malicious content propagation, secret leakage"Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-cloud-apps/ai-agent-protection), read 2026-11-16.
- **microsoft-41** (vendor): Microsoft says Prompt Shields detect user prompt attacks and document attacks, and in Foundry, with blocking enabled, block inputs classified as attacks before the model processes them.Quote: "When you enable blocking in Foundry, Prompt Shields block inputs classified as attacks before those inputs are processed by the model."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/azure/ai-services/content-safety/concepts/jailbreak-detection), read 2026-11-16.
- **microsoft-42** (vendor): Microsoft says in the Azure AI Content Safety API the service does not itself block or route content; the customer application decides.Quote: "In Azure AI Content Safety APIs, the API doesn't itself block or route content."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/azure/ai-services/content-safety/concepts/jailbreak-detection), read 2026-11-16.
- **microsoft-43** (vendor): Microsoft says Prompt Shields can scan document attacks at the tool response point for agents on supported tools (Preview) and that document attack detection is off by default.Quote: "For agents, Prompt Shields can also scan document attacks at the tool response intervention point for supported tools (Preview)."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/azure/ai-services/content-safety/concepts/jailbreak-detection), read 2026-11-16.
- **microsoft-44** (vendor): Microsoft states Prompt Shields do not judge whether content is harmful or accurate and that performance is primarily trained on English.Quote: "the user prompt attack and document attack models are primarily trained and evaluated on English content."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/azure/ai-services/content-safety/concepts/jailbreak-detection), read 2026-11-16.
- **microsoft-45** (vendor): Microsoft says guardrails work inside Microsoft Foundry or as a standalone offering through Content Safety Studio and APIs for models and agents outside Foundry.Quote: "Your models and agents are outside Foundry - or span multiple platforms - and you want a consistent guardrail experience."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/azure/ai-services/content-safety/overview), read 2026-11-16.
- **microsoft-46** (none-found): None found on 2026-11-16: no published latency figure or fail-open or fail-closed behavior for Prompt Shields or Content Safety on the Prompt Shields and overview pages.Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/azure/ai-services/content-safety/concepts/jailbreak-detection), read 2026-11-16.
- **microsoft-47** (vendor): Microsoft says Defender for Cloud AI threat protection works with Prompt Shields to alert on data leakage, jailbreak and credential theft, and sends alerts to Defender XDR; it supports text tokens only.Quote: "provide security alerts for threats like data leakage, data poisoning, jailbreak, credential theft, and more"Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/azure/defender-for-cloud/ai-threat-protection), read 2026-11-16.
- **microsoft-48** (vendor): Microsoft says Defender for AI Services has a 30-day free trial capped at 75 billion tokens scanned and is limited to commercial clouds.Quote: "Defender for AI Services includes a 30-day free trial, capped at 75 billion tokens scanned."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/azure/defender-for-cloud/ai-threat-protection), read 2026-11-16.
- **microsoft-49** (vendor): Microsoft lists Agent 365 at $15.00 per user per month, paid yearly, on its product page.Quote: "$15.00 user/month, paid yearly (Annual commitment)"Source: [microsoft.com](https://www.microsoft.com/en-us/microsoft-agent-365), read 2026-11-16.
- **microsoft-50** (vendor): Microsoft's Azure AI Content Safety pricing page shows a free tier of 5,000 text records per month and a Standard tier whose price displays as a dash.The exact quote is on the source page.Source: [azure.microsoft.com](https://azure.microsoft.com/en-us/pricing/details/content-safety/), read 2026-11-16.
- **microsoft-51** (none-found): None found on 2026-11-16: no price for Defender for Cloud Apps or Defender for Endpoint standalone is readable on the learn pages; they point to Microsoft 365 E5 inclusion and separate pricing pages that returned 403 to a script.Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/microsoft-defender-endpoint), read 2026-11-16.
- **microsoft-52** (vendor): Microsoft documentation on learn.microsoft.com is readable without sign-in, with last-updated dates (Agent 365 overview 2026-11-12).Quote: "Last updated on 2026-11-12"Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/microsoft-agent-365/overview), read 2026-11-16.
- **microsoft-53** (observed): Microsoft's MCP Gateway source is public on GitHub under the MIT license and offers telemetry, access control and observability hooks.Quote: "Enterprise-ready integration points including telemetry, access control and observability."Source: [github.com](https://github.com/microsoft/mcp-gateway), read 2026-11-16.
- **microsoft-56** (third-party): On 2026-10-20, Help Net Security quoted Satya Nadella, CEO of Microsoft, on treating frontier models as insider risks, and said Project Perception shares signals with Agent 365.Quote: "Satya Nadella , CEO of Microsoft, wrote"Source: [helpnetsecurity.com](https://www.helpnetsecurity.com/2026/10/20/microsofts-ai-agents-threats-guardian-agents/), read 2026-11-16.
- **microsoft-58** (none-found): None found on 2026-11-16: no named customer list or analyst page for shadow AI discovery, MCP security or AI runtime security was read on the Microsoft pages fetched.Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/whats-new-in-microsoft-defender-endpoint), read 2026-11-16.
- **microsoft-59** (vendor): Microsoft owns and sells these products itself: Defender for Endpoint is described as the endpoint security pillar of Microsoft Defender.Quote: "As the endpoint security pillar of Microsoft Defender , Defender for Endpoint feeds endpoint signals into the unified Defender portal."Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/defender-endpoint/microsoft-defender-endpoint), read 2026-11-16.
- **microsoft-61** (vendor): Microsoft Purview says DSPM for AI policies need at least 24 hours to collect data before results appear.Quote: "Allow at least 24 hours for these new policies to collect data to display the results in Data Security Posture Management for AI"Source: [learn.microsoft.com](https://learn.microsoft.com/en-us/purview/dspm-for-ai), read 2026-11-16.
- **microsoft-63** (third-party): On 2026-10-19, SiliconANGLE reported Microsoft extended Conditional Access to agents so that rising risk can hold higher-impact actions for human approval and, at the top of the scale, isolate or stop the runtime.Quote: "Higher-impact actions can be held for verified human approval, and at the top of the scale the runtime is isolated or stopped entirely."Source: [siliconangle.com](https://siliconangle.com/2026/10/19/microsoft-targets-the-agentic-insider-with-new-controls-for-rogue-ai-agents/), read 2026-11-16.
- **microsoft-64** (third-party): On 2026-10-19, SiliconANGLE covered the launch and named Satya Nadella as Microsoft Chairman and Chief Executive and Hayete Gallot as executive vice president of Microsoft Security.Quote: "Microsoft Chairman and Chief Executive Satya Nadella laid out the thinking behind the controls in an Oct. 10 essay"Source: [siliconangle.com](https://siliconangle.com/2026/10/19/microsoft-targets-the-agentic-insider-with-new-controls-for-rogue-ai-agents/), read 2026-11-16.
- **microsoft-66** (third-party): On 2026-10-19, SiliconANGLE reported Microsoft gave no availability dates or pricing for the new agent controls announced that day.Quote: "Microsoft did not give availability dates or pricing for the new controls"Source: [siliconangle.com](https://siliconangle.com/2026/10/19/microsoft-targets-the-agentic-insider-with-new-controls-for-rogue-ai-agents/), read 2026-11-16.

## Related pages

- [All vendor profiles](/vendors/)
- [Lasso Security profile](/vendors/lasso-security/)
- [Noma Security profile](/vendors/noma-security/)
- [The ranking](/ranking/)
- [How scores are set](/method/)
- [Best tool by situation](/best-for/)
- [Runtime Action Type Matrix](/runtime-action-type-matrix/)
- [Enforcement Point Map](/enforcement-point-map/)
