AI Security · Agent Runtime

AI Agent Runtime Security: What the CrowdStrike and Sysdig Moves Mean for IT Leaders

By Infonaligy · Updated November 9, 2026 · 8 min read

AGENTS, WATCHED

AI agent runtime security is the set of controls that watch what an agent actually does while it runs, and that can pause, block or stop an action before it causes harm. It differs from prompt filters, which inspect text going in and out, and from pre-deployment review, which happens once. In early November 2026, two security vendors made moves that point to where this market is heading: a "test, watch, stop" lifecycle for every agent you run.

This post is about what those moves mean for buying decisions. For the broader case on why agents opened a security gap in the first place, see the AI agent monitoring gap. If you need the operational detail on shutting an agent down, read our guide to the AI agent kill switch and containment. If your concern is which tools and systems an agent can reach, start with MCP governance for agent tool access.

What did CrowdStrike and Sysdig announce?

CrowdStrike agreed to buy two AI security companies, one for agent observability and one for automated AI red teaming, and Sysdig made its runtime product for AI agents generally available. Together they cover testing, watching and stopping agents.

  • CrowdStrike and Fiddler AI: On November 3, 2026, at Fal.Con Europe, CrowdStrike announced a definitive agreement to acquire Fiddler AI, which focuses on agent observability and explainability. CrowdStrike says the capability will become Falcon Observe.
  • CrowdStrike and Dreadnode: The same day, CrowdStrike also agreed to acquire Dreadnode, which builds automated AI security testing and red teaming. CSO Online reported that the deals extend CrowdStrike's coverage across the agent lifecycle. According to The Wall Street Journal, as reported by CSO Online, the deals are valued at about $400 million for Fiddler and $95 million for Dreadnode.
  • Sysdig AI Defense: Sysdig made AI Defense generally available. According to SecurityWeek, it discovers agents, MCP servers and plugins on endpoints and cloud workloads, builds a live inventory from runtime telemetry, and lets teams confirm, block or kill risky agent actions.

The specific products matter less than the pattern. Security platforms are treating agents as a workload class that needs its own testing, telemetry and enforcement, the way endpoints and containers did before them.

What does runtime security for AI agents actually mean?

It means judging an agent by its actions in production, not by its prompt or its design document. A runtime control sees the tool call, the data touched and the system targeted, and it can intervene in that moment.

Most organizations already have two other layers, and both are useful but incomplete.

Control layerWhat it seesWhat it misses
Prompt and output filtersText entering and leaving the modelHarmful actions that look harmless as text, such as a valid API call to the wrong system
Pre-deployment reviewThe agent's design, permissions and test results at launchDrift after launch, new tools added later, and inputs no one tested
Runtime securityLive tool calls, data access, process and network activityDepends on good inventory and clear policy to act on

An agent that reads a poisoned email and then exports a customer list may pass every prompt filter, because no single message looks malicious. Only a control watching the export, the destination and the agent's normal pattern can catch it.

Which three capabilities should you demand?

Ask for testing before deployment, observability during operation, and enforcement at the moment of action. A product or platform that covers only one of these leaves you assembling the rest yourself.

  • Pre-deployment testing and red teaming: Automated attempts to make the agent misbehave through prompt injection, tool misuse and data exfiltration paths, run before launch and again after every meaningful change. A one-time penetration test does not keep up with agents that gain new tools monthly.
  • Behavioral observability with explainable decision trails: A record of each step the agent took, what input led to it and why it chose that action, in a form your analysts can read during an incident. Logs of API calls alone do not explain intent. Our guide to AI agent observability covers what that telemetry should include.
  • Runtime enforcement: The ability to require confirmation for high-risk actions, block actions that violate policy, and kill an agent session or process outright. Enforcement should live outside the agent, so a manipulated agent cannot talk its way past it.

The headline

Vendors are converging on a test, watch, stop lifecycle for AI agents. Whatever you buy or build, make sure you can red team agents before launch, read an explainable trail of what they did, and confirm, block or kill risky actions at runtime from outside the agent itself.

Should you wait for your EDR vendor, buy a point tool, or build controls yourself?

It depends on how many agents you run, where they run, and how fast they are growing. Most mid-size organizations end up with a mix, using platform controls where they exist and filling gaps deliberately.

OptionFits best whenWatch for
Wait for your EDR or XDR vendor's bundleAgents are few, mostly on managed endpoints, and your SOC already lives in that consoleRoadmap timing after acquisitions, coverage for agents in SaaS and cloud you do not manage, and licensing changes
Buy a point tool nowAgents are multiplying across cloud workloads, developer machines and MCP servers, and you need inventory and enforcement this quarterAnother console for the SOC, overlap with your platform later, and how cleanly it exports telemetry
Build controls in your agent platformYou run custom agents on infrastructure you control and can enforce policy at the tool gatewayEngineering upkeep, gaps for third-party agents, and the need to still feed the SOC

Building in the platform pairs well with AI DevOps practices such as tool allowlists, scoped credentials and approval steps in the pipeline. For agents we design as custom AI agents or run on hosted AI infrastructure, we put enforcement at the tool layer from day one, so a later vendor product becomes a second line of defense rather than the only one.

What questions should you ask vendors?

Ask how the product discovers agents, what it can actually stop, and where its data goes. Clear answers to these questions separate runtime security from repackaged logging.

  • How do you discover agents, MCP servers and plugins we did not register, and on which platforms (endpoints, cloud workloads, SaaS)?
  • Can you show the decision trail for a single agent action, including the input that triggered it?
  • Which actions can you confirm, block or kill, and does enforcement happen outside the agent process?
  • How do you test agents before deployment, and can tests rerun automatically when tools or prompts change?
  • What happens to agent telemetry: where is it stored, for how long, and can it flow to our SIEM or SOC tooling?
  • If you were acquired or are integrating an acquisition, what changes for current customers over the next year?

What should IT leaders do in the next 90 days?

Inventory your agents, define what each is allowed to do, require approvals for risky actions, and send agent telemetry to the SOC. None of these steps require waiting for a vendor roadmap.

  • Days 1 to 30, inventory: List every agent, MCP server and plugin in use, with an owner, its permissions and the systems it reaches. Include agents inside SaaS products and on developer machines. Our post on finding shadow AI agents covers where to look.
  • Days 31 to 60, define allowed actions and approvals: For each agent, write down the actions it may take on its own, the actions that need a human to confirm, and the actions it may never take. Wire confirmation into high-risk actions such as payments, external sharing, deletions and permission changes.
  • Days 61 to 90, route telemetry and test: Send agent action logs to the same place your SOC watches endpoints and identity, write a first detection or two for unusual agent behavior, and run a red team exercise against your highest-risk agent.

Our AI security team helps organizations run this plan and choose between platform, point tool and build options, and our AI consulting practice ties it to your broader agent roadmap.

Frequently asked questions

What is AI agent runtime security?

It is the set of controls that monitor what an AI agent does while it runs, including tool calls, data access and network activity, and that can require confirmation for, block or kill a risky action before it completes. It complements prompt filters and pre-deployment review rather than replacing them.

What did CrowdStrike announce about AI agent security in November 2026?

On November 3, 2026, at Fal.Con Europe, CrowdStrike announced definitive agreements to acquire Fiddler AI, an agent observability and explainability company whose technology is to become Falcon Observe, and Dreadnode, which builds automated AI security testing and red teaming. The deals were reportedly valued at about $400 million and $95 million.

Are prompt filters enough to secure AI agents?

No. Prompt filters inspect text, but many harmful agent actions look harmless as text, such as a valid API call that sends data to the wrong place. You also need runtime visibility into actions and the ability to stop them.

Should we wait for our EDR vendor to add agent security?

If you run few agents, mostly on managed endpoints, waiting can be reasonable. If agents are spreading across cloud workloads, SaaS and developer machines, start with inventory, approvals and telemetry now, and consider a point tool or platform-level controls to cover the gap.

What is the first step toward AI agent runtime security?

Build an inventory of every agent, MCP server and plugin in use, with an owner, permissions and the systems each one can reach. Without an inventory, there is nothing reliable to test, observe or enforce against.

Infonaligy is based in Dallas–Fort Worth, serves the regions listed on our locations page, and delivers AI agent security work remotely to clients nationwide.

Test, watch, stop

Find out whether you could stop a risky agent action today.

Book a 30-minute assessment and we'll review which agents you run, what they can reach, where approvals and telemetry are missing, and whether a platform bundle, point tool or built-in controls fit best.

DFW · remote nationwide · governed by default · 800-985-1365