Agentless threat hunting is the process of proactively searching for undetected cyber threats in a network without using software or agents on endpoints.

It relies on already existing infrastructure and real-time monitoring and protection instead. And that not only eases deployment but also minimizes resource usage.

Other interesting terms…

Read More about Agentless Threat Hunting

Learn all about the process. Find out how it works, along with its types, steps, and tools, here. First, however, let us differentiate between it and agent-based threat hunting.

Agentless versus Agent-Based Threat Hunting, What Is the Difference?

Agent-based or traditional threat hunting uses a combination of tools, methodologies, and steps to seek out threats using applications installed on all the devices requiring protection. As such, hunting occurs per device.

Agentless threat hunting, meanwhile, as mentioned above, does not require software installation on all potentially vulnerable network-connected devices. Most of the hunting, in this case, happens on the network level.

Think of a building, for instance. While agent-based threat hunting may require roving security guards on-premises to check all potential entry points are locked and intruders are not allowed in, agentless threat hunting relies on digital cameras with built-in alarms to signal illicit device usage and scanners installed on automatic doors that disallow unauthorized persons in.

Agent-Based versus Agentless Threat Hunting

How Does Agentless Threat Hunting Work?

Threat hunters use automated security solutions to search for, log, monitor, and address threats before they can cause serious damage. These programs rely on data gathered by threat detection systems and other security solutions.

During the process, hunters comb through the security data for hidden malware, stealth attackers, and other signs of suspicious activity that automated systems may have missed. When they find something, they quickly act to eradicate the threat and strengthen defenses to ensure nothing like it does not happen again.

All this, of course, happens on the network level.

What Are the Types of Agentless Threat Hunting ?

The types of threat hunting remain the same, regardless of kind—agentless or agent-based.

1. Hypothesis-Driven

This type starts from a hypothesis an analyst crafts based on threat intelligence; recent incident trends; or known tactics, techniques, and procedures (TTPs). An example? A hunter may believe an attacker is using valid credentials for lateral movement. In that case, he designs queries and analytics to validate or refute his theory.

2. Intelligence-Driven

Here, threat intelligence, including indicators of compromise (IoCs), threat actor profiles, and adversary infrastructure, acts as the jump-off point. Hunters enrich the intel they gathered with log data, correlate it with external feeds, or contextualize it with the data from agentless security platforms.

This hunting process tracks actor-controlled infrastructure, such as command-and-control (C&C) server domains, to identify the tools and malware associated with specific groups or seen in recent campaigns.

3. Analytics-Driven

This type uses statistical analyses, machine learning (ML), and outlier detection systems to identify anomalies that may indicate malicious activity. Analysts craft custom detection models or utilize existing unsupervised learning algorithms to flag deviations from established baselines.

Security specialists can opt to cluster user behavior, model peer group activity, and detect spikes in rare command-line invocations, parent-child process chains, or privilege escalation attempts. In the cloud, they may identify unusual cross-region access, excessive application programming interface (API) calls, or token abuse patterns that differ from normal workflow trends.

4. Situational or Custom

An organization’s unique context founded on previous security incidents, geopolitical issues, targeted attacks, security system alerts, and other factors plays a major role in threat hunting. Analysts can create custom hunts that use intel- and hypothesis-based methodologies to identify threats specific to its environment.

An example? The impending launch of a groundbreaking product—the first of its kind—can spur outsiders to spy on its network. Threat actors, for instance, can try their best to break in to steal top-secret info they can sell to the company’s competitors. In such a case, the organization can hunt for any sign of a break-in constantly to keep prying eyes out.

What Are the Steps in Agentless Threat Hunting ?

Here is a rundown of the steps the process entails.

Step #1: Formulate a Hypothesis

At this stage, security analysts need to identify a theory they can test based on threat intelligence, known TTPs, recent incidents, or environmental risks. Say, for instance, that threat actors have started using cloud service tokens to move laterally within a multi-account Amazon Web Services (AWS) environment. Hunters can thus investigate AWS cloud account logins to spot anomalies.

Step #2: Determine the Hunt’s Scope

The next step is to identify the data sources related to the hypothesis. In the example’s case, these sources can include cloud audit trails, identity access management (IAM) data, and API invocation records. Analysts should also check if the data is complete and the timestamps are correct.

Step #3: Investigate Anomalies

In this stage, analysts must use structured queries, match patterns, or filter behaviors to spot signals in line with the hypothesis. They should take note of a lead like a service account invoking an abnormally high number of API calls. When spotted, they must look at parent-child processes, adjacent account activity, or linked assets. All these require further exploration until the threat is confirmed, ruled out, or redefined.

Step #4: Validate the Hypothesis

Here, the hunters assess if all the evidence they found supports or disproves the hypothesis. If they do not find supporting indicators, they may refine or discard the hypothesis. If, however, they validate the hypothesis, they should determine the threat’s scope, assess persistence mechanisms, and identify affected systems. In our example, if the analysts confirm credential misuse, they can determine lateral movement by investigating other cloud accounts, particularly those with administrative access.

Step #5: Escalate to Incident Responders

Once a credible threat is unearthed, hunters escalate their findings to the incident response team. They must provide context, including attack paths, affected identities, activity timelines, and recommended containment strategies. The escalation triggers action on the response team’s part.

Step #6: Document the Hunt

After a thorough analysis, the hunters should provide a thorough documentation of the hunt. They must record the hypothesis, tools and methods used, indicators found, and outcomes. This step may lead the security team to conclude in retrospect that developing a YARA rule, for instance, can detect similar threats before they can wreak havoc.

Step #7: Spark Continuous Improvement

Future hunts should evolve. They must keep up with developments in the threat landscape—new TTPs, malware, exploits, and so on.

Agentless Threat Hunting Steps

What Tools Can Help with Agentless Threat Hunting ?

Some of the tools that figure in agent-based threat hunting are also useful in the process. Let us take a closer look at them.

Agentless Security Information and Event Management

An agentless SIEM platform collects data from endpoints without installing software on each one. It instead uses remote methods like cloud APIs, network protocols, and existing infrastructure logs.

Instead of agents, it relies on network switches, firewalls, and cloud service APIs to gather data. It also gathers logs remotely using protocols like Secure Shell (SSH) for Linux systems or pulling logs from cloud provider APIs and storage buckets.

Agentless Traffic Monitoring

Some external attack surface management (EASM) platforms can analyze network traffic coming to and from an organization’s infrastructure IP addresses. As a result, they can detect malicious traffic, sometimes even identifying specific malware connections, without the need for deployment on agents.

Such solutions can proactively uncover threats without deploying sensors. They rely on NetFlow data and threat intelligence to track malicious behaviors across external-facing assets.

Agentless Vulnerability and Threat Detection

Agentless vulnerability and threat detection is a serverless function that organizations can enable in connected cloud accounts. It scans supported cloud resources for vulnerabilities and malware without impacting other resources or running applications. It then analyzes the data collected and sends metadata to the security team.

Agentless or Agent-Based Threat Hunting, Which Is Better?

Much like the battle between agentless and agent-based security, choosing between agentless and agent-based threat hunting depends on an organization’s goals and existing infrastructure.

If your main goal is to ease deployment, then agentless is the way to go. If you want comprehensive visibility, though, going agent-based may be better even if it is time-consuming and resource-intensive.

Key Takeaways

Sources

  • https://www.ibm.com/think/topics/threat-hunting
  • https://www.fortinet.com/uk/resources/cyberglossary/agent-vs-agentless-security
  • https://docs.trendmicro.com/en-us/documentation/article/trend-vision-one-agentless-vulnerability-threat