CIS-1355 · Security Engineering · Module 5

Endpoint Vulnerabilities, Attacks, and Defenses

Recognize malware behavior, trace application vulnerabilities, and choose controls that protect endpoints.

Group malware by what it does

Malware is software or code designed to carry out harmful or unauthorized activity. The five categories below describe behavior. A single attack can combine several of them.

Kidnap
Screen-locking ransomware blocks normal access to the device. File-encrypting ransomware denies access to data and may reach mounted shares or attached storage. Extortion can combine encryption with data theft and threats to publish stolen files.
Eavesdrop
A software or hardware keylogger captures keystrokes. Spyware can monitor activity or collect sensitive information such as browser data, screen contents, or keystrokes. Its capabilities vary.
Masquerade
A Trojan appears legitimate or useful while performing a malicious operation. A remote access Trojan (RAT) gives an attacker unauthorized remote control or access.
Launch
A virus replicates by infecting a host such as a file. A worm spreads by replicating across systems, often by exploiting a network-accessible weakness. A bot is a remotely controlled device, and a botnet is a collection of them that can be used for attacks such as distributed denial of service.
Sidestep
A logic bomb waits for a trigger before running its payload. A rootkit can hide attacker activity or maintain privileged access. A backdoor bypasses the normal access path or authentication controls.

Fileless techniques can execute code in memory or abuse legitimate scripts, utilities, and services, often called living-off-the-land binaries (LOLBins) when they are executables already available in the environment. These techniques can evade scans that look only at files. Script inspection and behavior-based monitoring can still detect them. Bloatware is unwanted preinstalled or bundled software. It can add unnecessary services, permissions, or update requirements even when the software is legitimate.

CISA and partners · StopRansomware Guide (PDF) · Microsoft · Script inspection and fileless threats

Interpret indicators in context

An indicator of attack (IoA) is behavior that suggests an attempt or attack sequence, such as an unusual process launching a script and rapidly rewriting files. An indicator of compromise (IoC) is an observable artifact associated with compromise, such as a verified malware hash, a known malicious domain contacted by the host, or an unauthorized persistence entry. An IoC may be found during an active incident or afterward.

Account lockouts, concurrent sessions, impossible-travel alerts, blocked content, resource spikes, unavailable services, and missing or unusually timed logs deserve investigation in context. Compare them with normal activity, scheduled jobs, and published or documented indicators. Correlate identity, endpoint, and network records before deciding what happened. For example, a VPN can produce an impossible-travel alert without account theft.

Recognize application vulnerabilities and attacks

A buffer overflow writes past an allocated buffer and can corrupt memory, crash a process, or allow code execution. Improper exception handling can expose internal information, leave resources insecure, or let operations continue after a failed check. A race condition occurs when correctness depends on the timing of operations on shared state. One example is when a file changes between a permission check and its use. Use memory-safe code where practical, handle errors securely, and make critical operations atomic or properly synchronized.

Directory traversal
Manipulates a path to reach files outside the intended directory. Prefer mapping approved identifiers to files and enforce path and access restrictions.
Cross-site scripting (XSS)
Causes attacker-controlled content to execute in a victim’s browser in the site’s context. Use context-aware output encoding and sanitize HTML when users are allowed to supply markup.
SQL injection (SQLi)
Turns untrusted input into part of a database command. Use parameterized queries to keep data separate from SQL instructions.
Cross-site request forgery (CSRF)
Induces a victim’s browser to send an unwanted request to a site where the victim is authenticated. Use anti-CSRF tokens for state-changing requests and configure SameSite cookies where appropriate.
Server-side request forgery (SSRF)
Induces a server to request a destination chosen by an attacker, potentially reaching internal services. Restrict allowed destinations and outbound access.
Replay
Resends a previously valid message or request. Protocols can use nonces, sequence numbers, or time windows with duplicate detection to reject reuse.

OWASP · Cheat Sheet Series · OWASP · SQL Injection Prevention

Layer endpoint and browser defenses

Protect session cookies with Secure so they are sent over HTTPS, HttpOnly to restrict script access, and an appropriate SameSite policy to limit cross-site sending. HTTP response headers such as Content Security Policy (CSP) and HTTP Strict Transport Security (HSTS) can also constrain browser behavior. CSP can limit executable content, while HSTS directs supporting browsers to use HTTPS for the site.

MDN · Secure cookie configuration · NIST · Enterprise Patch Management · NIST · Application allowlisting

Practice case

A workstation starts rewriting files. EDR alerts when a document application launches a script that rapidly renames files on a shared drive. Antivirus reports no malicious file. File-server logs independently show unexpected writes from the workstation’s account.

Which observations suggest an attack in progress? What would you do first to limit damage? What evidence would help establish the affected files and whether any data was taken?

Review a possible answer

The unusual parent-child process and rapid writes suggest a possible ransomware attack. The antivirus result can coexist with script-based execution. Follow the response plan to isolate the workstation and coordinate with the file-server team to stop ongoing writes. Preserve process, script, identity, and file-server telemetry while containing the incident.

Use file-server logs and backup or snapshot comparisons to scope the changes. Review outbound network records and relevant cloud activity for signs of transfer. File changes show an impact on the share. Establishing exfiltration requires evidence of a transfer.

Read and watch