CompTIA CAS-005 Dumps Questions Answers

Model poisoning occurs when an attacker manipulates the training data or the training process of an AI model so that its predictions are deliberately inaccurate or biased. In the SecurityX CAS-005 objectives, this is part of understanding emerging technology threats, specifically AI/ML vulnerabilities. This differs from:

Social engineering, which manipulates humans rather than AI models.

Output handling, which deals with how outputs are processed but doesn’t cause inaccuracy at the model level.

Prompt injections, which manipulate the model at query time, not during training.Because model poisoning directly corrupts the AI model itself, it is the clearest reason AI outputs could be inaccurate.

Analysis and Remediation Options for Each IoC:

IoC 1:

Evidence:

Source: Apache_httpd

Type: DNSQ

Dest: @10.1.1.1:53,@10.1.2.5

Data: update.s.domain, CNAME 3a129sk219r9slmfkzzz000.s.domain, 108.158.253.253

Analysis:

Analysis: The service is attempting to resolve a malicious domain.

Reason: The DNS queries and the nature of the CNAME resolution indicate that the service is trying to resolve potentially harmful domains, which is a common tactic used by malware to connect to command-and-control servers.

Remediation:

Remediation: Implement a blocklist for known malicious ports.

Reason: Blocking known malicious domains at the DNS level prevents the resolution of harmful domains, thereby protecting the network from potential connections to malicious servers.

IoC 2:

Evidence:

Src: 10.0.5.5

Dst: 10.1.2.1, 10.1.2.2, 10.1.2.3, 10.1.2.4, 10.1.2.5

Proto: IP_ICMP

Data: ECHO

Action: Drop

Analysis:

Analysis: Someone is footprinting a network subnet.

Reason: The repeated ICMP ECHO requests to different addresses within a subnet indicate that someone is scanning the network to discover active hosts, a common reconnaissance technique used by attackers.

Remediation:

Remediation: Block ping requests across the WAN interface.

Reason: Blocking ICMP ECHO requests on the WAN interface can prevent attackers from using ping sweeps to gather information about the network topology and active devices.

IoC 3:

Evidence:

Proxylog:

GET /announce?info_hash=%01dff%27f%21%10%c5%wp%4e%1d%6f%63%3c%49%6d&peer_id%3dxJFS

Uploaded=0&downloaded=0&left=3767869&compact=1&ip=10.5.1.26&event=started

User-Agent: RAZA 2.1.0.0

Host: localhost

Connection: Keep-Alive

HTTP200 OK

Analysis:

Analysis: An employee is using P2P services to download files.

Reason: The HTTP GET request with parameters related to a BitTorrent client indicates that the employee is using peer-to-peer (P2P) services, which can lead to unauthorized data transfer and potential security risks.

Remediation:

Remediation: Enforce endpoint controls on third-party software installations.

Reason: By enforcing strict endpoint controls, you can prevent the installation and use of unauthorized software, such as P2P clients, thereby mitigating the risk of data leaks and other security threats associated with such applications.

A crisis management plan defines coordinated communication and response strategies for high-profile incidents that may harm an organization’s public reputation and shareholder confidence. CAS-005 GRC content includes crisis communication planning for regulatory compliance and public relations in the wake of breaches.

A Data Subject Access Request (A) addresses individual data rights, not overall crisis handling.

Business Impact Analysis (B) helps assess potential operational and financial impacts but does not manage public perception during an incident.

Supply chain management (C) is preventative for vendor risks, not responsive to current crises.