Incident Score: Analysis & Impact (DELNEUSHACRECAM1789209035)
The details regarding individual company incidents & reports gives you full view from every side.
Rankiteo Score Impact Analysis
Key Highlights From The Incident Analysis
- Timeline of Deloitte's Vulnerability and lateral movement inside company's environment.
- Overview of affected data sets, including SSNs and PHI, and why they materially increase incident severity.
- How Rankiteo’s incident engine converts technical details into a normalized incident score.
- How this cyber incident impacts Deloitte Rankiteo cyber scoring and cyber rating.
- Rankiteo’s MITRE ATT&CK correlation analysis for this incident, with associated confidence level.
Full Incident Analysis Transcript
In this Rankiteo incident briefing, we review the Deloitte breach identified under incident ID DELNEUSHACRECAM1789209035.
The analysis begins with a detailed overview of Deloitte's information like the linkedin page: https://www.linkedin.com/company/deloitte, the number of followers: 0, the industry type: Business Consulting and Services and the number of employees: 533222 employees
After the initial compromise, the video explains how Rankiteo's incident engine converts technical details into a normalized incident score. The incident score before the incident was 803 and after the incident was 801 with a difference of -2 which is could be a good indicator of the severity and impact of the incident.
In the next step of the video, we will analyze in more details the incident and the impact it had on Deloitte and their customers.
On 01 April 2026, a cybersecurity incident called "Enterprise AI Agents Pose Unprecedented Security Risks" came to light.
In late 2024, enterprise AI evolved into autonomous agents capable of acting, planning, and executing tasks without human intervention, creating a new threat surface.
The disruption is felt across the environment, affecting File systems, APIs and Databases, and exposing True, plus an estimated financial loss of $4.7 million (average cost of AI agent-related data breach).
Formal response steps have not been shared publicly yet.
The case underscores how teams are taking away lessons such as Traditional security tools are inadequate for AI agents; runtime threats require behavioral-based security models; enterprises lack visibility into agent deployments; governance frameworks are immature, and recommending next steps like Adopt NeuralTrust’s *Agentic AI Security at Runtime* framework (Observe, Enforce, Detect, Respond), Implement real-time monitoring of agent actions and Apply semantic-level controls to restrict unintended behavior.
Finally, we try to match the incident with the MITRE ATT&CK framework to see if there is any correlation between the incident and the MITRE ATT&CK framework.
The MITRE ATT&CK framework is a knowledge base of techniques and sub-techniques that are used to describe the tactics and procedures of cyber adversaries. It is a powerful tool for understanding the threat landscape and for developing effective defense strategies.
MITRE ATT&CK® Correlation Analysis
Rankiteo's analysis has identified several MITRE ATT&CK tactics and techniques associated with this incident, each with varying levels of confidence based on available evidence. Under the Initial Access tactic, the analysis identified Supply Chain Compromise: Compromise Software Dependencies and Development Tools (T1195.002) with moderate to high confidence (70%), supported by evidence indicating open-source agent frameworks like LangChain, CrewAI, AutoGen, Semantic Kernel and Phishing: Spearphishing Attachment (T1566.001) with moderate to high confidence (80%), supported by evidence indicating malicious payloads embedded in documents or API responses redirect agent behavior. Under the Execution tactic, the analysis identified Command and Scripting Interpreter: JavaScript (T1059.007) with moderate confidence (60%), supported by evidence indicating aI agents execute tasks dynamically based on external inputs (documents, APIs) and User Execution: Malicious File (T1204.002) with moderate to high confidence (70%), supported by evidence indicating prompt injection attacks manipulate agent behavior via adversarial content. Under the Persistence tactic, the analysis identified Compromise Client Software Binary (T1554) with moderate confidence (60%), supported by evidence indicating agents retain memory across sessions, enabling persistent corruption and Event Triggered Execution: Accessibility Features (T1546.008) with moderate confidence (50%), supported by evidence indicating unmonitored prototype agents remain in production, creating blind spots. Under the Privilege Escalation tactic, the analysis identified Valid Accounts: Cloud Accounts (T1078.004) with moderate to high confidence (80%), supported by evidence indicating compromised agent with access to email, databases, or file storage becomes trusted insider and Abuse Elevation Control Mechanism: Bypass User Account Control (T1548.002) with moderate confidence (60%), supported by evidence indicating multi-agent trust propagation inherits elevated permissions. Under the Defense Evasion tactic, the analysis identified Masquerading: Match Legitimate Name or Location (T1036.005) with moderate to high confidence (70%), supported by evidence indicating prompt injection evades static analysis by manipulating agent behavior and Impair Defenses: Disable or Modify Tools (T1562.001) with moderate confidence (60%), supported by evidence indicating 48% of AI agents operate without meaningful security controls. Under the Credential Access tactic, the analysis identified Steal Application Access Token (T1528) with moderate to high confidence (70%), supported by evidence indicating tool & permission abuse grants attackers broad access via compromised agents. Under the Discovery tactic, the analysis identified Account Discovery: Cloud Account (T1087.004) with moderate confidence (60%), supported by evidence indicating agents interact with file systems, APIs, databases, and email inboxes. Under the Collection tactic, the analysis identified Data from Local System (T1005) with moderate to high confidence (80%), supported by evidence indicating real enterprise data handled by AI agents is compromised and Automated Collection (T1119) with moderate to high confidence (70%), supported by evidence indicating agents dynamically collect data from external inputs (documents, APIs). Under the Exfiltration tactic, the analysis identified Exfiltration Over C2 Channel (T1041) with moderate to high confidence (70%), supported by evidence indicating average cost of AI agent-related data breach reached $4.7 million and Transfer Data to Cloud Account (T1537) with moderate confidence (60%), supported by evidence indicating compromised agents interact with cloud-based systems (APIs, databases). Under the Impact tactic, the analysis identified Resource Hijacking (T1496) with moderate to high confidence (70%), supported by evidence indicating compromised agents act as trusted insider threats, enabling cascading attacks and Data Destruction (T1485) with moderate confidence (50%), supported by evidence indicating memory/context poisoning corrupts agent operations undetected. These correlations help security teams understand the attack chain and develop appropriate defensive measures based on the observed tactics and techniques.
Sources & References
- Deloitte Rankiteo Cyber Incident Details: https://www.rankiteo.com/company/deloitte/incident/DELNEUSHACRECAM1789209035
- Deloitte CyberSecurity Rating page: https://www.rankiteo.com/company/deloitte
- Deloitte Rankiteo Cyber Incident Blog Article: https://blog.rankiteo.com/delneushacrecam1789209035-deloitte-shattered0io-neuraltrust-crewai-semantic-kernel-vulnerability-december-2025/
- Deloitte CyberSecurity Score History: https://www.rankiteo.com/company/deloitte/history
- Deloitte CyberSecurity Incident Source: https://www.cybersecurity-insiders.com/48-of-your-ai-agents-are-running-without-meaningful-security-controls/
- Rankiteo A.I CyberSecurity Rating methodology: https://www.rankiteo.com/Images/rankiteo_algo.pdf
- Rankiteo TPRM Scoring methodology: https://static.rankiteo.com/model/rankiteo_tprm_methodology.pdf