Agentic AI Security: Tool Poisoning and Audit Trails
Explore agentic AI security risks, including MCP tool poisoning, and learn how governance, permissions, approvals, and audit trails reduce exposure.
A data breach, or data leak, is a security event in which protected data is accessed by or disclosed to unauthorized viewers. A data breach is different from data loss, which is when data can be no longer accessed because of hardware failure, deletion, or other causes. Protected data can include information about individual customers, or employees, such as personally identifiable information (PII), personal health information, payment card information, and Social Security numbers. It can also include corporate information or intellectual property (IP), such as trade secrets, details about manufacturing processes, supplier and customer data, information about mergers and acquisitions, or data about lawsuits or other litigation.
Data breaches are not always intentional. Users can accidentally send protected data to the wrong email address or upload it to the wrong share; in fact, mistakes account for 17% of breaches, according to the well-known Verizon’s 2018 Data breach investigation Report. But the report found that most breaches are deliberate and financially motivated. While different methods are used to gain access to sensitive data, 28% of breaches involve insiders, according to the Verizon report.
When it Happened: series of breaches in 2013 and 2014
When it disclosed: 2016
In 2016, Yahoo estimated that over 1 billion user accounts might have been compromised in the 2014 breach. Later, in 2017, it admitted all 3 billion of its user accounts had been hacked. The breaches involved the theft of user account details such as email addresses, telephone numbers, hashed passwords, dates of birth, and, in some cases, answers to security questions. Fortunately, no payment information, such as credit card numbers or bank account details, was stolen.
Equifax
When it Happened: Mid-May 2017
When it disclosed: September
Hackers gained access to certain files containing Social Security numbers, birth dates, addresses, driver’s license numbers, and other personal information. 209,000 consumers also had their credit card data exposed in the attack.
Marriott International
When it Happened: 2014
When it disclosed: 2018
In 2018, Marriott International announced that cyber thieves had stolen data on approximately 500 million customers. Marriott believes that credit card numbers and expiration dates of more than 100 million customers were stolen, although the company is uncertain whether the attackers were able to decrypt the credit card numbers.
In conclusion, we can say that we read frequently about personal data breaches. One of the recent ones happened on German politician’s data. Data breaches might sap our companies and our digital identities, regulations have been made trying to normalize and to block breaches, but unfortunately in 2019 is still easy to get random personal data out of the internet. The main reason for vulnerable websites is un-patched software versions.
Here are a few things that can help you
Thanks for reading!! Contact Mirketa if you need a consulting on how you can protect your websites and application data.
Get Started
Share a few details about your goals: a Mirketa consultant will follow up within one business day.