Policy control threat detection

A method comprising:

• analyzing, by a first compliance agent of two or more compliance agents, the first compliance agent comprising a first machine learning generated model executing on one or more processors, first components associated with an application at a first phase of a software development cycle for the application as input to the first compliance agent to determine compliance or non-compliance of the first components according to a first set of policies as output from the first compliance agent, wherein the first machine learning generated model is trained based on the first set of policies,
• analyzing, by a second compliance agent of the two or more compliance agents, the second compliance agent comprising a second machine learning generated model executing on the one or more processors, second components associated with the application at a second phase of the software development cycle for the application as input to the second compliance agent to determine compliance or non-compliance of the second components according to a second set of policies as output from the second compliance agent, wherein the second set of policies is different than the first set of policies, and wherein the second machine learning generated model is trained based on the second set of policies;
• and
• determining, by a compliance analyzer, a risk of the application across the software development cycle based on outputs from the two or more compliance agents, including the output from the first compliance agent and the output from the second compliance agent.

A non-transitory computer readable medium comprising instructions that, when executed, cause one or more processors to:

• analyze, by a first compliance agent of two or more compliance agents, the first compliance agent comprising a first machine learning generated model executing on the one or more processors, first components associated with an application at a first phase of a software development cycle for the application as input to the first compliance agent to determine compliance or non-compliance of the first components according to a first set of policies as output from the first compliance agent, wherein the first machine learning generated model is trained based on the first set of policies;
• analyze, by a second compliance agent of the two or more compliance agents, the second compliance agent comprising a second machine learning generated model executing on the one or more processors, second components associated with the application at a second phase of the software development cycle for the application as input to the second compliance agent to determine compliance or non-compliance of the second components according to a second set of policies as output from the second compliance agent, wherein the second set of policies is different than the first set of policies, and wherein the second machine learning generated model is trained based on the second set of policies;
• and
• determine, by a compliance analyzer, a risk of the application across the software development cycle based on outputs from the two or more compliance agents including the output from the first compliance agent and the output from the second compliance agent.