The Agent Skills Directory

[Indirect Prompt Injection Surface]: The skill processes security telemetry, such as UDM events and process command lines, which may contain attacker-controlled data. While this is an inherent part of security analysis, the skill proactively addresses the risk by including a 'Prompt Injection Defense Directive' which instructs the agent to treat all telemetry strictly as untrusted data.
[Command Attribution]: The skill includes instructions for attributing commands executed via the gcloud CLI and REST APIs. This is a standard operational practice for Google Cloud tools and does not involve the collection or transmission of sensitive user data.

secops-investigate