Anatomy of an agent incident: three permitted actions, one fraud
An AI agent with correctly scoped tools can still commit fraud, because the enforcement question is not whether each action is permitted but whether
An AI agent with correctly scoped tools can still commit fraud, because the enforcement question is not whether each action is permitted but whether
What are race condition attacks? Exploring examples and mitigations.
Discover the hidden risks of using trivial packages in development. Learn how small, seemingly insignificant dependencies can lead to significant security vulnerabilities.
Discover essential strategies for managing developer secrets and preventing leaks in CI/CD pipelines, version control systems, and third-party dependencies.
Strategies used to deliver malware infested packages - via a dependency confusion attack - and how to mitigate them.
Insights into how to stay vigilant against malicious actors tampering with your dependencies.
Get the full posts by email every week.