RSAC and BSidesSF 2026 Takeaways: AI Agents, SOC Architecture, and Supply-Chain Risk
Summary
A conference recap argues that MCP-enabled AI agents and decoupled data architectures are reshaping security operations, while AI-related data-access concerns and recent supply-chain compromises heighten risks.
Key points
- The author says the gap between a working MVP and a production-ready security tool, plus ongoing maintenance, complicates build-versus-buy decisions.
- Security teams are using Model Context Protocol (MCP) to let AI agents query tools and, in some cases, take actions such as isolating endpoints or starting response playbooks.
- The article describes growing interest in decoupled architectures that combine security data lakes, data pipelines, and federated search rather than relying on a SIEM as a one-stop shop.
- The author identifies agentic AI’s access to databases, documents, and APIs as a data-governance concern, particularly when permissions or identities are over-scoped.
- The article reports a TeamPCP supply-chain campaign involving Trivy, Checkmarx KICS, and LiteLLM, and says a hijacked Axios maintainer account pushed a RAT through poisoned npm releases.
Article Details
- Topic
- RSAC and BSidesSF 2026 takeaways on AI security, security operations, and software supply chain risks
MITRE ATT&CK
People
Threat Actors
Vendors
Aqua SecurityMarch alone gave us the TeamPCP cascading campaign that started with compromising Aqua’s Trivy scanner, spread to Checkmarx KICS, and then used stolen CI/CD credentials to poison LiteLLM on PyPI (a library present inCheckmarxgave us the TeamPCP cascading campaign that started with compromising Aqua’s Trivy scanner, spread to Checkmarx KICS, and then used stolen CI/CD credentials to poison LiteLLM on PyPI (a library present in 36% of
Tools
KICSus the TeamPCP cascading campaign that started with compromising Aqua’s Trivy scanner, spread to Checkmarx KICS, and then used stolen CI/CD credentials to poison LiteLLM on PyPI (a library present in 36% of cloudTrivyRecent software supply chain incidents (Trivy, TeamPCP, LiteLLM, npm/PyPI compromises) highlight escalating risk to tooling and dependencies.