AWS says AI agents lack business context and security, launches two services to patch the gaps

2026-06-21 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Cybersecurity & Data Privacy · Depth: Intermediate, medium

Summary

Amazon Web Services (AWS) launched two new services, AWS Continuum and AWS Context, at the AWS Summit in New York on June 21, 2026, to enhance AI agent production readiness. AWS Continuum automates the full lifecycle of code vulnerability management, from detection and prioritization to validation and remediation, initially for pilot customers. It uses specialized security models like Anthropic's Claude Mythos to identify and fix risks, operating in a learning mode before enforcement. AWS Context builds a shared knowledge graph from enterprise data, including databases, documents, emails, and chat messages, providing AI agents with essential business context to prevent incorrect recommendations. The AWS DevOps Agent also gained Release Readiness Review and dynamic test plan generation for AI-generated code, available in preview in the US East region. AWS further released its coding agent Kiro as a native iOS app and expanded Bedrock AgentCore with connectors to S3, SharePoint, Confluence, Google Drive, and security filters.

Key takeaway

For MLOps Engineers deploying AI agents, you should integrate AWS's new services to mitigate critical production risks. AWS Continuum can automate vulnerability management, reducing security backlogs and matching the pace of AI-generated code. Simultaneously, AWS Context provides agents with essential business knowledge via a knowledge graph, preventing erroneous outputs. Consider using the AWS DevOps Agent's new features for rigorous testing of AI-generated code before deployment, ensuring stability and compliance.

Key insights

AWS addresses AI agent production challenges by integrating security automation and business context via new services.

Principles

• AI agents require explicit business context.
• Automated security must match AI-generated code speed.
• Validate AI-generated code in production-like environments.

Method

AWS Continuum automates vulnerability management: discover, prioritize by business impact, validate exploitability in isolated tests, then remediate with specific countermeasures.

In practice

• Use AWS Context to build knowledge graphs for agents.
• Implement AWS Continuum for automated vulnerability fixes.
• Deploy AWS DevOps Agent for AI-generated code review.

Topics

• AI Agents
• AWS Continuum
• AWS Context
• Code Security
• Knowledge Graphs
• DevOps Agent

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, MLOps Engineer, AI Security Engineer

Related on AIssential

• AWS unveils agent security, data access tools — The rapid adoption of AI agents necessitates new security and data access tools to manage evolving enterprise risks.
• AWS enters the context layer race with a graph that learns from agents, not manual curation — AWS's new context layer automatically builds and refines knowledge graphs through agent interaction, reducing manual curation.
• AI needs a strong data fabric to deliver business value — Enterprise AI success hinges on data context and judgment, not just speed, necessitating a robust data fabric.
• AWS Announces General Availability of DevOps Agent for Automated Incident Investigation — AWS DevOps Agent uses generative AI to autonomously investigate incidents and automate operational tasks across diverse environments.
• AWS Quick's personal knowledge graph is making orchestration decisions most control planes can't see — AWS Quick evolves into a proactive, desktop-native AI agent with a persistent knowledge graph, raising enterprise governance concerns.
• AWS Introduces New Cybersecurity, Coding Tools at NY Summit — AWS is addressing AI-driven software engineering challenges with new AI products, including a code vulnerability tool.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.