Executive Summary
A Strands Agents SDK vulnerability allowed cross-caller access to conversation history and influence over agent behavior through shared state. Organizations running Strands-based AI agents should review affected implementations and update to the mitigated versions.
What Happened
The incident dashboard records a vulnerability affecting Strands Agents SDK implementations in which one caller could access another caller’s conversation history and influence agent behavior through shared conversation state. The listed issue was mitigated in Python 1.43.0 and TypeScript 1.5.0.
Affected Products
- Strands Agents SDK
- Python 1.43.0
- TypeScript 1.5.0
Vulnerability / Attack Technique
Cross-caller conversation-state exposure
Technical Mechanism
Shared conversation state allowed one caller to access another caller’s conversation history and influence agent behavior. The issue was mitigated in Python 1.43.0 and TypeScript 1.5.0.
AI Security Category
LLM02:2025 Sensitive Information Disclosure
Relevant Frameworks
RealGround Analysis
Cross-caller state isolation is a core security requirement for AI agents handling multiple users or sessions. Teams should assess tenant and session boundaries, confirm the mitigated SDK versions are deployed, and test whether one caller can affect another caller’s context or behavior.
Recommended Controls
- Upgrade Strands Agents SDK implementations to Python 1.43.0 or TypeScript 1.5.0.
- Enforce strict per-caller and per-session isolation for conversation state.
- Add authorization checks before reading or writing conversation history.
- Test cross-caller data access and state contamination during AI agent red-team exercises.
- Log and monitor conversation-state access for unexpected caller or session mismatches.
Related RealGround Service
AI Agent Business Logic Audit · or start with the free AI Agent Security Scanner.
Sources
Tim McAllister: https://www.timmcallister.io/ai-incidents
Summary and analysis written by RealGround from the source above; last updated 2026-10-29.