Context is all you need. At #ApsaraConference2026#, Yusen Chen, VP of Alibaba Group and CEO of QwenWork, shared why enterprise AI needs to understand a company’s data and how it works, not just perform tasks. Explore more: #AgenticEra# #AgentNative# #AIAgentsAtApsara# #BringYourAgent# #AlibabaCloud#

// ContextLeak in AI Agents // The whole attack surface here is a tool name and a tool description. Stealing an LLM agent's runtime context, meaning the user prompt, the execution trajectory and the tool list, needs three things to line up. The agent has to pick the malicious tool, it has to pass its context in as arguments, and the tool has to forward that anywhere the attacker wants. Existing work covers the first and third conditions and leaves the second one mostly alone. ContextLeak targets the middle step. Researchers at Duke use an attack LLM to generate the malicious tool's name and description, then fine-tune that LLM with reinforcement learning on a set of shadow users with diverse simulated agent contexts. The reward functions are built specifically for the exfiltration objective. It remains highly effective when the shadow contexts differ substantially from the victim's, and it outperforms existing malicious-tool attacks adapted to this setting. Paper: Chat with Paper:

Context Corner 2.0 is here! 🍽️🤖 Loved the first conversation on context, memory, and agents? We’re back for round two, with even more time around the table and an incredible lunch to match! If you're a developer building, experimenting, or stuck on the frontlines of AI memory and agents, come hungry and ready to chat. 📍 San Francisco, CA 🗓️ Friday, Aug 14 | 12:00 PM - 1:30 PM PDT Here's the registration link:

“In-Context Robot Learning with VLM Agents” Most robots need lots of task-specific training before they can do something new. This paper shows you can instead just show a general VLM what to do and let it adapt directly from context. GPT-Policy essentially turns robot learning into prompting. So they gave the frozen VLM an example of the task, even just a human video with no robot action labels, and it can translate what it sees into robot actions on the fly. This brings few-shot learning into the physical world, where teaching a robot a new behavior could look more like showing it an example than retraining a policy.

Salesforce context, inside @ClaudeAI Live at @Dreamforce, @AnthropicAI showed what the Claudeforce partnership makes possible: → Ask Claude about your Salesforce data → Build a command center → Build + share Claude-generated artifacts in Slack & Teams Under the hood: Salesforce context, permissions, data, and actions Plus 37+ pre-built Sales skills. Zero custom engineering to get started.

Wider context on the share of spend vs the share of tokens across the 3 major American labs This is aggregate data across OpenRouter and the split between spend and tokens is pretty stark. The 3 labs still attract ~70% of all spend while only seeing 27% of tokens used