Connect with us

Tech

ElizaOS Exposed: Researchers Discover How to Manipulate Its Memory and Alter Its Operations

Published

1 year ago

on

TL;DR

  • Princeton researchers discovered how to manipulate the memory of AI agents like ElizaOS to alter their financial decisions.
  • The memory injection attack allows fake memories to be inserted into AI systems, causing harmful transactions.
  • CrAIBench, the tool created by Princeton, measures AI agents’ resistance to contextual manipulations and social attacks.

A group of researchers from Princeton University, in collaboration with the Sentient Foundation, identified a critical vulnerability in artificial intelligence agents operating on blockchains. The study focused on ElizaOS, a popular open-source framework used to automate financial operations on decentralized networks, and revealed a method for manipulating its memory.

How AI Agents Are Manipulated

The attack, known as memory injection, allows false data to be inserted into an AI agent’s persistent memory. This information is stored and influences the system’s future decisions without triggering any alert. While it does not directly compromise the blockchains, it causes harmful transactions driven by data that was externally manipulated. The researchers successfully demonstrated the effectiveness of this technique by using social platforms to generate fake memories within ElizaOS.

The agents most affected are those that adjust their activity based on social perception. In these cases, attackers create fake profiles and post coordinated messages that artificially alter the sentiment around a token. This causes the AI to purchase overvalued assets, only to get trapped in a price drop planned by the attackers themselves. This type of maneuver, known as a Sybil attack, becomes more effective when combined with memory manipulation.

ElizaOS post

ElizaOS Works with Researchers to Find a Solution

The Princeton team thoroughly examined all the functionalities of ElizaOS to design realistic and complete attacks, which revealed the broad range of available vectors when an AI has multiple plugins and access to financial operations. From these trials, the researchers developed CrAIBench, a testing system that measures the resistance of different AI agents to contextual manipulations.

The results have already been shared with Eliza Labs, the company responsible for the framework. Discussions about possible solutions are ongoing. The study concludes that protecting these systems requires improvements in both memory management and language model capabilities. So they can better distinguish between legitimate data and malicious instructions.

The post ElizaOS Exposed: Researchers Discover How to Manipulate Its Memory and Alter Its Operations appeared first on The Cryptocurrency Post.

Related Topics:
Continue Reading

Tech

Bittensor SN10 Work Tied to Pareto AI and Engy AI at Exploit Summit

Published

2 days ago

on

October 1, 2026

By

An official Exploit Summit post said @ExploitSummit that @xavi3rlu announced a partnership between pareton_ai and @engyai aimed at optimizing Engy inference latency and throughput using SN10 work.

The announcement frames the collaboration as a technical optimization effort rather than a broader product launch. According to the conference account, the work is intended to improve Engy’s inference performance through SN10-related optimization, with the stated goal of supporting Engy customers.

That is the clearest confirmed detail currently available: the partnership was presented publicly at the event, and the stated focus was on latency and throughput. No additional operational specifics were included in the post, such as implementation timing, deployment scope or whether the optimization is already live for users.

SN10’s role in the announcement

The mention of SN10 places Bittensor-related optimization work at the center of the update, but the announcement does not spell out technical mechanics beyond performance improvements for inference workloads. The post does not describe any token, market or governance implications, and it does not indicate whether the collaboration changes access, pricing or product availability.

For readers following AI-linked crypto projects, the update is best understood as a partnership claim around infrastructure performance. It signals that Pareto AI and Engy AI are positioning SN10 work as part of an effort to make inference more efficient, but the announcement itself stays narrow in scope.

The post Bittensor SN10 Work Tied to Pareto AI and Engy AI at Exploit Summit appeared first on The Cryptocurrency Post.

Continue Reading

Tech

BNB Chain Agent Studio v4 adds NodeOpsHQ for prompt-to-agent deployment

Published

1 week ago

on

September 25, 2026

By

BNB Chain said NodeOps is now a deployment option in BNB Agent Studio v4, allowing builders to move from a prompt to a running agent without creating a cloud account or configuring infrastructure.

The update is part of a broader v4 release for BNB Agent Studio, which BNB Chain described as aimed at reducing the manual steps between writing a prompt and getting an agent live. In the company’s framing, the new NodeOps path removes a setup layer that would normally sit between development and deployment.

NodeOps also said agents can go live in BNB Agent Studio v4 using its infrastructure without manual setup. BNB Chain’s blog post described NodeOps alongside AWS and Azure as live deployment options.

The same release also adds stablecoin payment support for agents, with USDT, USDC and USD1 now working alongside $U. BNB Chain said wallet setup, funding and password handling are now automatic, while documentation, error messages and feedback tools have been improved to reduce friction for developers.

The changes make the studio feel more like a streamlined build environment than a setup-heavy toolchain, but the announcement remains centered on deployment convenience rather than a broader shift in the agent market. For developers using the platform, the practical change is that the path from prompt to live agent now appears to involve fewer manual steps.

The post BNB Chain Agent Studio v4 adds NodeOpsHQ for prompt-to-agent deployment appeared first on The Cryptocurrency Post.

Continue Reading

Tech

BNB Chain releases Agent Studio v4 with simplified AI agent deployment and stablecoin support

Published

1 week ago

on

September 24, 2026

By

BNB Chain has released Agent Studio v4, updating its AI agent builder with a simpler deployment path, automatic wallet setup and broader stablecoin payment support.

In its announcement, BNB Chain said developers can now deploy through NodeOps without setting up cloud infrastructure themselves. The company also said the updated flow reduces manual steps between creating an agent and getting it running.

The release adds support for USDT, USDC, USD1 and $U as payment options for agents, expanding the stablecoin set available inside the studio. BNB Chain said the change is meant to let developers use the assets they already hold without an extra conversion step.

Agent Studio v4 also automates parts of wallet setup that previously required more manual work. According to the company, TWAK wallet creation now runs in the background, new wallets are funded automatically with test BNB and a test stablecoin, and local wallets receive generated passwords instead of requiring manual setup.

BNB Chain framed the update as part of a broader effort to reduce the friction between prompting an agent and having it live. The company described NodeOps as one of three deployment paths now available in the studio, alongside options for developers using their own cloud accounts.

For teams building on BNB Chain, the main change is practical rather than theoretical: fewer setup steps, more deployment options and a wider choice of stablecoins for agent transactions.

The post BNB Chain releases Agent Studio v4 with simplified AI agent deployment and stablecoin support appeared first on The Cryptocurrency Post.

Continue Reading

Trending