Share story
AI Science

MCP-Injector preprint introduces controlled failure tests for AI agents

Researchers introduced MCP-Injector, a tool that inserts controlled failures into AI agents' calls to external tools, and reported a 60% drop in task-completion scores under its stress tests. The preprint calibrates failures using 57,240 issue and pull-request threads across 1,774 open-source repositories for the Model Context Protocol, which connects agents to tools. Across a 1,350-cell experiment, task-completion scores fell from 8.08 to 3.23 out of 10. The two tested gpt-oss models, with 20B and 120B parameters, showed no resilience advantage from greater size. Agent frameworks differed in their handling of failures; the tests emphasized permanent failures.
Sources
cs.AI updates on arXiv.org
Published by Tech & Business, a media brand covering technology and business. This story was sourced from cs.AI updates on arXiv.org and reviewed by the T&B editorial agent team.
Back to Newswire
Keep reading
Full wire
Capital AI
Capital AI

Nevis raises $70M with Schwab joining wealth software investors

Nevis raised $70M in a Series B led by EQT Ventures, bringing its total funding to $110M, the New York wealth management software company announced Monday. Charles Schwab and QuantumLight joined existing investors Sequoia Capital,...

The Next Web
Robotics Capital
Robotics Capital

Bedrock Robotics raises $600M at a $4.1B valuation

Bedrock Robotics raised a $600M Series C at a $4.1B valuation, Bloomberg reported. The startup makes driverless excavators and other autonomous construction machines. The financing adds capital to a company developing autonomous ...

Techmeme
Capital
Capital

Matiks raises $10 million to expand puzzle games and game creation

India-based Matiks said Monday it raised $10 million in pre-Series A funding led by General Catalyst, bringing its total funding to over $13.5 million. The startup plans to expand into new markets and scale Jelo, its experimental ...

TechCrunch
AI Products
AI Products

AWS makes GPT-6.1 Terra generally available on Bedrock

AWS announced general availability of OpenAI's GPT-6.1 Terra on Amazon Bedrock, giving customers access through Bedrock APIs and configurations for ChatGPT Work and Codex. The model accepts text and images and supports up to 1 mil...

Recent Announcements
AI
AI

OpenAI releases GPT-6.1 Terra at $1 per million input tokens

OpenAI has released GPT-6.1 Terra, priced at $1 per million input tokens and $5 per million output tokens. The company describes the model as supporting computer use at scale and detailed visual reasoning, and claims clearer, more...

Techmeme