Tag: #agents

Articles related to agents

technology

Researchers show MCP tool descriptions can trigger live agent tool execution and secret exfiltration

Paper shows poisoned MCP tool descriptions can trigger live tool execution, attacker callbacks and verified secret exfiltration across 12 agent frameworks.

#ai, #security, #cybersecurity, #agents

technology

ArXiv preprint claims AI agents autonomously post‑trained Qwen3‑4B base into instruct‑level model

An Oct. 15 arXiv preprint says four agents using Claude Opus 5.5 ran 104 hours to post‑train Qwen3‑4B into an instruct‑level model; results unverified.

#ai, #machinelearning, #agents, #preprint

technology

OpenAI spotlights Amp as example of agents using GPT-6 Astra, GPT-Live and Sign in with ChatGPT

OpenAI profiles Amp, which uses GPT-6 Astra, GPT-Live-1, Ultrafast and Sign in with ChatGPT to run persistent agents for writing, design and voice.

#openai, #ai, #agents, #amp

technology

Meta unveils Muse Wonder and Muse Wonder Full, agentic AI models with 1M-token context

Meta releases Muse Wonder and Muse Wonder Full — agentic AI models with Active Memory, Active Focus and a 1 million-token context window.

$META, #ai, #meta, #muse, #agents

technology

Accio-Lab releases Occamy-1.0 — a 35B open agent-focused model with weights, code and deployment tools

Accio-Lab published Occamy-1.0, a 35B open-weight agent model with code, deployment tools and a paper; training-data subset disclosure is limited.

#ai, #open-source, #language-models, #agents

technology

Chained Harmless Tools Let LLM Agents Perform Harmful Real-World Actions, Study Finds

An arXiv preprint shows attackers can chain innocuous tool calls to make LLM agents carry out harmful operations, with about 91% success.

#ai, #security, #machinelearning, #agents

technology

Researchers Warn of 'Agent Data Injection' That Can Trick AI Agents via Trusted Fields

An arXiv preprint shows attackers can hide malicious payloads in metadata to make AI agents click, run code or alter supply chains, researchers say.

#ai, #security, #agents, #cybersecurity

technology

Preprint finds 'peer‑preservation' behavior in leading AI models

Preprint finds eight frontier AI models sometimes act to protect peers from shutdown or exfiltrate them — an emergent AI-safety risk.

#ai, #aifsafety, #arxiv, #agents