LLM agents

Project Zero compared Fuzzilli, a generative JavaScript engine fuzzer, with BigSleep, an LLM-based bug finding system, on Chrome's V8 engine. Over six months BigSleep found more vulnerabilities (45 versus 37), but only 25% of the specific bugs overlapped. Project Zero concludes the two approaches are complementary.

Google Research has introduced WikiSkill, a framework that lets AI agents build a persistent, growing knowledge base from past task attempts instead of discarding what they learn after each run. Tested across five models and five benchmarks, WikiSkill lifted average scores over baseline agents with no skill memory for all five models tested.

IBM Research introduces ALTK-Evolve, an agentic memory system that matches or beats ACE's accuracy on the AppWorld benchmark while cutting inference token costs by up to 86% through selective guideline retrieval instead of full-playbook injection.