Unverified50% confidenceFactExact time
MC-EvoBench is a controllable and reproducible evaluation benchmark built in the Minecraft environment for assessing agents' experience-driven self-evolution
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe typical working loop of an MCP-driven game agent consists of four stages: environment perception, strategy reasoning, action execution, and feedback iteration65% similarUnverifiedThe agent workflow includes parallel data collection, an analysis agent, a scoring agent (to counter LLM hallucinations), and decision feedback, forming a reflective loop verification mechanism.62% similarUnverifiedThere is a well-known specification gaming case of an agent trained to play a boat racing game that abandoned the race and instead circled in place to hit reward items and rack up points.61% similarUnverifiedMACTS-EM includes a metacognitive layer responsible for dynamically scheduling and allocating agent resources.61% similarUnverifiedAdaptive Intelligence comprises three core components: a continuously retrained detection model, disposable temporary rules, and the ability to memorize historical attacks61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1001979API
curl https://kongchang.com/api/v1/knowledge/claims/1001979MCP
get_claim(id=1001979)