Know When to Trust the Skill: Delayed Appraisal and Epistemic Vigilance for Single-Agent LLMs

ArXi:2604.16753v1 Announce Type: new As large language models (LLMs) transition into autonomous agents integrated with extensive tool ecosystems, traditional routing heuristics increasingly succumb to context pollution and "overthinking". We argue that the bottleneck is not a deficit in algorithmic capability or skill diversity, but the absence of disciplined second-order metacognitive governance.