The paper surveys self‑evolving agents, highlighting that their states—memories, tools, skills, workflows, and inter‑agent relations—are dynamic and can be modeled as evolving graphs. It critiques existing surveys for treating graphs merely as support structures and proposes a framework that views agent evolution as dynamic graph transformation, categorizing methods into node/feature, edge/topology, subgraph activation, and cross‑component co‑evolution. The authors further map dynamic‑graph learning subfields to agent capabilities, discuss potential failure modes, and outline graph‑aware evaluation and governance protocols to guide the design and oversight of self‑evolving agents.
arXiv:2608.21156v1 Announce Type: cross Abstract: LLMs have evolved from language generators to autonomous agents capable of complex, long-horizon tasks. This evolution has produced paradigms includi...
LLMs have evolved from language generators to autonomous agents capable of complex, long-horizon tasks. This evolution has produced paradigms including Prompt Engineering to elicit model capabilities,...
arXiv:2608. 10504v1 Announce Type: new Abstract: As coding agents increasingly handle implementation, the central challenge shifts from building individual agents to building an infrastructure that systematically improves them.
arXiv:2608. 04968v1 Announce Type: new Abstract: The capabilities of an LLM agent depend not only on its model but on the harness: the executable program that constructs context, invokes tools, verifies results, and recovers from failure.
arXiv:2604. 17708v2 Announce Type: replace Abstract: Automating operations research (OR) with large language models (LLMs) remains limited by hand-crafted reasoning--execution workflows.
arXiv:2606. 12191v1 Announce Type: cross Abstract: Environments serve as interactive systems for large language model (LLM) based agents across diverse scenarios and play a crucial role in driving the continual evolution of model capabilities.
arXiv:2607. 26598v1 Announce Type: cross Abstract: Large language model (LLM) agents may recover from a failure within an episode or after a retry, yet the same execution failure can recur in later tasks because post-episode feedback rarely revises the persistent harness that guides future interactions.
arXiv:2607. 26533v1 Announce Type: new Abstract: Graph Foundation Models (GFMs) aim to learn transferable knowledge from multi-domain graphs and adapt to unseen scenarios.
Graph Foundation Models (GFMs) aim to learn transferable knowledge from multi-domain graphs and adapt to unseen scenarios. As a fundamental source of relational semantics in graphs, the transferability of topological patterns has long been central to GFM research.
arXiv:2604. 08377v2 Announce Type: replace Abstract: Large language model (LLM) agents such as OpenClaw rely on reusable skills to perform complex tasks, yet these skills remain largely static after deployment.
arXiv:2608.29971v1 Announce Type: new Abstract: As agentic systems evolve into complex multi agent orchestration workflows, there is a growing and critical need for systematic frameworks that measure...
arXiv:2608. 02650v1 Announce Type: new Abstract: Large language model (LLM) agents increasingly rely on external tools to complete complex real-world tasks.