arXiv:2606. 21140v2 Announce Type: replace-cross Abstract: Rapid advances in large language models have improved the task-solving capabilities of command-line-interface (CLI)-based agents, whose CLIs determine how models invoke tools, maintain interaction history, and recover from failures.
CivBench is an open‑source benchmark that evaluates language‑model agents in the long‑horizon, tool‑mediated game Civilization VI using the Model Context Protocol (MCP). Each episode lasts over 300 turns, generating thousands of tool calls across a 76‑tool action space, and includes a narration layer that translates visual game state into structured text. The study characterises agent behaviour across four model families, introducing Proactive Monitoring Rate (PMR) and RAG@10 as interface‑level metrics, and finds that agents often under‑monitor strategic state and fail to execute near‑term commitments despite tool access and explicit guidance.
arXiv:2607. 10059v1 Announce Type: new Abstract: Agent systems based on large language models (LLMs) are increasingly deployed for autonomous tasks, yet existing evaluations mostly focus on task success rather than whether agents know when to abstain.
The paper surveys AI agents that operate primarily through command-line terminals, defining them as systems whose main action loop involves executing terminal commands, receiving textual feedback, and interacting with a stateful environment. It introduces a seven‑dimensional terminal competence profile to link system architecture, learning, and evaluation, and highlights how behavior is jointly shaped by the model, interface, harness, runtime, and environment. The authors argue for explicit reporting of system and runtime conditions, supported by replayable traces and process‑level evidence, to better understand and benchmark terminal‑mediated agency.
The study investigates how individual components of a coding harness—planning, action space, and context management—affect autonomous coding agents’ performance. By fixing the execution loop and varying these components across 176 settings on SWE‑Bench Verified and Terminal‑Bench 2.1, the authors find that context management is most valuable when context windows are tight, staging rule‑based elision before LLM summarization yields the best efficiency, planning serves as an accuracy scaffold for weaker models and a cost saver for stronger ones, and predefined tools help models with limited bash skills while bash‑capable models benefit from a bash‑only interface. Trajectory‑level analysis shows that context management lengthens execution paths, planning alters where trajectories terminate, and the action space determines code granularity, offering a modular framework for future harness design.
arXiv:2606. 24551v1 Announce Type: new Abstract: Computer-use agents can execute software tasks through either graphical interfaces or programmatic command interfaces, but existing evaluations confound interaction modality with differences in tasks, initial states, verifiers, and permitted actions.
arXiv:2609.13458v1 Announce Type: cross Abstract: Embodied language grounding requires more than identifying the referent of an instruction: recovered semantics must also control the action an agent...
AnyAct introduces a universal action layer that consolidates diverse tool capabilities into a self‑evolving action space for AI agents operating in open‑world environments. It tackles the scale dilemma, tool non‑stationarity, and heterogeneous feedback by using hierarchical progressive retrieval and test‑time reliability evolution, while a heterogeneous observation grounding module unifies multi‑modal feedback. Evaluations on LiveMCPBench and the newly created OSMCP benchmark show state‑of‑the‑art performance, with significant gains in task success rate and reduced execution steps, especially for models with limited native capabilities.
The paper introduces a framework for evaluating large‑language‑model agents by focusing on the interaction between the model and its execution harness. It defines five complementary metrics—Valid Completion Rate, Model‑Harness Fit, Cross‑Model Fit, Interaction Residuals, and Token Load—to assess how reliably agents finish tasks, how well harnesses support specific models, and the overall suitability across models. Using these metrics, the authors evaluate four general‑purpose harnesses and two native systems on six benchmarks, finding that harness rankings differ by model and task and that some harnesses consistently achieve high valid completion rates.
arXiv:2605. 10555v2 Announce Type: replace Abstract: As AI agents transition from research prototypes to enterprise production systems, the tool interfaces they consume remain rooted in human-oriented CRUD paradigms.
The paper investigates how the representation of tool interfaces—referred to as a tool schema—affects the performance of large language model agents. By systematically transforming a native tool schema with nine operators, the authors demonstrate that agents exhibit significant variability in success rates across different but functionally equivalent schema variants, a phenomenon they call schema bias. Experiments on eleven LLMs, including closed models, reveal that success can range from complete failure to 97% solely due to schema differences, and that training only mitigates bias when the specific variant is included in the training data.
FACET is a framework for synthesizing terminal tasks that preserves source intent and ensures cross‑artifact consistency. It reconstructs agent skills into coherent scenarios, repairs the execution environment, and uses the resulting container state as shared grounding for the instruction, solution, and verifier. By validating and repairing artifacts through execution, FACET produces complex tasks with dense executable checks and data‑efficient supervision, improving performance on Terminal‑Bench 2.1.