Large language models are deployed as long-horizon agents that compose diverse tools across MCP services into executable workflows. Existing MCP environment-scaling methods are data-centric: they first synthesize a database and scenario, then derive tools and tasks from them. As a result, tool interfaces may not match real ones, tool compositions are not grounded in execution, and tasks rarely span more than one environment. We propose ToolSpace, a tool-centric framework that reverses this order. ToolSpace mines data, state, and retry dependencies from adjacent calls in execution traces of real MCP services and open agent corpora, using deterministic rules rather than LLM judgment, and clusters the resulting weighted tool-composition graph into environments. Databases are synthesized only for state-mutating environments, while query and stateless tools replay recorded returns. ToolSpace then performs state-aware random walks with rollback and decompiles each chain into a natural-language task with assertion-based checks, keeping only samples that replay successfully from the initial state. Applied to AutomationBench and tau3-Bench Banking, the pipeline produces harder Pro versions on which 7 frontier models drop by up to 38.2 points. Across seven long-horizon tool-use benchmarks, ToolSpace raises Qwen3.6-35B-A3B from 23.0 to 41.0 average score (43.2 with RL) and GLM-5.2 from 44.6 to 56.6, on par with Claude Opus 5. Ablations show that execution-witnessed edges clearly outperform LLM-judged ones, and performance keeps improving with data and environment scale.
Figures & tables
Figure 1: Overview of ToolSpace. (a) Design Philosophy: a tool space layer between agents and MCP environments. (b) The tool-composition graph with a state-aware walk. (c) Generalization across 7 benchmarks. (d) Performance on two Pro benchmarks.
Figure 2: Overview of the ToolSpace framework. Stage 1 builds the Tool Space from real traces; Stage 2 synthesizes cross-environment tasks by State-Aware Random Walks, retaining only samples that pass replay verification.
Figure 3: The Interactive demo of state-aware random walking. An example walk from the interactive demo: 28 tool calls across 6 environments, completed after 2 rollbacks and kept after replay verification. Candidate successors and their sampling probabilities P(v′∣v) are shown on the right. Demo: https://toolspace-agent.github.io/#demo
Figure 4: ToolSpace graph statistics. (a) Environment-name word cloud. (b) Environment distribution across MCP domains. (c) Number of tools per environment. (d) Proportions of data, state, and retry dependencies. (e) Tool distribution by operational category. (f) Environment distribution by operational category.
Figure 5: Synthesized task and trajectory statistics. Distributions across 21,385 trajectories with at least five tool calls: (a) trajectory token length, (b) tool calls per trajectory, and (c) environments per trajectory.
Figure 6: Average tool calls of frontier models on (a) AutomationBench and (b) τ3 -Bench Banking (Standard vs. Pro).
Table 1: Performance on seven tool-use benchmarks. Bold/underline mark the best/second-best score per column within each same-scale pool (the 8B, 27–35B, and 744B pools, each including the ToolSpace model(s) at that scale). “–” indicates results not officially reported.
Frontier model
AutomationBench
τ3 -Bench Banking
Std.
Pro
Δ↓
Std.
Pro
Δ↓
Claude Opus 5.5
59.2
26.3
−32.9
51.5
35.7
−15.8
GPT 6.1 Sol
50.8
24.7
−26.1
46.4
29.6
−16.8
DeepSeek-V4.1-Flash-1017
60.5
22.3
−38.2
51.3
36.9
−14.4
DeepSeek-V4.1-Flash
54.8
20.0
−34.8
50.3
35.7
−14.6
GLM 5.3
52.5
22.7
−29.8
40.2
20.4
−19.8
Table 2: Results on AutomationBench Pro and τ3 -Bench Banking Pro. Std. denotes the standard split; Pro denotes our upgraded split. Δ is Pro − Standard. Std. scores are taken from the Atria Dawn public leaderboard ( https://atria-asi.ai/#benchmark-results ). The scores of DeepSeek-V4.1-Flash follow their official release notes.
Figure 7: Data and environment scaling results. (a) Data scaling by training-data fraction. (b) Environment scaling by training-environment fraction. Zero denotes the untuned Qwen3.6-35B-A3B backbone.
Figure 9: Environment relevance and synthesis cost. (a) Effect of environment relevance: Full, Top 50%, and Random 50% environments; Δ is the Top–Random gap. (b) Environment synthesis cost proxy: synthesized files by type over 500 sampled environments.
Variant
Automation Bench
Skills Bench
BFCL V4
τ3 -Bench Banking
OfficeQA Pro
Full ToolSpace
45.8
42.4
74.3
34.0
48.9
w/o state edges
36.5
38.0
70.1
21.5
44.0
w/o data edges
38.2
37.5
62.4
29.8
40.6
w/o retry edges
43.9
41.2
72.8
32.6
47.9
w/o rollback
41.0
39.8
71.5
30.1
46.3
Uniform walk (no w )
39.6
38.9
69.8
28.4
45.1
Table 3: Component ablation. Darker shading marks a larger within-column drop from Full.
Tool-augmented LLM agents commonly rely on step-wise atomic tool calls, where each invocation, observation, and value transfer is exposed in the main reasoning trace. This creates an execution-granularity mismatch: locally deterministic tool workflows are unfolded into repeated model-visible decisions, consuming context and forcing the model to manage low-level dataflow in the trace. We introduce HyperTool, a unified executable MCP-style tool interface that changes the model-visible unit of tool execution. A model invokes HyperTool with a code block that can call existing tools through their original schemas, manipulate returned values, and pass intermediate results locally, folding deterministic tool subroutines into a single outer call. To train models to use this interface, we synthesize HyperTool-format trajectories from cross-tool compositional tasks and verify them in real MCP environments. On MCP-Universe, HyperTool improves average accuracy from 15.69% to 35.29% on Qwen3-32B and from 9.93% to 33.33% on Qwen3-8B, and surpass GPT-OSS and Kimi-k2.5 on average accuracy, showing that our HyperTool can substantially improve multi-step tool use.
Yaxin Du, Yifan Zhou, Yujie Ge +7
Shanghai Jiao Tong University · IQuest Research · Beijing University of Aeronautics and Astronautics
Equipping LLMs with tool-use capabilities via Agentic Reinforcement Learning (Agentic RL) is bottlenecked by two challenges: the lack of scalable, robust execution environments and the scarcity of realistic training data that captures implicit human reasoning. Existing approaches depend on costly real-world APIs, hallucination-prone LLM simulators, or synthetic environments that are often single-turn or depend on pre-collected documents. Moreover, synthetic trajectories are frequently over-specified, resembling instruction sequences rather than natural human intents, reducing their effectiveness for RL training. We introduce EnvFactory, a fully automated framework that addresses both challenges. EnvFactory autonomously explores and verifies stateful, executable tool environments from authentic resources, and synthesizes natural multi-turn trajectories through topology-aware sampling and calibrated refinement, producing grounded queries with implicit intents. Using only 85 verified environments across 7 domains, EnvFactory generates 2,575 SFT and RL trajectories. Despite using significantly fewer environments than prior work, which are often 5 times more, EnvFactory achieves superior training efficiency and downstream performance, improving Qwen3-series models by up to +15% on BFCLv3, +8.6% on MCP-Atlas, and +6% on conversational benchmarks including τ2-Bench and VitaBench. By fully automating both environment construction and trajectory synthesis, EnvFactory provides a scalable, extensible, and robust foundation for Agentic RL.
Minrui Xu, Zilin Wang, Mengyi DENG +12
LARK, HKUST (GZ) · University of Cambridge · UCL +2
We explore how to build performant large language model (LLM) tool-use agents that complete complex, multi-step tasks. Specifically, we study how data composition in synthetically generated tool-use environments affects the training of such agents. To do so, we propose the SynthAPI pipeline, which explicitly controls the diversity of generated environments along the axes of tools, tasks, and the databases on which the tools operate. By systematically varying SynthAPI's generation controls in 458,752 tasks across 849 trained models of four different sizes, we derive a scaling law that predicts downstream tool-use agent generalization based on the scale and composition of the synthesized environments. These predictable relationships reveal several lessons for scaling tool-use agent data. As data budgets grow, scaling tool and database diversity becomes more effective than increasing the number of tasks, and increasing data complexity yields larger gains. Scaling according to these findings yields agents that outperform models trained on data from prior work in large-scale tool-use agent and synthetic environment generation across four evaluation benchmarks.