ClawProBench: Trace-Aware Evaluation of AI Agents with Runtime Coverage and Frozen Workplace-Style Holdouts

Hugging Face Daily Papers Papers

Summary

ClawProBench introduces a trace-aware benchmark for evaluating AI agents, emphasizing runtime execution and safety to reveal limitations in final-answer rankings.

Agent benchmarks often evaluate only final answers even when agents run on stateful runtimes. We argue this under-specifies what is being evaluated: the proper unit is a declared model-plus-runtime configuration whose failures can occur in evidence acquisition, runtime routing, safety boundaries, or repeated execution. We present ClawProBench, a trace-aware benchmark for runtime-native agent evaluation instantiated on OpenClaw, a live agent runtime with workspace tools and native surfaces for browsing, memory, messaging, scheduling, skills, and subagents. ClawProBench defines two tracks: a 102-scenario full profile with live workspace and native-runtime routing tasks, and a frozen 68-scenario holdout with closed-world JSON output contracts for robust ranking. Trials are scored from execution traces via a safety-gated formula combining correctness, process quality, and efficiency, preserving failure evidence for audit. Our anonymous artifact includes benchmark definitions, scoring code, manifests and sanitized traces. We evaluate 68 configurations on the full profile and 37 on holdout. The top safety-gated average trace score is 0.7671. Native-runtime tasks underperform workspace-live tasks (0.5238 vs. 0.6415). On holdout, pass@k-any outperforms strict three-trial pass (0.6638 vs. 0.2890), while full-profile and holdout rankings show weak alignment (Spearman 0.1300). Rankings based purely on correctness differ substantially from process-aware, safety-gated and strict-pass views. Final-answer leaderboards may hide native-surface weaknesses, one-off successes and trace-local agent failure modes.
Original Article
View Cached Full Text

Cached at: 08/25/26, 08:38 PM

Paper page - ClawProBench: Trace-Aware Evaluation of AI Agents with Runtime Coverage and Frozen Workplace-Style Holdouts

Source: https://huggingface.co/papers/2608.22510

Abstract

ClawProBench evaluates agent configurations via execution traces across live and frozen tracks, revealing that final-answer rankings obscure native-runtime failures and process-quality differences.

Agent benchmarksoften evaluate only final answers even when agents run on stateful runtimes. We argue this under-specifies what is being evaluated: the proper unit is a declared model-plus-runtime configuration whose failures can occur in evidence acquisition, runtime routing, safety boundaries, or repeated execution. We present ClawProBench, a trace-aware benchmark for runtime-native agent evaluation instantiated onOpenClaw, a live agent runtime withworkspace toolsandnative surfacesfor browsing, memory, messaging, scheduling, skills, and subagents. ClawProBench defines two tracks: a 102-scenario full profile with live workspace and native-runtime routing tasks, and a frozen 68-scenario holdout with closed-world JSON output contracts for robust ranking. Trials are scored fromexecution tracesvia a safety-gated formula combining correctness,process quality, and efficiency, preserving failure evidence for audit. Our anonymous artifact includes benchmark definitions, scoring code, manifests and sanitized traces. We evaluate 68 configurations on the full profile and 37 on holdout. The top safety-gated average trace score is 0.7671. Native-runtime tasks underperform workspace-live tasks (0.5238 vs. 0.6415). On holdout,pass@k-anyoutperforms strict three-trial pass (0.6638 vs. 0.2890), while full-profile and holdout rankings show weak alignment (Spearman 0.1300). Rankings based purely on correctness differ substantially from process-aware, safety-gated and strict-pass views. Final-answer leaderboards may hide native-surface weaknesses, one-off successes and trace-local agent failure modes.

View arXiv pageView PDFProject pageGitHub822Add to collection

Get this paper in your agent:

hf papers read 2608\.22510

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2608.22510 in a model README.md to link it from this page.

Datasets citing this paper1

#### xyh110sym/clawprobench Updatedabout 11 hours ago • 2

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.22510 in a Space README.md to link it from this page.

Collections including this paper1

Similar Articles

WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation

Hugging Face Daily Papers

WildClawBench evaluates language and vision-language models on realistic long-horizon tasks using actual CLI environments with real tools. The benchmark reveals that even the best model achieves only 62.2% accuracy, indicating long-horizon agent evaluation remains challenging.