Ignorance or Incompetence? Constructing Knowledge-Gated, Verifiable Tasks for LLM Agents
2026-08-31 • Artificial Intelligence
Artificial IntelligenceComputation and Language
AI summaryⓘ
The authors created a new way to build tasks for AI agents that clearly separate the instructions from any special knowledge or rules needed to complete the task. They tested different setups and found that some tasks could only be solved when the agent had access to the private knowledge, proving the importance of this information. Their method also allows checking exactly when and how the private knowledge is used. The authors shared some of their tasks and tools publicly to support further research.
knowledge-gated taskprivate conventionstask instructionbenchmarkingprovenanceoraclecalibration tasksdeterministic solverrule corporaknowledge leakage
Authors
Hanlin Tian, Minhao Li, Yu Mi, Sihan Zhu, Zhao Yang, Yuxiang Wang, Hongquan Zhu, Qiufei Hu
Abstract
Professional agent tasks often depend on conventions that are absent from public corpora, yet benchmarks rarely control whether an agent has access to those conventions. We introduce a knowledge-gated task-construction protocol that separates a task instruction from a compact artefact containing private conventions, reference tables, and utility operators. Construction-time provenance, byte-identical task instructions across the provided- and withheld-artefact conditions, leak audits, and executable witnesses make dependence on the artefact explicit and testable. Across fifteen calibration tasks, one frontier agent configuration achieves a 68.0% pass rate with the artefact and 0% without it; on one task, a plausible but incorrect artefact also yields 0% across five trials. Deterministic solvers and rule corpora provide exact ground truth for structured tasks, while named criterion-level rubrics support outputs that cannot be checked by a single executable oracle. A configuration-relative calibration screen retains seven tasks satisfying our five-trial empirical knowledge-gating screen. These experiments validate the behavior of the construction protocol; they do not establish that the retained tasks improve post-training. We publicly release part of the task suite and supporting tooling at https://github.com/DatagridsAI/Knowledge-Gated-Task-Construction.