Skill: long-document schema and variable grounding

[Case 2]

• Initial Environment: A government-finance survey directory contains raw data files plus official documentation that defines revenue categories, units, and reporting structure. The analyst must read documentation before computing any total.
• Real Question: What was the total state government revenue in 2021?
• Why this demonstrates the capability: The question sounds simple, but the answer depends on identifying the correct revenue field, unit convention, and population scope from documentation. Without those definitions, a model can easily sum the wrong fields or mix derived and raw variables. The case therefore demonstrates variable grounding through long-form documentation.

Pipeline Execution Instructions

To synthesize data for this capability, you must strictly follow a 3-phase pipeline. Do not hallucinate steps. Read the corresponding reference file for each phase sequentially:

1. Phase 1: Environment Exploration Read the exploration guidelines to discover raw knowledge seeds: references/EXPLORATION.md

2. Phase 2: Trajectory Selection Once Phase 1 is complete, read the selection criteria to evaluate the trajectory: references/SELECTION.md

3. Phase 3: Data Synthesis Once a trajectory passes Phase 2, read the synthesis instructions to generate the final data: references/SYNTHESIS.md