Skill: bias and fairness diagnosis

[Case 2]

• Initial Environment: The environment contains structured healthcare-related data and a method library that supports both high-level and metric-specific bias questions. The agent can generate both numerical results and natural-language summaries.
• Real Question: Does insurance type influence the allocation of medical resources?
• Why this demonstrates the capability: The question is phrased at a general level, but answering it rigorously requires turning the abstract fairness concern into measurable comparisons or causal-effect style analyses. This makes the capability about diagnosing bias from structured data with suitable metrics and clear interpretation. It is therefore a strong example of bias and fairness diagnosis.

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