Skip to content

28 eval-informed mental models and critical-thinking skills for Claude Code, GitHub Copilot, Codex, Cursor, and other Agent Skills-compatible tools

1.4 KB
Raw Blame
{
  "skill": "second-order",
  "primary_value_surface": "objective correctness",
  "primary_metric": "consequence_prediction_accuracy",
  "secondary_metrics": ["cascade_depth_coverage", "unintended_effect_identification", "time_horizon_calibration", "intervention_simulation_fidelity"],
  "distractors": {
    "off_target_set": "Direct cause-effect relationships with no downstream effects; immediate decisions with no ripple effects; meta-questions about second-order thinking",
    "description": "Items where consequences are direct and immediate with no cascading effects to analyze"
  },
  "calibration_range": [0.40, 0.70],
  "calibration_exception": null,
  "min_sample_primary": 60,
  "min_sample_replication": 60,
  "planned_comparison": "placebo",
  "decision_gate": {
    "elevate": "primary_delta_pp >= 5 AND paired_test_p < 0.05 AND replication_delta_pp >= 5 AND replication_paired_test_p < 0.05 AND replication_direction_matches_primary",
    "directional_not_replicated": "primary_delta_pp >= 5 AND paired_test_p < 0.05 AND (replication_delta_pp < 5 OR replication_paired_test_p >= 0.05 OR replication_direction_opposite)",
    "no_lift": "primary_delta_pp < 5 OR paired_test_p >= 0.05",
    "quarantine_redirect": "primary_delta_pp < 0 OR cascade_depth_coverage < 0.50",
    "ceiling_needs_harder_data": "placebo_baseline > 0.70"
  },
  "eval_family": "Debugging/fault-localization"
}