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

1.5 KB
Raw Blame
{
  "skill": "archetypes",
  "primary_value_surface": "paired reasoning quality",
  "primary_metric": "pairwise_win_rate_vs_placebo",
  "secondary_metrics": ["judge_agreement", "format_length_controlled_win_rate", "archetype_match_accuracy", "intervention_appropriateness"],
  "distractors": {
    "off_target_set": "Novel problems with no historical precedent; one-off incidents without recurring structure; systems too simple to exhibit archetypal behavior; meta-questions about system archetypes",
    "description": "Items where the recurring pattern library doesn't apply — genuinely novel dynamics or trivially simple systems"
  },
  "calibration_range": [0.40, 0.70],
  "calibration_exception": null,
  "min_sample_primary": 80,
  "min_sample_replication": 80,
  "planned_comparison": "trigger",
  "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 AND format_length_control_pass",
    "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 format_length_bias_detected",
    "trigger_only": "trigger_card_delta_pp >= full_guide_delta_pp AND trigger_paired_test_p < 0.05",
    "ceiling_needs_harder_data": "placebo_baseline > 0.70"
  },
  "eval_family": "Systems/product/strategy pairwise"
}