OpenMindOpenMind 组织动力学实时模拟与决策引擎

MindOS Prompt 内部 API 预览

这页专门确认：后面 Prompt 技术说明接入时，我们不仅有结构页和示例页，还有一条内部 API 可以按 layer 返回示例载荷。

内部 Prompt 载荷 API

GET /api/assessment/prompt-payloads?layer=scoring

这条内部 API 只返回 Prompt 示例载荷，不返回题目权重、原始 score 或修正规则全文。

保密口径

这些示例只演示 Prompt 真正该吃到的结构化字段，不演示逐题权重、原始 score 或 B / D 模块内部规则全文。

先看 Prompt 服务端载荷再看 Prompt 示例载荷再看文档接入清单回 Prompt 总导航再看替换后验收清单再看报告 / API 验收

评分 / 汇总前置层 API 预览

先把答题结果整理成服务端稳定结构，再交给后续 Prompt 读取。

curl

curl "http://localhost:3000/api/assessment/prompt-payloads?layer=scoring"

允许示例字段

assessment_versionnormalized_answersdimension_scores_10resistance_scores_r1_r5function_group_scores_s1_s5four_forces_f1_f4rank_basequadrant_base

继续只留服务端

assessment_versionnormalized_answersdimension_scores_10resistance_scores_r1_r5function_group_scores_s1_s5four_forces_f1_f4rank_basequadrant_base

对外只出这些结果块

scoring_summaryrank_summarybase_quadrant_summary

Response

{
  "success": true,
  "data": {
    "source": "mindos_private_server_only",
    "principle": "这些示例只演示 Prompt 真正该吃到的结构化字段，不演示逐题权重、原始 score 或 B / D 模块内部规则全文。",
    "layer": "scoring",
    "title": "评分 / 汇总前置层",
    "description": "先把答题结果整理成服务端稳定结构，再交给后续 Prompt 读取。",
    "payload": {
      "assessment_version": "L3",
      "normalized_answers": [
        {
          "question_id": "L3_A_001",
          "option": "B"
        },
        {
          "question_id": "L3_B_012",
          "option": "A"
        }
      ],
      "dimension_scores_10": {
        "heart": 71,
        "cognition_mode": 63,
        "emotional_quotient": 68,
        "values": 74,
        "personality": 61,
        "motivation": 66,
        "orientation_force": 72,
        "awareness_force": 65,
        "connection_force": 58,
        "action_force": 69
      },
      "resistance_scores_r1_r5": {
        "r1": 58,
        "r2": 47,
        "r3": 61,
        "r4": 49,
        "r5": 42
      },
      "function_group_scores_s1_s5": {
        "s1": 70,
        "s2": 62,
        "s3": 59,
        "s4": 64,
        "s5": 73
      },
      "four_forces_f1_f4": {
        "f1": 69,
        "f2": 66,
        "f3": 57,
        "f4": 68
      },
      "rank_base": "D5",
      "quadrant_base": "高心智+高能量"
    },
    "contract": {
      "key": "scoring",
      "layer": "评分 / 汇总前置层",
      "allowedPayloadKeys": [
        "assessment_version",
        "normalized_answers",
        "dimension_scores_10",
        "resistance_scores_r1_r5",
        "function_group_scores_s1_s5",
        "four_forces_f1_f4",
        "rank_base",
        "quadrant_base"
      ],
      "privateFields": [
        "assessment_version",
        "normalized_answers",
        "dimension_scores_10",
        "resistance_scores_r1_r5",
        "function_group_scores_s1_s5",
        "four_forces_f1_f4",
        "rank_base",
        "quadrant_base"
      ],
      "publicOutputFields": [
        "scoring_summary",
        "rank_summary",
        "base_quadrant_summary"
      ],
      "guardrails": [
        "不把题目权重、score 映射和 B 模块原始修正规则直接喂给 Prompt。",
        "不允许 Prompt 逐题再算一次维度或功能组。"
      ]
    }
  }
}

解释层 Prompt API 预览

把位阶、画像、摩擦、风险等解释真正需要的修正层结果喂给 Prompt。

curl

curl "http://localhost:3000/api/assessment/prompt-payloads?layer=interpretation"

允许示例字段

assessment_versionrank_basequadrant_baserank_shift_summaryprofile_shift_summaryfriction_summaryrisk_summarymask_break_summaryb_module_revision_summaryd_module_consistency_summary

继续只留服务端

rank_shift_summaryprofile_shift_summaryfriction_summaryrisk_summarymask_break_summaryb_module_revision_summaryd_module_consistency_summary

对外只出这些结果块

rank_interpretationprofile_interpretationquadrant_interpretationfriction_interpretationrisk_interpretation

Response

{
  "success": true,
  "data": {
    "source": "mindos_private_server_only",
    "principle": "这些示例只演示 Prompt 真正该吃到的结构化字段，不演示逐题权重、原始 score 或 B / D 模块内部规则全文。",
    "layer": "interpretation",
    "title": "解释层 Prompt",
    "description": "把位阶、画像、摩擦、风险等解释真正需要的修正层结果喂给 Prompt。",
    "payload": {
      "assessment_version": "L3",
      "rank_base": "D5",
      "quadrant_base": "高心智+高能量",
      "rank_shift_summary": "B 模块显示风险防御偏强，表现位阶下修半级。",
      "profile_shift_summary": "画像从『推进型』修正为『谨慎推进型』。",
      "friction_summary": {
        "inner": "对犯错代价较敏感",
        "relation": "对协作中失控感有预警",
        "structure": "遇到模糊边界时容易先收缩"
      },
      "risk_summary": [
        "高标准导致犹豫",
        "对不确定性的容忍度偏低"
      ],
      "mask_break_summary": "D 模块显示公开表达比真实内在状态更保守。",
      "b_module_revision_summary": "风险防御和路径依赖阻抗偏高，需要在建议层做修正。",
      "d_module_consistency_summary": "自述与作答一致性中等，存在轻微防御性表达。"
    },
    "contract": {
      "key": "interpretation",
      "layer": "解释层 Prompt",
      "allowedPayloadKeys": [
        "assessment_version",
        "rank_base",
        "quadrant_base",
        "rank_shift_summary",
        "profile_shift_summary",
        "friction_summary",
        "risk_summary",
        "mask_break_summary",
        "b_module_revision_summary",
        "d_module_consistency_summary"
      ],
      "privateFields": [
        "rank_shift_summary",
        "profile_shift_summary",
        "friction_summary",
        "risk_summary",
        "mask_break_summary",
        "b_module_revision_summary",
        "d_module_consistency_summary"
      ],
      "publicOutputFields": [
        "rank_interpretation",
        "profile_interpretation",
        "quadrant_interpretation",
        "friction_interpretation",
        "risk_interpretation"
      ],
      "guardrails": [
        "解释层输出可以对外展示，但修正层内部判断细节继续只留在服务端。",
        "Prompt 只解释结构化结果，不暴露内部 if/then 和修正规则。"
      ]
    }
  }
}

报告生成 Prompt API 预览

把已经稳定的解释块和报告章节顺序交给报告层 Prompt 生成文案。

curl

curl "http://localhost:3000/api/assessment/prompt-payloads?layer=report"

允许示例字段

assessment_versionuser_contextreport_sectionsrank_interpretationprofile_interpretationfriction_interpretationproduct_recommendation_basismicro_intervention_basiscta_blocks

继续只留服务端

deep_report_section_orderproduct_recommendation_basismicro_intervention_basisreport_render_hints

对外只出这些结果块

report_sectionsproduct_recommendationsmicro_interventionscta_blocks

Response

{
  "success": true,
  "data": {
    "source": "mindos_private_server_only",
    "principle": "这些示例只演示 Prompt 真正该吃到的结构化字段，不演示逐题权重、原始 score 或 B / D 模块内部规则全文。",
    "layer": "report",
    "title": "报告生成 Prompt",
    "description": "把已经稳定的解释块和报告章节顺序交给报告层 Prompt 生成文案。",
    "payload": {
      "assessment_version": "L3",
      "user_context": {
        "age": 34,
        "role": "销售负责人",
        "industry": "企业服务"
      },
      "report_sections": [
        "用户概览",
        "四象限定位",
        "摩擦分析",
        "十二画像分析",
        "产品与微干预建议"
      ],
      "rank_interpretation": "当前位阶属于稳定推进阶段，但在高风险情境下会明显转保守。",
      "profile_interpretation": "画像更偏谨慎推进型，适合先确定边界再行动。",
      "friction_interpretation": "最大摩擦来自对犯错成本的预期过高。",
      "product_recommendation_basis": [
        "位阶 D5",
        "高心智+高能量",
        "风险防御阻抗偏高"
      ],
      "micro_intervention_basis": [
        "降低单次决策压力",
        "先做低风险试探动作",
        "强化协作场景下的边界表达"
      ],
      "cta_blocks": {
        "primary": "先去销帮做一条真实机会判断",
        "secondary": "回工作台看完整档案",
        "support": "如果仍不确定，再看产品建议和微干预"
      }
    },
    "contract": {
      "key": "report",
      "layer": "报告生成 Prompt",
      "allowedPayloadKeys": [
        "assessment_version",
        "user_context",
        "report_sections",
        "rank_interpretation",
        "profile_interpretation",
        "friction_interpretation",
        "product_recommendation_basis",
        "micro_intervention_basis",
        "cta_blocks"
      ],
      "privateFields": [
        "deep_report_section_order",
        "product_recommendation_basis",
        "micro_intervention_basis",
        "report_render_hints"
      ],
      "publicOutputFields": [
        "report_sections",
        "product_recommendations",
        "micro_interventions",
        "cta_blocks"
      ],
      "guardrails": [
        "报告生成层不重复解释底层公式，只消费已整理好的解释块。",
        "前台 / 开放平台最终只拿章节结果，不拿服务端内部依据字段。"
      ]
    }
  }
}