feat: 安徽运八需求全流水线输出同步 + 知识库/agents/资源文件更新

agents/design/case_designer.md

@@ -4,6 +4,7 @@ zone: design
 description: 将测试点转化为可执行测试用例，严格遵循模板表头，预期结果双验证
 tools: Read, Write, Glob, Bash
 depends_on: ["testpoint-designer", "data-builder"]
+ai_generative: true
 produces: ["output/test_cases/{{BASE_NAME}}_测试用例.md"]
 ---
 

agents/design/data_builder.md

@@ -4,6 +4,7 @@ zone: design
 description: 为测试用例构造精确的测试数据集，自动提取 + 人工补充
 tools: Read, Write, Glob
 depends_on: ["requirement-analyzer", "testpoint-designer"]
+ai_generative: true
 produces: ["output/analysis/{{BASE_NAME}}_测试数据.md"]
 ---
 

agents/design/testpoint_designer.md

@@ -4,6 +4,7 @@ zone: design
 description: 基于分析+风险+策略+历史经验，设计全面测试点矩阵，标注来源
 tools: Read, Write, Glob
 depends_on: ["test-strategist", "requirement-analyzer", "risk-assessor"]
+ai_generative: true
 produces: ["output/test_points/{{BASE_NAME}}_测试点.md"]
 ---
 

agents/review/case_reviewer.md

@@ -4,6 +4,7 @@ zone: review
 description: 按 review_checklist 逐项评审用例，错误分级（阻断/建议/优化），标注到具体行号
 tools: Read, Write, Glob
 depends_on: ["case-designer"]
+ai_generative: true
 produces: ["output/analysis/{{BASE_NAME}}_评审报告.md"]
 ---
 

agents/review/coverage_auditor.md

@@ -4,6 +4,7 @@ zone: review
 description: 需求→测试点→用例三级追溯覆盖审计，识别覆盖缺口和过度覆盖
 tools: Read, Write, Glob
 depends_on: ["case-reviewer"]
+ai_generative: true
 produces: ["output/analysis/{{BASE_NAME}}_覆盖率审计.md"]
 ---
 

agents/review/quality_gatekeeper.md

@@ -4,6 +4,7 @@ zone: review
 description: 综合评审+覆盖率进行三级裁决：PASS / PASS_WITH_FIX / BLOCKED
 tools: Read, Write
 depends_on: ["case-reviewer", "coverage-auditor"]
+ai_generative: true
 produces: ["output/analysis/{{BASE_NAME}}_质量裁决.md"]
 ---