speccore 5.67.87 → 5.67.89
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
|
@@ -19,6 +19,12 @@ disable-model-invocation: false
|
|
|
19
19
|
- 先列出发现的问题（如"已有旧分析结果"、"路径配置为空"）
|
|
20
20
|
- 给用户清晰的选择：继续 / 覆盖 / 跳过 / 停止
|
|
21
21
|
- **由用户决定**，不是你替用户决定
|
|
22
|
+
5. **analyze 必须走完整 prompt→填充→apply 流程**: 分析需求文档时，禁止直接跳到代码修复或跳过文档生成。正确流程:
|
|
23
|
+
- 步骤1: `speccore analyze --prompt -I <iter> --type <type>` → 获取 SPECCORE_PROMPT
|
|
24
|
+
- 步骤2: Read 各端 PRD 文档和源码路径，理解需求
|
|
25
|
+
- 步骤3: 对每个文档（ANALYSIS/TECH/TEST/REVIEW/RISK/DEPS/MONITOR）**填入实质内容**，禁止写"待填充"或复制模板空壳
|
|
26
|
+
- 步骤4: `speccore analyze --apply '{"ANALYSIS.md":"...",...}' -I <iter>` 写入
|
|
27
|
+
- ❌ 禁止行为: 跳过文档填充直接写代码、生成 137 行重复空壳、用 "API=0 接口" 为由跳过分析
|
|
22
28
|
`init → doc2spec → analyze → split → plan → execute → pr → done → spec2doc`
|
|
23
29
|
示例: 用户说"执行完再分析" → 你应该排列为 `analyze → execute`，并在计划中说明"根据工作流逻辑，先分析再执行"
|
|
24
30
|
- **全自动**: "全自动执行/一键完成" → 所有步骤不等确认，全部自动跑
|