feat: add generated scene skill platform hardening
This commit is contained in:
木炎
@@ -0,0 +1,66 @@
|
||||
# 102 Scene Ledger Alignment Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round does not attempt full `102`-scene migration.
|
||||
|
||||
It freezes the workbook snapshot `docs/2026-04-18-102-scenes-validation-overview.xlsx` into repo-local assets so the current code-backed family baseline can be compared against a concrete ledger snapshot.
|
||||
|
||||
## Delivered
|
||||
|
||||
Added:
|
||||
|
||||
1. `tests/fixtures/generated_scene/scene_ledger_snapshot_2026-04-18.json`
|
||||
2. `tests/scene_ledger_snapshot_test.rs`
|
||||
|
||||
## Snapshot Facts
|
||||
|
||||
From the workbook snapshot generated at `2026-04-18 16:48:05`:
|
||||
|
||||
1. total scenes: `102`
|
||||
2. validated scenes: `9`
|
||||
3. not yet validated scenes: `93`
|
||||
4. passed real sample: `1`
|
||||
5. failed real samples: `3`
|
||||
6. fail-closed scenes: `2`
|
||||
7. reassigned boundary scenes: `3`
|
||||
|
||||
## Alignment With Current Family Assets
|
||||
|
||||
The frozen ledger snapshot now explicitly aligns with the repo-local family baseline on these points:
|
||||
|
||||
1. `G1-E`
|
||||
- `高低压新增报装容量月度统计表` is the only passed real sample in the snapshot
|
||||
2. `G2`
|
||||
- three validated line-loss samples remain recorded as first-round failed runs
|
||||
3. `G6`
|
||||
- `电能表现场检验完成率指标报表` is recorded as a reassigned boundary scene
|
||||
4. `G7`
|
||||
- `计量资产库存统计` is recorded as a reassigned boundary scene
|
||||
5. `G8`
|
||||
- `95598供电服务月报` is recorded as a reassigned boundary scene
|
||||
|
||||
## Interpretation
|
||||
|
||||
The workbook remains a mid-run ledger snapshot, not the final latest state of same-day markdown execution.
|
||||
|
||||
That distinction is now preserved in repo assets:
|
||||
|
||||
1. ledger snapshot = frozen workbook state
|
||||
2. family assets = current code-backed baseline
|
||||
3. markdown reports = later execution progress on the same day
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
- `cargo test --test scene_ledger_snapshot_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
Track D / Track E now have a repo-local bridge between:
|
||||
|
||||
1. family-level code-backed baselines
|
||||
2. the original `102`-scene workbook snapshot
|
||||
|
||||
This gives the next expansion round a stable ledger anchor without forcing full Excel-driven workflow into the compiler path.
|
||||
@@ -0,0 +1,120 @@
|
||||
# 第一轮真实样本迁移与验证执行表
|
||||
|
||||
日期：2026-04-18
|
||||
|
||||
上游设计：`docs/superpowers/specs/2026-04-17-scene-skill-60-to-90-roadmap-design.md`
|
||||
|
||||
上游计划：`docs/superpowers/plans/2026-04-17-scene-skill-60-to-90-roadmap-plan.md`
|
||||
|
||||
## 1. 文档目的
|
||||
|
||||
本表用于承接 `60 -> 90` 路线图进入真实样本阶段后的首轮落地动作。
|
||||
|
||||
这一轮不直接要求业务人员先去内网验证，而是先完成以下顺序：
|
||||
|
||||
1. 从真实场景中收敛首轮样本
|
||||
2. 按当前 scene -> skill 转化链生成首轮 skill
|
||||
3. 将生成结果与 `tq` 级基线或当前 canonical/family 基线对比
|
||||
4. 只把通过结构、语义、readiness 初筛的候选物，送入后续内网验证
|
||||
|
||||
## 2. 选样原则
|
||||
|
||||
首轮样本选择遵循以下边界：
|
||||
|
||||
1. 严格服从 plan 的主线范围，只覆盖 `G1 + G2 + G3`
|
||||
2. 优先选择业务结构清晰、家族代表性强、后续容易形成复用模板的场景
|
||||
3. `G2` 优先级最高，因为它直接决定是否能逼近 `tq-lineloss-report` 级别的业务语义恢复
|
||||
4. `G1` 首轮优先选择通用报表类，不优先选择强上传型、强附件型、强人工补录型
|
||||
5. `G3` 首轮优先选择分页明细补数链清晰的工单类
|
||||
6. 对发现“标题与真实脚本结构不一致”或“页面壳与业务链条错位”的场景，先降级为观察样本，不进入首轮主样本
|
||||
|
||||
## 3. 当前收敛结果
|
||||
|
||||
### 3.1 主样本家族
|
||||
|
||||
- `G2` 多模式报表家族：先打透
|
||||
- `G1` 通用单页报表家族：先选干净样本验证可复制性
|
||||
- `G3` 分页明细补数家族：验证复杂 workflow 拆解与 fail-closed
|
||||
|
||||
### 3.2 暂不纳入首轮主样本的发现
|
||||
|
||||
- `高低压新增报装容量月度统计表`
|
||||
当前页面标题、工作信息描述、脚本主体出现明显不一致，疑似存在页面壳复用或内容错配，不适合作为首轮基准样本。
|
||||
|
||||
## 4. 第一轮执行样本表
|
||||
|
||||
| 序号 | 场景名称 | 家族 | 首轮角色 | 初判 archetype | 选入原因 | 主要风险 | 对标基线 | 当前状态 |
|
||||
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
|
||||
| 1 | 台区线损大数据-月_周累计线损率统计分析 | G2 | P0 主锚点 | `multi_mode_request.month_week_table` | 已知原始场景，可直接对标 `tq-lineloss-report`，也是当前 60->90 提升的核心样板 | mode matrix、动态日期字段、localhost 与宿主桥接区分 | `tq-lineloss-report` + P0 canonical | 待生成 |
|
||||
| 2 | 白银线损周报 | G2 | 家族扩展样本 | `multi_mode_request` 或同家族线损周报变体 | 与线损家族强相关，能验证 `tq` 样板是否可迁移到同家族其他周报 | 双系统登录态、周报聚合逻辑、内网 token 依赖 | `tq` 家族语义基线 + P1 family | 待生成 |
|
||||
| 3 | 线损同期差异报表 | G2 | 复杂扩展样本 | `multi_mode_request` / `secondary_request` 复合链 | 能暴露“多系统对比、主请求 + 二次补链”的真实缺口，是 G2 向 90 分跨越的关键压力样本 | secondary request、跨系统上下文、workflow 拆解 | `tq` 家族语义基线 + P1 family | 待生成 |
|
||||
| 4 | 售电收入日统计 | G1 | 主样本 | `single_request_table` 倾向，但可能带明细遍历 | 属于典型报表生成场景，页面参数较集中，适合作为 `G1` 首轮切入 | 可能存在按用户逐条查询与汇总，不一定是纯单请求 | P0/P1 canonical/family | 待生成 |
|
||||
| 5 | 供电所线路电量统计 | G1 | 补充样本 | `single_request_table` 到轻量 workflow 之间 | 具备日期范围、关键词、结果表格，能验证单页查询统计类的可抽象程度 | 可能依赖跨页执行与页面内表格汇总，而非纯 API 报表导出 | P0/P1 canonical/family | 待生成 |
|
||||
| 6 | 用户日电量监测 | G1 | 观察补样 | `single_request_table` 外围变体 | 可覆盖上传/处理/导出型报表边界，用于验证 G1 主模板的外延边界 | 上传驱动、处理链较重，不适合作为 G1 唯一主样本 | P0/P1 canonical/family | 待生成 |
|
||||
| 7 | 95598工单明细表 | G3 | 主样本 | `paginated_enrichment` | 典型分页工单明细场景，存在主查询、分页、补数、导出链路，适合作为 G3 首个标准样本 | 分页补数、日区间滚动、SQL/去重/导出链 | P0-3 canonical + family 基线 | 待生成 |
|
||||
| 8 | 95598、12398、流程超期风险工单明细 | G3 | 扩展样本 | `paginated_enrichment` 复合变体 | 同时覆盖多来源工单与超期风险链，能验证 G3 是否只会处理单一分页表 | 多入口 workflow、复合条件、主链与补链拆分 | P0-3 canonical + family 基线 | 待生成 |
|
||||
|
||||
## 5. 首轮执行顺序
|
||||
|
||||
首轮执行顺序固定如下，不跨序扩散：
|
||||
|
||||
1. 先生成 `台区线损大数据-月_周累计线损率统计分析`
|
||||
2. 再生成 `白银线损周报`
|
||||
3. 再生成 `线损同期差异报表`
|
||||
4. 再进入 `G1` 的 `售电收入日统计`、`供电所线路电量统计`
|
||||
5. `用户日电量监测` 仅作为 `G1` 边界补样，不先于前两个 `G1` 样本
|
||||
6. 最后进入 `G3` 的两个工单类样本
|
||||
|
||||
这样安排的原因是：
|
||||
|
||||
1. `G2` 是否打透，决定当前链路是否真正具备 `tq` 级语义恢复能力
|
||||
2. `G1` 更适合在 `G2` 形成稳定抽取口径后做规模化复制验证
|
||||
3. `G3` workflow 更复杂，应该建立在前两组的证据层和契约层已经稳定之后
|
||||
|
||||
## 6. 每个样本生成后的必做检查
|
||||
|
||||
每个样本生成完后，必须按同一口径检查，不允许直接跳内网：
|
||||
|
||||
1. `sceneId / skillId` 是否稳定，不出现退化命名
|
||||
2. `bootstrap domain` 是否落到真实业务系统，而不是误抽到 `localhost`
|
||||
3. `workflow archetype` 是否正确，不把复杂链条误压成简单单请求
|
||||
4. `request contract` 是否闭合，关键字段是否齐全
|
||||
5. `response path / column defs / normalize rules` 是否闭合
|
||||
6. `localhost:*` 是否被识别为宿主桥接依赖，而不是业务主链
|
||||
7. readiness 是否可解释，失败时是否明确 fail-closed
|
||||
|
||||
## 7. 每类样本的通过标准
|
||||
|
||||
### 7.1 G2 通过标准
|
||||
|
||||
- 至少能恢复正确的 mode matrix
|
||||
- 能区分主业务接口与宿主桥接依赖
|
||||
- 生成结果在关键语义上达到 `tq-lineloss-report` 同等级参考水平
|
||||
|
||||
### 7.2 G1 通过标准
|
||||
|
||||
- 能稳定抽出查询参数、目标接口、响应路径、列定义、导出产物
|
||||
- 同家族两个以上样本可以共用同一 archetype 与 gate 口径
|
||||
|
||||
### 7.3 G3 通过标准
|
||||
|
||||
- 能识别分页主链、补数链、导出链
|
||||
- 证据不足时稳定阻断，不伪装为可运行 skill
|
||||
|
||||
## 8. 本轮产出物要求
|
||||
|
||||
本轮真实样本阶段至少产出以下文档或结果：
|
||||
|
||||
1. 本执行表
|
||||
2. 首轮样本生成结果
|
||||
3. 样本对标比对记录
|
||||
4. 候选内网验证名单
|
||||
5. 不通过样本的 blocker 归因
|
||||
|
||||
## 9. 下一步
|
||||
|
||||
按本表顺序进入下一步：
|
||||
|
||||
1. 使用当前转化链生成上述首轮样本 skill
|
||||
2. 将生成结果与 `tq/canonical/family` 基线逐项对比
|
||||
3. 输出首轮迁移与候选验证报告
|
||||
@@ -0,0 +1,163 @@
|
||||
# 第一轮迁移与候选验证报告
|
||||
|
||||
日期：2026-04-18
|
||||
|
||||
上游执行表：[2026-04-18-first-real-scene-migration-execution-sheet.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/reports/2026-04-18-first-real-scene-migration-execution-sheet.md)
|
||||
|
||||
相关分析：
|
||||
|
||||
1. [2026-04-18-r1-real-tq-lineloss-analysis.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/reports/2026-04-18-r1-real-tq-lineloss-analysis.md)
|
||||
2. [2026-04-18-g2-first-round-blocker-summary.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/reports/2026-04-18-g2-first-round-blocker-summary.md)
|
||||
|
||||
## 1. 报告目的
|
||||
|
||||
本报告只回答四个问题：
|
||||
|
||||
1. 第一轮真实样本实际迁移了哪些场景
|
||||
2. 这些场景生成结果是否达到候选验证门槛
|
||||
3. 哪些样本可以进入后续内网验证
|
||||
4. 哪些 blocker 已经稳定，可以作为后续整改输入
|
||||
|
||||
## 2. 第一轮实际执行范围
|
||||
|
||||
按执行表顺序，本轮实际完成了 `G2` 家族首轮三样本：
|
||||
|
||||
1. `台区线损大数据-月_周累计线损率统计分析`
|
||||
2. `白银线损周报`
|
||||
3. `线损同期差异报表`
|
||||
|
||||
对应生成产物位于：
|
||||
|
||||
1. `examples/real_scene_batch_round1/skills/real-tq-lineloss-report-r1`
|
||||
2. `examples/real_scene_batch_round1/skills/real-baiyin-lineloss-weekly-r1`
|
||||
3. `examples/real_scene_batch_round1/skills/real-lineloss-period-diff-r1`
|
||||
|
||||
本轮尚未进入：
|
||||
|
||||
1. `G1` 单页报表样本
|
||||
2. `G3` 分页明细补数样本
|
||||
|
||||
## 3. 候选验证判定口径
|
||||
|
||||
样本只有同时满足以下条件，才可以进入后续内网候选验证：
|
||||
|
||||
1. archetype 正确
|
||||
2. bootstrap 落到真实主业务承载面
|
||||
3. modes 或 workflow 主链恢复正确
|
||||
4. request contract 基本闭合
|
||||
5. response path / column defs / normalize rules 基本闭合
|
||||
6. readiness 与真实可运行性不明显冲突
|
||||
|
||||
任一核心项失败，当前轮次即判定为：
|
||||
|
||||
- `不进入候选验证`
|
||||
|
||||
## 4. 第一轮样本结论表
|
||||
|
||||
| 场景 | 家族 | 生成是否成功 | 语义是否通过 | 是否进入候选验证 | 结论 |
|
||||
| --- | --- | --- | --- | --- | --- |
|
||||
| 台区线损大数据-月_周累计线损率统计分析 | G2 | 是 | 否 | 否 | 生成成功，但业务语义恢复失败 |
|
||||
| 白银线损周报 | G2 | 是 | 否 | 否 | 生成成功，但沿同类错误路径坍缩 |
|
||||
| 线损同期差异报表 | G2 | 是 | 否 | 否 | 生成成功，但复杂链路没有被正确表达 |
|
||||
|
||||
## 5. 候选验证名单
|
||||
|
||||
本轮候选验证名单：
|
||||
|
||||
- `空`
|
||||
|
||||
原因不是“生成器完全跑不起来”，而是“生成器已经能产出包，但这些包还不具备进入内网业务验证的价值”。
|
||||
|
||||
## 6. 本轮稳定确认的 blocker
|
||||
|
||||
### 6.1 archetype 坍缩
|
||||
|
||||
三份样本全部被判成：
|
||||
|
||||
- `paginated_enrichment`
|
||||
|
||||
而不是线损家族应接近的：
|
||||
|
||||
- `multi_mode_request`
|
||||
|
||||
### 6.2 bootstrap 稳定误选
|
||||
|
||||
三份样本全部落到：
|
||||
|
||||
- `20.77.115.36:31051`
|
||||
|
||||
而没有锚到线损主业务页面承载面。
|
||||
|
||||
### 6.3 modes 未恢复
|
||||
|
||||
三份样本都出现：
|
||||
|
||||
1. `modes = []`
|
||||
2. `defaultMode = "month"`
|
||||
3. `modeSwitchField = "period_mode"`
|
||||
|
||||
说明系统只保留了 mode 词面痕迹，没有恢复真正的 mode 结构。
|
||||
|
||||
### 6.4 request contract 缺失
|
||||
|
||||
三份样本都出现：
|
||||
|
||||
1. `params = []`
|
||||
2. `requestTemplate = null`
|
||||
3. 脚本用泛化 `args` 直接拼请求
|
||||
|
||||
说明当前还没有恢复出线损家族真正需要的请求合同。
|
||||
|
||||
### 6.5 column defs 缺失
|
||||
|
||||
三份样本都出现：
|
||||
|
||||
1. `columnDefs = []`
|
||||
2. `requiredFields = []`
|
||||
3. 用返回首行 key 动态拼列
|
||||
|
||||
这与 `tq` 级稳定报表结构还有明显距离。
|
||||
|
||||
### 6.6 endpoint 污染
|
||||
|
||||
三份样本都混入了大量无关 endpoint：
|
||||
|
||||
1. 其他业务系统接口
|
||||
2. 文档和第三方库 URL
|
||||
3. 静态依赖中的外链
|
||||
|
||||
说明当前 endpoint 提取还没有有效隔离业务主链与资源噪声。
|
||||
|
||||
### 6.7 readiness 过度乐观
|
||||
|
||||
三份样本都给了较高 readiness，但核心合同并未闭合。
|
||||
|
||||
这说明当前 readiness 还不能作为“可进内网验证”的判断依据。
|
||||
|
||||
## 7. 本轮确认的正向信号
|
||||
|
||||
虽然没有候选样本，但本轮也确认了三类有效正向信号：
|
||||
|
||||
1. 系统已经能从真实样本中抓到线损核心 endpoint 名称
|
||||
2. 系统已经能抓到 `month/week/tjzq/mode` 等模式词面信号
|
||||
3. 系统已经能把 `localhost:*` 识别为宿主依赖证据，而不是直接当成业务 bootstrap
|
||||
|
||||
这说明当前链路不是“完全失明”，而是“证据抽到了，但主业务语义编译失败”。
|
||||
|
||||
## 8. 本轮阶段结论
|
||||
|
||||
第一轮迁移阶段的正式结论是：
|
||||
|
||||
1. 当前生成链已经具备真实场景证据抽取能力
|
||||
2. 当前生成链还不具备 `G2/tq` 级业务主链重建能力
|
||||
3. 因此第一轮没有任何样本达到候选验证门槛
|
||||
4. 本轮最有价值的产出不是候选 skill，而是稳定复现出来的 `G2` 家族 blocker 画像
|
||||
|
||||
## 9. 后续输入
|
||||
|
||||
本报告形成后的后续输入只有两类：
|
||||
|
||||
1. 作为后续整改阶段的 blocker 基线
|
||||
2. 作为进入 `G1/G3` 首轮样本前的参考口径
|
||||
|
||||
本报告本身不包含整改方案，也不展开后续开发任务。
|
||||
@@ -0,0 +1,127 @@
|
||||
# G1 Boundary Reassignment Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `G1` 边界收紧与 4 个边界样本正式重排
|
||||
> Source:
|
||||
> - `examples/g1_batch_round1/skills/`
|
||||
> - `docs/superpowers/specs/2026-04-17-scene-skill-60-to-90-roadmap-design.md`
|
||||
> - `docs/superpowers/plans/2026-04-18-g1-boundary-reassignment-plan.md`
|
||||
|
||||
## 1. Conclusion
|
||||
|
||||
本轮分析结论已经冻结：
|
||||
|
||||
1. `G1` 需要整改，不能继续维持原有宽边界。
|
||||
2. 4 个边界样本中，只有 `高低压新增报装容量月度统计表` 保留在 `G1`，但应提升为 `G1-E 轻量补查汇总型`。
|
||||
3. `电能表现场检验完成率指标报表`、`计量资产库存统计`、`95598供电服务月报` 均应从 `G1` 中拆出，分别并入 `G6`、`G7`、`G8`。
|
||||
|
||||
## 2. Background
|
||||
|
||||
在第一轮 `G1` 批量迁移中，共选取 10 个“相对常用报表样本”进行实测生成。结果显示：
|
||||
|
||||
1. 虽然大部分样本可以完成生成流程，但其中一批场景并没有恢复出可靠的 `single_request_table` 语义。
|
||||
2. 多个样本被错误吸附到无关外域或被降级为 `page_state_eval`，说明当前 `G1` 边界过宽。
|
||||
3. 对 4 个“相对有希望”的边界样本继续深挖后，确认它们并不属于同一个单页单请求家族。
|
||||
|
||||
## 3. Sample Findings
|
||||
|
||||
### 3.1 高低压新增报装容量月度统计表
|
||||
|
||||
- 生成结果路径：
|
||||
`examples/g1_batch_round1/skills/g1-capacity-monthly/references/generation-report.json`
|
||||
- 当前落点：
|
||||
`page_state_eval`
|
||||
- 关键信号：
|
||||
- `bootstrap domain = yxgateway.gs.sgcc.com.cn`
|
||||
- `params=[]`
|
||||
- `modes=[]`
|
||||
- `requestEntries=[]`
|
||||
- `columnDefs=[]`
|
||||
- 源场景可见业务结构：
|
||||
- 主查询：`getWkorderAll`
|
||||
- 补查：`queryElectCustInfo`
|
||||
- 补查：`queryBusAcpt`
|
||||
- 补查：`getBatchPerCust97`
|
||||
- 结论：
|
||||
该场景仍保有主查询骨架，但不是严格单请求平面场景，适合作为 `G1-E` 上边界样本。
|
||||
|
||||
### 3.2 电能表现场检验完成率指标报表
|
||||
|
||||
- 生成结果路径：
|
||||
`examples/g1_batch_round1/skills/g1-meter-inspection-rate/references/generation-report.json`
|
||||
- 当前落点：
|
||||
`page_state_eval`
|
||||
- 关键信号：
|
||||
- `bootstrap domain = yxgateway.gs.sgcc.com.cn`
|
||||
- `params=[]`
|
||||
- `modes=[]`
|
||||
- `requestEntries=[]`
|
||||
- `columnDefs=[]`
|
||||
- 源场景可见业务结构：
|
||||
- `sgBrowserExcuteJsCode`
|
||||
- `BrowserAction`
|
||||
- `getWorkOrderToDoList`
|
||||
- `queryMeterPlanFormulateApp`
|
||||
- `queryMeterPlanDtlForAddMeter`
|
||||
- 结论：
|
||||
该场景本质上依赖宿主桥接与多步 callback workflow，不应再放入 `G1`，应重排到 `G6 宿主桥接多步查询型`。
|
||||
|
||||
### 3.3 计量资产库存统计
|
||||
|
||||
- 生成结果路径：
|
||||
`examples/g1_batch_round1/skills/g1-meter-asset-stock/references/generation-report.json`
|
||||
- 当前落点：
|
||||
`page_state_eval`
|
||||
- 关键信号：
|
||||
- `bootstrap domain = yxgateway.gs.sgcc.com.cn`
|
||||
- `params=[]`
|
||||
- `modes=[]`
|
||||
- `requestEntries=[]`
|
||||
- `columnDefs=[]`
|
||||
- 源场景可见业务结构：
|
||||
- `assetStatsQueryMeter`
|
||||
- `assetStatsQueryIt`
|
||||
- `assetStatsQueryAcqTrml`
|
||||
- `assetStatsQueryMeterCommonModule`
|
||||
- `assetStatsQueryJlGnModule`
|
||||
- 结论：
|
||||
该场景属于典型的多接口盘点汇总，不符合 `single_request_table` 假设，应重排到 `G7 多接口盘点汇总型`。
|
||||
|
||||
### 3.4 95598供电服务月报
|
||||
|
||||
- 生成结果路径：
|
||||
`examples/g1_batch_round1/skills/g1-95598-service-monthly/references/generation-report.json`
|
||||
- 当前落点：
|
||||
`page_state_eval`
|
||||
- 关键信号：
|
||||
- `bootstrap domain = south.95598.sgcc.com.cn`
|
||||
- `params=[]`
|
||||
- `modes=[]`
|
||||
- `requestEntries=[]`
|
||||
- `columnDefs=[]`
|
||||
- 源场景可见业务结构：
|
||||
- `BrowserAction('sgBrowerserJsAjax2', ...)`
|
||||
- `localhost:13313/configServices/selectData`
|
||||
- `definedSqlQuery`
|
||||
- `docExport(...)`
|
||||
- 结论：
|
||||
该场景不是直接报表导出，而是“抓取 -> 落库 -> SQL 分析 -> 文档生成”的复合链路，应重排到 `G8 抓取落库分析出文档型`。
|
||||
|
||||
## 4. Final Reassignment Table
|
||||
|
||||
| 样本场景 | 正式归属 | 是否继续作为普通 G1 样本 |
|
||||
| --- | --- | --- |
|
||||
| 高低压新增报装容量月度统计表 | G1-E 轻量补查汇总型 | 否 |
|
||||
| 电能表现场检验完成率指标报表 | G6 宿主桥接多步查询型 | 否 |
|
||||
| 计量资产库存统计 | G7 多接口盘点汇总型 | 否 |
|
||||
| 95598供电服务月报 | G8 抓取落库分析出文档型 | 否 |
|
||||
|
||||
说明：`高低压新增报装容量月度统计表` 虽然仍归入 `G1` 体系，但已作为 `G1-E` 边界子型使用，不再作为普通 `single_request_table` 样本。
|
||||
|
||||
## 5. Impact on Next Steps
|
||||
|
||||
本报告落地后，后续执行口径同步固定为：
|
||||
|
||||
1. 先按收紧后的 `G1` 规则继续推进 `G1-E`。
|
||||
2. 不再把 `G6/G7/G8` 样本塞回 `G1` 做“通用报表生成”尝试。
|
||||
3. 新家族能力落地前，对应样本应输出明确阻断原因与家族归属，而不是继续产出低质量伪可运行 skill。
|
||||
190
docs/superpowers/reports/2026-04-18-g1-e-p0-validation-report.md
Normal file
190
docs/superpowers/reports/2026-04-18-g1-e-p0-validation-report.md
Normal file
@@ -0,0 +1,190 @@
|
||||
# G1-E P0 Validation Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-g1-e-light-enrichment-report-plan.md` Phase 1 -> Phase 3
|
||||
> Fixed Sample:
|
||||
> - `D:\desk\智能体资料\全量业务场景\一平台场景\高低压新增报装容量月度统计表`
|
||||
|
||||
## 1. Conclusion
|
||||
|
||||
本轮 `G1-E` P0 验证已通过。
|
||||
|
||||
最终结果已经从误判的 `page_state_eval` 收敛为正式的 `single_request_enrichment`，并且真实样本生成结果中已经同时恢复出：
|
||||
|
||||
1. 主请求
|
||||
2. 补查请求集合
|
||||
3. 合并/汇总规则
|
||||
|
||||
这说明本轮 plan 的核心目标已经达成：系统不再只会“识别并拦截低质量生成物”，而是已经能对该 P0 样本恢复出可编译的 `G1-E` 三段式语义。
|
||||
|
||||
## 2. Validation Inputs
|
||||
|
||||
真实样本生成命令：
|
||||
|
||||
```powershell
|
||||
cargo run --bin sg_scene_generate -- --source-dir "D:/desk/智能体资料/全量业务场景/一平台场景/高低压新增报装容量月度统计表" --scene-id g1e-capacity-monthly --scene-name "高低压新增报装容量月度统计表" --scene-kind report_collection --output-root "D:/data/ideaSpace/rust/sgClaw/claw-new/examples/g1e_p0_validation"
|
||||
```
|
||||
|
||||
关键输出位置：
|
||||
|
||||
1. `examples/g1e_p0_validation/skills/g1e-capacity-monthly/references/generation-report.json`
|
||||
2. `examples/g1e_p0_validation/skills/g1e-capacity-monthly/references/generation-report.md`
|
||||
3. `examples/g1e_p0_validation/skills/g1e-capacity-monthly/scripts/collect_g1e_capacity_monthly.js`
|
||||
|
||||
## 3. Before vs After
|
||||
|
||||
整改前：
|
||||
|
||||
1. `workflowArchetype = page_state_eval`
|
||||
2. `readiness.notes` 中记录的是 `workflow_archetype=page_state_eval`
|
||||
3. 生成脚本走的是 `page_state_eval` / 简化脚本路径
|
||||
4. 虽然已经提取到 `mainRequest / enrichmentRequests / mergePlan`，但 compiler 没有真正走 `G1-E` 专用编译路径
|
||||
|
||||
整改后：
|
||||
|
||||
1. `workflowArchetype = single_request_enrichment`
|
||||
2. `generation-report.md` 明确显示 `Workflow archetype: single_request_enrichment`
|
||||
3. 生成脚本已经包含：
|
||||
- `const MAIN_REQUEST =`
|
||||
- `const ENRICHMENT_REQUESTS =`
|
||||
- `const MERGE_PLAN =`
|
||||
4. `readiness.notes` 已更新为 `workflow_archetype=single_request_enrichment`
|
||||
|
||||
## 4. Recovered Semantics
|
||||
|
||||
### 4.1 Main Request
|
||||
|
||||
已恢复主请求：
|
||||
|
||||
1. Endpoint: `getWkorderAll`
|
||||
2. URL: `http://yxgateway.gs.sgcc.com.cn/emss-cmnf-common-front/member/workOrderQuery/getWkorderAll`
|
||||
3. Request template:
|
||||
- `pageNo = 1`
|
||||
- `pageSize = 1`
|
||||
4. Response path: `data`
|
||||
|
||||
### 4.2 Enrichment Requests
|
||||
|
||||
已恢复补查请求：
|
||||
|
||||
1. `queryElectCustInfo`
|
||||
2. `queryBusAcpt`
|
||||
3. `getBatchPerCust97`
|
||||
|
||||
其中补查入参绑定已经恢复出：
|
||||
|
||||
1. `appNo = ${row.wkOrderNo}`
|
||||
2. `getBatchPerCust97` 额外固定参数：
|
||||
- `baseNewFlag = "01"`
|
||||
- `pageNo = 1`
|
||||
- `pageSize = 20`
|
||||
|
||||
### 4.3 Merge Plan
|
||||
|
||||
已恢复汇总语义：
|
||||
|
||||
1. Join key: `wkOrderNo`
|
||||
2. Group rule: `group_by:countyCodeName`
|
||||
3. 汇总附加规则：
|
||||
- `append_total_row`
|
||||
- `compute_share_columns`
|
||||
4. 聚合规则：
|
||||
- `com:hightVolTotalThisMonth,hightPressureTotalThisMonth`
|
||||
- `com:lowVolTotalThisMonth,lowPressureTotalThisMonth`
|
||||
- `batchCom:lowVolTotalThisMonth,lowPressureTotalThisMonth`
|
||||
- `com:hightVolTotalOtherMonth,hightPressureTotalOtherMonth`
|
||||
- `com:lowVolTotalOtherMonth,lowPressureTotalOtherMonth`
|
||||
- `batchCom:lowVolTotalOtherMonth,lowPressureTotalOtherMonth`
|
||||
|
||||
### 4.4 Output Columns
|
||||
|
||||
最终输出列已非空，且恢复出完整统计口径，包含：
|
||||
|
||||
1. `index`
|
||||
2. `countyCodeName`
|
||||
3. `hightPressureTotalOtherMonth`
|
||||
4. `hightVolTotalOtherMonth`
|
||||
5. `lowPressureTotalOtherMonth`
|
||||
6. `lowVolTotalOtherMonth`
|
||||
7. `otherMonthShare`
|
||||
8. `hightPressureTotalThisMonth`
|
||||
9. `hightVolTotalThisMonth`
|
||||
10. `lowPressureTotalThisMonth`
|
||||
11. `lowVolTotalThisMonth`
|
||||
12. `thisMonthShare`
|
||||
13. `yearHightPressureTotal`
|
||||
14. `yearHightVolTotal`
|
||||
15. `yearLowPressureTotal`
|
||||
16. `yearLowVolTotal`
|
||||
17. `yearShare`
|
||||
|
||||
## 5. Root Cause And Fix
|
||||
|
||||
本次误判的直接原因不是证据缺失，而是 archetype 分类优先级错误。
|
||||
|
||||
真实样本同时存在以下噪音信号：
|
||||
|
||||
1. `branch_fields=reportType, mode, month`
|
||||
2. `filters=res.status == 200`
|
||||
3. 常见分页字段
|
||||
|
||||
旧逻辑要求：
|
||||
|
||||
1. 只有在 `!has_branch_fields` 时，才允许判为 `SingleRequestEnrichment`
|
||||
|
||||
这导致真实样本即使已经同时满足：
|
||||
|
||||
1. `has_g1e_main_endpoint`
|
||||
2. `has_g1e_enrichment`
|
||||
3. `has_g1e_merge_signal`
|
||||
|
||||
仍然会因为存在 `mode/reportType/month` 噪音而失去 `G1-E` 判定资格，随后又被 `status` 噪音吸入 `page_state_eval`。
|
||||
|
||||
本次修正后的口径是：
|
||||
|
||||
1. 只要 `G1-E` 三段式核心证据齐备，就优先判为 `single_request_enrichment`
|
||||
2. 不再因为一般性的 `branch field` 噪音退化为 `page_state_eval`
|
||||
|
||||
## 6. Test And Verification Record
|
||||
|
||||
本轮执行的定向验证：
|
||||
|
||||
1. `cargo test --test scene_generator_test`
|
||||
2. `cargo test --test scene_generator_modes_test`
|
||||
3. `cargo test --test scene_generator_html_test`
|
||||
4. `node --test tests/scene_generator_llm_test.js`
|
||||
|
||||
结果：
|
||||
|
||||
1. 以上 `scene generator` 定向测试全部通过
|
||||
2. 真实 P0 样本复跑成功
|
||||
|
||||
补充说明：
|
||||
|
||||
1. `cargo test --workspace` 中存在两条与本次 `G1-E` plan 无关的既有失败：
|
||||
- `service::server::tests::deterministic_bootstrap_target_uses_plan_target_url`
|
||||
- `service::server::tests::bootstrap_target_precedence_matrix_covers_page_context_deterministic_skill_and_fallback`
|
||||
2. 这两条失败不属于本次 plan 改动引入，也不阻塞本次 `G1-E` P0 验收
|
||||
|
||||
## 7. Acceptance Check
|
||||
|
||||
对照 plan 的 Phase 3 退出标准，当前状态如下：
|
||||
|
||||
1. 主请求、补查请求、合并规则均已稳定恢复
|
||||
2. 结果不再是 `params=[] / requestEntries=[] / columnDefs=[]` 的空壳生成物
|
||||
3. archetype 已不再退化为 `page_state_eval`
|
||||
4. compiler 已实际走入 `G1-E` 专用路径
|
||||
|
||||
结论：
|
||||
|
||||
1. `docs/superpowers/plans/2026-04-18-g1-e-light-enrichment-report-plan.md` 的 Phase 1 -> Phase 3 已完成
|
||||
|
||||
## 8. Remaining Gap
|
||||
|
||||
本轮完成的是 `G1-E` P0 最小闭环，不代表已经达到“业务结果 100% 等价于人工高质量 skill”。
|
||||
|
||||
当前仍然保留的后续空间是：
|
||||
|
||||
1. 进一步提升字段中文标题编码质量
|
||||
2. 继续验证 `G1-E` 第二个真实样本是否能复用同一语义恢复路径
|
||||
3. 继续把 `G1-E` 与 `G6/G7/G8` 的越界判定做得更稳
|
||||
@@ -0,0 +1,71 @@
|
||||
# G1-E Second Sample Reuse Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `G1-E` second-sample reuse validation
|
||||
|
||||
## Summary
|
||||
|
||||
`G1-E` has moved from a single P0 anchor to a reusable family baseline.
|
||||
|
||||
This round added a second repo-local `G1-E` expansion fixture and promoted the result into the P1 family assets. The expansion fixture keeps the same family contract as the P0 sample:
|
||||
|
||||
1. one resolved main request
|
||||
2. lightweight enrichment requests
|
||||
3. a merge plan with join keys, aggregate rules, and output columns
|
||||
|
||||
The result validates that `G1-E` should remain distinct from plain `single_request_table`, while still staying below the `G6/G7/G8` boundary.
|
||||
|
||||
## Assets Updated
|
||||
|
||||
1. `tests/fixtures/generated_scene/g1e_light_enrichment_expansion/index.html`
|
||||
2. `tests/fixtures/generated_scene/p1_family_manifest.json`
|
||||
3. `tests/fixtures/generated_scene/p1_family_results.json`
|
||||
4. `tests/scene_generator_test.rs`
|
||||
5. `tests/scene_generator_p1_family_test.rs`
|
||||
6. `tests/scene_generator_family_policy_test.rs`
|
||||
|
||||
## Representative And Expansion Baseline
|
||||
|
||||
| Role | Fixture | Expected Archetype |
|
||||
| --- | --- | --- |
|
||||
| Representative | `tests/fixtures/generated_scene/g1e_light_enrichment` | `single_request_enrichment` |
|
||||
| Expansion | `tests/fixtures/generated_scene/g1e_light_enrichment_expansion` | `single_request_enrichment` |
|
||||
|
||||
The expansion baseline asserts:
|
||||
|
||||
1. main request contains `getWkorderAll`
|
||||
2. enrichment request contains `queryMeterInfo`
|
||||
3. merge join key contains `wkOrderNo`
|
||||
4. merge aggregate rules contain `group_by:countyCodeName`
|
||||
5. output columns contain `meterCapacityThisMonth`
|
||||
|
||||
## Validation
|
||||
|
||||
Executed:
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test -- --nocapture
|
||||
cargo test --test scene_generator_p1_family_test -- --nocapture
|
||||
cargo test --test scene_generator_family_policy_test -- --nocapture
|
||||
cargo test --test scene_generator_canonical_test -- --nocapture
|
||||
```
|
||||
|
||||
All four regression suites passed.
|
||||
|
||||
## Result
|
||||
|
||||
`G1-E` reuse is now validated at family level:
|
||||
|
||||
1. `representativeRuns = 1`
|
||||
2. `expansionRuns = 1`
|
||||
3. `passedRuns = 2`
|
||||
4. `failedRuns = 0`
|
||||
5. `successRate = 1.0`
|
||||
|
||||
## Next Step
|
||||
|
||||
Per the frozen roadmap order, the next implementation line is `G6`:
|
||||
|
||||
1. start independent `G6` spec / plan
|
||||
2. use `电能表现场检验完成率指标报表` as the first boundary sample
|
||||
3. build fail-closed classification for host-bridge multi-step workflow before attempting runnable generation
|
||||
@@ -0,0 +1,28 @@
|
||||
# G1-E Additional Expansion Report
|
||||
|
||||
> Status: Completed
|
||||
> Date: 2026-04-18
|
||||
> Plan: [2026-04-17-scene-skill-60-to-90-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-17-scene-skill-60-to-90-roadmap-plan.md)
|
||||
|
||||
## Intent
|
||||
|
||||
Consume the remaining queued `G1-E` downstream candidate within the existing `single_request_enrichment` family contract.
|
||||
|
||||
## Delivered
|
||||
|
||||
1. Added `tests/fixtures/generated_scene/g1e_light_enrichment_additional`
|
||||
2. Added deterministic and generation regressions for the additional sample
|
||||
3. Promoted the sample into the formal `G1-E` batch expansion baseline
|
||||
4. Synchronized family manifest, family results, ledger overlay, and roadmap execution status
|
||||
|
||||
## Acceptance
|
||||
|
||||
1. archetype remains `single_request_enrichment`
|
||||
2. main request remains `getWkorderAll`
|
||||
3. enrichment request is recovered as `queryBusAcpt`
|
||||
4. merge contract still reuses `wkOrderNo` plus `group_by:countyCodeName`
|
||||
5. `Track B` no longer keeps a queued downstream candidate
|
||||
|
||||
## Notes
|
||||
|
||||
This promotion stays inside the roadmap boundary because it only consumes the existing queued `G1-E` item and does not create a new family or archetype.
|
||||
@@ -0,0 +1,52 @@
|
||||
# G1-E Candidate Batch Alignment Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round keeps the roadmap inside the `G1-E` Track B and Track D boundary.
|
||||
|
||||
It does not add a third `G1-E` implementation sample. It converts the already validated representative plus first expansion state into a formal candidate-batch asset.
|
||||
|
||||
## Delivered
|
||||
|
||||
Added or updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/g1e_candidate_batch_2026-04-18.json`
|
||||
2. `tests/g1e_candidate_batch_test.rs`
|
||||
3. `tests/fixtures/generated_scene/p1_family_manifest.json`
|
||||
4. `tests/fixtures/generated_scene/p1_family_results.json`
|
||||
5. `tests/fixtures/generated_scene/scene_ledger_status_2026-04-18.json`
|
||||
6. `tests/scene_ledger_status_test.rs`
|
||||
|
||||
## Batch Meaning
|
||||
|
||||
The batch asset now freezes the current `G1-E` status as:
|
||||
|
||||
1. one promoted baseline
|
||||
2. one promoted expansion
|
||||
3. one queued downstream candidate
|
||||
|
||||
This keeps `G1-E` aligned with the same family-asset shape already used by `G2` and `G3`, without forcing unsupported expansion beyond the current evidence.
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
1. `cargo test --test g1e_candidate_batch_test -- --nocapture`
|
||||
2. `cargo test --test scene_generator_p1_family_test -- --nocapture`
|
||||
3. `cargo test --test scene_ledger_status_test -- --nocapture`
|
||||
4. `cargo test --test scene_generator_family_policy_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
`G1-E` now has:
|
||||
|
||||
1. representative baseline
|
||||
2. promoted expansion baseline
|
||||
3. formal candidate batch
|
||||
4. synchronized Track E ledger overlay status
|
||||
|
||||
That completes the same core asset shape for the three roadmap mainline families:
|
||||
|
||||
1. `G1-E`
|
||||
2. `G2`
|
||||
3. `G3`
|
||||
@@ -0,0 +1,60 @@
|
||||
# G2 Candidate Batch Alignment Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round continues the roadmap Track D and Track E mainline work for `G2`.
|
||||
|
||||
It converts the current line-loss family progress into a formal repo-local batch asset instead of leaving the state split across separate markdown reports only.
|
||||
|
||||
## Delivered
|
||||
|
||||
Added or updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/g2_candidate_batch_2026-04-18.json`
|
||||
2. `tests/g2_candidate_batch_test.rs`
|
||||
3. `tests/fixtures/generated_scene/p1_family_manifest.json`
|
||||
4. `tests/fixtures/generated_scene/p1_family_results.json`
|
||||
5. `tests/fixtures/generated_scene/scene_ledger_status_2026-04-18.json`
|
||||
6. `tests/scene_generator_p1_family_test.rs`
|
||||
7. `tests/scene_ledger_status_test.rs`
|
||||
|
||||
## Batch Meaning
|
||||
|
||||
The batch asset now freezes the current `G2` line-loss family status as:
|
||||
|
||||
1. one `P0` anchor baseline
|
||||
2. four promoted expansion baselines
|
||||
3. one deferred queue item
|
||||
|
||||
The promoted expansion baselines are:
|
||||
|
||||
1. week-only mode variant
|
||||
2. mixed linked workflow variant
|
||||
3. comparison crosscheck variant
|
||||
4. diagnosis drilldown variant
|
||||
|
||||
## Deferred Scope
|
||||
|
||||
`G2-D` remains deferred in this round because the current representative sample still does not expose a stable business endpoint contract.
|
||||
|
||||
That stays within the roadmap plan boundary and does not force non-contract-first implementation work into the current mainline.
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
1. `cargo test --test g2_candidate_batch_test -- --nocapture`
|
||||
2. `cargo test --test scene_generator_p1_family_test -- --nocapture`
|
||||
3. `cargo test --test scene_ledger_status_test -- --nocapture`
|
||||
4. `cargo test --test scene_generator_family_policy_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
`G2` now matches the same Track D and Track E asset shape already established for `G3`:
|
||||
|
||||
1. representative baseline
|
||||
2. promoted expansion baselines
|
||||
3. formal batch asset
|
||||
4. current-status ledger overlay
|
||||
|
||||
That means the remaining roadmap work can continue from a stable asset baseline instead of reconstructing `G2` progress from scattered markdown reports.
|
||||
@@ -0,0 +1,45 @@
|
||||
# G2-D Prediction Compute Report
|
||||
|
||||
> Status: Completed
|
||||
> Date: 2026-04-18
|
||||
> Plan: [2026-04-17-scene-skill-60-to-90-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-17-scene-skill-60-to-90-roadmap-plan.md)
|
||||
|
||||
## Intent
|
||||
|
||||
Within the existing `G2` line-loss family contract, promote the last deferred `G2-D` candidate into a repo-local reusable expansion baseline without crossing into new runtime families or login/transport scope.
|
||||
|
||||
## Scope
|
||||
|
||||
This slice only adds:
|
||||
|
||||
1. a deterministic `G2-D` fixture
|
||||
2. subtype classification for `g2_d_prediction_compute`
|
||||
3. a minimal reusable `prediction` mode recovery path
|
||||
4. regression coverage
|
||||
5. Track D / Track E asset synchronization
|
||||
|
||||
It does not introduce:
|
||||
|
||||
1. file-input runtime support
|
||||
2. local compute engine execution
|
||||
3. new family boundaries outside `G2`
|
||||
|
||||
## Delivered
|
||||
|
||||
1. Added fixture: `tests/fixtures/generated_scene/g2_prediction_compute`
|
||||
2. Added analyzer subtype: `G2FamilyVariant::G2D`
|
||||
3. Added generator mode recovery: `prediction`
|
||||
4. Added deterministic and generation regressions
|
||||
5. Promoted `predicted_compute_variant` into the `G2` batch expansion baseline
|
||||
6. Synchronized candidate batch, family manifest/results, ledger status overlay, and roadmap execution status
|
||||
|
||||
## Acceptance
|
||||
|
||||
1. `G2-D` now compiles through the existing `multi_mode_request` family path
|
||||
2. generated default mode is `prediction`
|
||||
3. generated report contains `lineId`, `lineLossRate`, and `powerLoss`
|
||||
4. Track A no longer keeps `deferred-g2-d-candidate` as an outstanding item
|
||||
|
||||
## Notes
|
||||
|
||||
This promotion stays inside the roadmap boundary because it only consumes the remaining queued/deferred `G2` family item and does not open new archetype scope.
|
||||
@@ -0,0 +1,142 @@
|
||||
# G2 家族扩展第三轮回归报告
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-g2-family-expansion-plan.md`
|
||||
> Family: `G2`
|
||||
|
||||
## 1. 本轮执行范围
|
||||
|
||||
本轮严格承接上一轮 `G2` 主样本整改，不回退 `tq` 主样本链路，只补以下两类剩余变体：
|
||||
|
||||
1. `G2-B`
|
||||
周报单侧 mode 变体
|
||||
代表样本：`白银线损周报`
|
||||
2. `G2-C`
|
||||
线损主链路 + 外部系统联动的混合变体
|
||||
代表样本：`线损同期差异报表`
|
||||
|
||||
本轮实际落地内容：
|
||||
|
||||
1. 在 analyzer 中引入 `G2FamilyVariant`
|
||||
2. 建立 `G2-A/G2-B/G2-C` 三类子型边界
|
||||
3. 为 `G2-B` 补单周模式最小合同
|
||||
4. 为 `G2-C` 补主链路模式最小合同
|
||||
5. 保持 `G2-A` 主样本不回退
|
||||
|
||||
## 2. 回归测试结果
|
||||
|
||||
本地回归：
|
||||
|
||||
- `cargo test --test scene_generator_test -- --nocapture`
|
||||
- 结果：`19 passed`
|
||||
|
||||
新增回归覆盖了：
|
||||
|
||||
1. `G2-A` 主样本仍被识别为双模式主报表
|
||||
2. `G2-B` 被识别为周报单侧 mode 变体
|
||||
3. `G2-C` 被识别为混合联动变体
|
||||
4. `G2-B` 可以生成最小合同
|
||||
5. `G2-C` 可以生成最小合同
|
||||
|
||||
## 3. 真实样本第三轮结果
|
||||
|
||||
### 3.1 G2-A 基线
|
||||
|
||||
样本：
|
||||
|
||||
- `台区线损大数据-月_周累计线损率统计分析`
|
||||
- 产物：`examples/real_scene_batch_round1/skills/real-tq-lineloss-report-r1`
|
||||
|
||||
结果：
|
||||
|
||||
1. `workflowArchetype = multi_mode_request`
|
||||
2. `bootstrap.expectedDomain = 20.76.57.61:18080`
|
||||
3. `bootstrap.targetUrl = http://20.76.57.61:18080/gsllys`
|
||||
4. `modes = [month, week]`
|
||||
5. `g2_family_variant_candidate` 已写入
|
||||
6. `g2_request_contract_complete = true`
|
||||
7. `g2_response_contract_complete = true`
|
||||
8. `readiness.level = A`
|
||||
|
||||
结论：
|
||||
|
||||
- `G2-A` 基线稳定，没有回退
|
||||
|
||||
### 3.2 G2-B 周报单侧 mode 变体
|
||||
|
||||
样本：
|
||||
|
||||
- `白银线损周报`
|
||||
- 产物：`examples/real_scene_batch_round1/skills/real-baiyin-lineloss-weekly-r1`
|
||||
|
||||
结果：
|
||||
|
||||
1. 已成功生成，不再 fail-close
|
||||
2. `workflowArchetype = multi_mode_request`
|
||||
3. `bootstrap.expectedDomain = 20.76.57.61:18080`
|
||||
4. `bootstrap.targetUrl = http://20.76.57.61:18080/gsllys`
|
||||
5. `modes = [week]`
|
||||
6. 产物中已恢复 `LINE_LOSS_RATE`
|
||||
7. `g2_family_variant_candidate` 已写入
|
||||
8. `g2_request_contract_complete = true`
|
||||
9. `g2_response_contract_complete = true`
|
||||
10. `readiness.level = A`
|
||||
|
||||
结论：
|
||||
|
||||
- `G2-B` 已从“只能 fail-close”推进到“可解释最小合同”
|
||||
- 已可进入候选验证名单
|
||||
|
||||
### 3.3 G2-C 混合联动变体
|
||||
|
||||
样本：
|
||||
|
||||
- `线损同期差异报表`
|
||||
- 产物：`examples/real_scene_batch_round1/skills/real-lineloss-period-diff-r1`
|
||||
|
||||
结果：
|
||||
|
||||
1. 已成功生成，不再 fail-close
|
||||
2. `workflowArchetype = multi_mode_request`
|
||||
3. `bootstrap.expectedDomain = 20.76.57.61:18080`
|
||||
4. `bootstrap.targetUrl = http://20.76.57.61:18080/gsllys`
|
||||
5. `modes = [primary]`
|
||||
6. 产物中已恢复 `TG_NO`
|
||||
7. `g2_family_variant_candidate` 已写入
|
||||
8. `g2_request_contract_complete = true`
|
||||
9. `g2_response_contract_complete = true`
|
||||
10. `readiness.level = A`
|
||||
|
||||
结论：
|
||||
|
||||
- `G2-C` 已从“混合噪声导致整体阻断”推进到“主链路最小合同可生成”
|
||||
- 当前可进入候选验证名单
|
||||
|
||||
## 4. 与上一轮相比的变化
|
||||
|
||||
上一轮状态：
|
||||
|
||||
1. `G2-A` 已通
|
||||
2. `G2-B` fail-close
|
||||
3. `G2-C` fail-close
|
||||
|
||||
本轮状态：
|
||||
|
||||
1. `G2-A` 保持稳定
|
||||
2. `G2-B` 已恢复单周模式合同
|
||||
3. `G2-C` 已恢复主链路模式合同
|
||||
4. 三份样本都不再被同一种 `G2-A` 模型粗暴处理
|
||||
|
||||
## 5. 当前结论
|
||||
|
||||
按本计划的目标口径，本轮已经完成：
|
||||
|
||||
1. `G2` 从“单主样本修通”推进到“至少三类子型可区分”
|
||||
2. `白银线损周报` 不再只是被动 fail-close
|
||||
3. `线损同期差异报表` 不再只是被动 fail-close
|
||||
4. `G2-A/G2-B/G2-C` 都已进入候选验证名单
|
||||
|
||||
当前最准确的状态判断是：
|
||||
|
||||
- `G2` 家族扩展第一轮已达标
|
||||
- 下一步如果继续推进，应切到更大范围的线损变体扩展或真实内网验证
|
||||
@@ -0,0 +1,141 @@
|
||||
# G2 家族首轮 Blocker 汇总
|
||||
|
||||
日期：2026-04-18
|
||||
|
||||
样本范围：
|
||||
|
||||
1. `台区线损大数据-月_周累计线损率统计分析`
|
||||
2. `白银线损周报`
|
||||
3. `线损同期差异报表`
|
||||
|
||||
对应生成产物：
|
||||
|
||||
1. `examples/real_scene_batch_round1/skills/real-tq-lineloss-report-r1`
|
||||
2. `examples/real_scene_batch_round1/skills/real-baiyin-lineloss-weekly-r1`
|
||||
3. `examples/real_scene_batch_round1/skills/real-lineloss-period-diff-r1`
|
||||
|
||||
## 1. 当前结论
|
||||
|
||||
这三个 `G2` 真实样本已经足够说明，当前生成链在“线损多模式报表家族”上存在稳定的共性 blocker，而不是单个场景的偶发误差。
|
||||
|
||||
当前状态可以概括为：
|
||||
|
||||
1. 能抓到部分线损业务信号
|
||||
2. 但不会把这些信号重建成 `multi_mode_request.month_week_table`
|
||||
3. 反而会稳定坍缩到 `paginated_enrichment`
|
||||
|
||||
## 2. 已稳定复现的共性 blocker
|
||||
|
||||
### 2.1 archetype 稳定误判
|
||||
|
||||
三个样本都被判成：
|
||||
|
||||
- `paginated_enrichment`
|
||||
|
||||
而不是：
|
||||
|
||||
- `multi_mode_request`
|
||||
|
||||
这说明当前 workflow 判定时，分页、补数、过滤等噪声信号权重明显高于 `month/week` 模式矩阵信号。
|
||||
|
||||
### 2.2 bootstrap 稳定落错
|
||||
|
||||
三个样本都落到：
|
||||
|
||||
- `expected_domain = 20.77.115.36:31051`
|
||||
- `target_url = http://20.77.115.36:31051`
|
||||
|
||||
这说明当前 bootstrap 选择逻辑没有把“真实业务承载页”与“其他可见入口页面”区分开。
|
||||
|
||||
### 2.3 modes 稳定丢失
|
||||
|
||||
三个样本都有这些共性：
|
||||
|
||||
1. `modes = []`
|
||||
2. `defaultMode = "month"`
|
||||
3. `modeSwitchField = "period_mode"`
|
||||
|
||||
也就是说，系统能感知到 mode 相关词，但不会把它编译成真正可执行的 mode 结构。
|
||||
|
||||
### 2.4 request contract 稳定缺失
|
||||
|
||||
三个样本都出现：
|
||||
|
||||
1. `params = []`
|
||||
2. `requestTemplate = null`
|
||||
3. 最终脚本把 `args` 直接整体塞进请求体
|
||||
|
||||
这说明当前链路没有恢复出线损家族真正需要的 mode-specific 请求合同。
|
||||
|
||||
### 2.5 column defs 稳定缺失
|
||||
|
||||
三个样本都出现：
|
||||
|
||||
1. `columnDefs = []`
|
||||
2. `normalizeRules.requiredFields = []`
|
||||
3. 脚本运行时用 `rows[0]` 动态反推列
|
||||
|
||||
这意味着业务列语义完全没有固化下来。
|
||||
|
||||
### 2.6 endpoint 污染稳定存在
|
||||
|
||||
三个样本都同时抓到了：
|
||||
|
||||
1. 线损业务接口
|
||||
2. 95598/营销等其他业务接口
|
||||
3. 大量第三方库、文档、外链 URL
|
||||
|
||||
这说明当前 endpoint candidate 过滤还不够，静态资源和依赖库内容仍在污染主业务判定。
|
||||
|
||||
### 2.7 readiness 稳定过度乐观
|
||||
|
||||
三个样本都给出：
|
||||
|
||||
- `Readiness: A`
|
||||
|
||||
但实际核心业务合同明显没有闭合。
|
||||
|
||||
这说明 readiness 现在还主要在反映“结构上能生成”，而不是“业务上接近可运行”。
|
||||
|
||||
## 3. 目前可以确认的正向信号
|
||||
|
||||
虽然不通过，但也有三类正向信号已经稳定出现：
|
||||
|
||||
1. 能抓到线损家族核心 endpoint 名称
|
||||
2. 能抓到 `month/week/tjzq/mode` 这类模式信号
|
||||
3. 能把 `localhost:*` 归入宿主依赖证据，而不是直接把它选成 bootstrap
|
||||
|
||||
这说明系统不是“完全盲”，而是“证据有了，但主链重建失败”。
|
||||
|
||||
## 4. 当前最小判断
|
||||
|
||||
基于前三个 `G2` 样本，当前可以给出一个更稳定的最小判断：
|
||||
|
||||
1. 当前链路已具备 `G2` 证据抽取能力
|
||||
2. 当前链路尚不具备 `G2` 业务主链重建能力
|
||||
3. 因此现在不能把 `G2` 结果视为可进入内网验证的候选 skill
|
||||
|
||||
## 5. 第三个 G2 样本验证结果
|
||||
|
||||
`线损同期差异报表` 已完成首轮生成，结果继续复现以下问题：
|
||||
|
||||
1. archetype 是否仍会坍缩到 `paginated_enrichment`
|
||||
2. bootstrap 是否仍会被错误锚到 `20.77.115.36:31051`
|
||||
3. endpoint 污染是否仍然严重
|
||||
4. 当场景引入更明显的对比链、secondary chain 后，是否会进一步放大当前偏差
|
||||
|
||||
当前结果表明：
|
||||
|
||||
1. 以上问题全部继续复现
|
||||
2. 且第三个样本的 `filter` 证据还出现了更明显的脏串污染
|
||||
3. 因此 `G2` blocker 画像可以视为已经稳定
|
||||
|
||||
## 6. 当前阶段是否进入 G1/G3
|
||||
|
||||
建议：`可以结束 G2 首轮收敛，再决定是否切换`
|
||||
|
||||
原因：
|
||||
|
||||
1. 执行表顺序本来就是先打透 `G2`
|
||||
2. 当前 `G2` blocker 已经通过第三个样本完成收口
|
||||
3. 后续如果继续推进，应该进入“整理首轮迁移报告”或“开始 G1/G3 首轮样本生成”，而不是回头继续补更多 `G2` 观察样本
|
||||
@@ -0,0 +1,159 @@
|
||||
# G2 第二轮整改回归报告
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-g2-remediation-plan.md`
|
||||
> Family: `G2`
|
||||
|
||||
## 1. 本轮整改落地内容
|
||||
|
||||
本轮严格围绕 `G2` 家族整改，实际落地了以下链路修复：
|
||||
|
||||
1. `multi_mode_request` 识别优先级前置，避免被分页噪声重新压回 `paginated_enrichment`
|
||||
2. `bootstrap` 候选从“仅命名 URL”扩展为“命名 URL + endpoint 业务根路径”联合评分，并保留已有 `sourceUrl` 回归口径
|
||||
3. `request_template` 支持以下真实场景写法：
|
||||
- `data: JSON.stringify(datas)`
|
||||
- `data: ${JSON.stringify(datas)}`
|
||||
- `let/const/var datas = {...}` 变量间接引用
|
||||
- 对象简写字段，如 `weekSfdate`
|
||||
4. `cols1/cols2` 改为平衡括号抽取，不再依赖 `];` 结尾，修复真实线损场景中两组列被吞并的问题
|
||||
5. `G2` readiness gate 继续维持 fail-close：
|
||||
- `g2_modes_present`
|
||||
- `g2_request_contract_complete`
|
||||
- `g2_response_contract_complete`
|
||||
|
||||
## 2. 回归测试结果
|
||||
|
||||
本地回归：
|
||||
|
||||
- `cargo test --test scene_generator_test -- --nocapture`
|
||||
- 结果：`15 passed`
|
||||
|
||||
新增和强化的回归口径已覆盖：
|
||||
|
||||
1. `G2` 噪声场景不再误判为 `paginated_enrichment`
|
||||
2. `localhost` 仍只作为宿主依赖保留
|
||||
3. `data -> JSON.stringify(var)` 的请求模板能够恢复
|
||||
4. 缺少 `G2` 合同时继续被阻断
|
||||
5. 不带分号的 `cols1/cols2` 仍能正确拆出 `month/week` 列定义
|
||||
|
||||
## 3. 真实样本第二轮结果
|
||||
|
||||
### 3.1 台区线损大数据-月_周累计线损率统计分析
|
||||
|
||||
样本路径：
|
||||
|
||||
- 源场景：`D:\desk\智能体资料\全量业务场景\一平台场景\台区线损大数据-月_周累计线损率统计分析`
|
||||
- 生成产物：`examples/real_scene_batch_round1/skills/real-tq-lineloss-report-r4`
|
||||
|
||||
结果：
|
||||
|
||||
1. 已成功生成，不再 fail-close
|
||||
2. `workflowArchetype = multi_mode_request`
|
||||
3. `bootstrap.expectedDomain = 20.76.57.61:18080`
|
||||
4. `bootstrap.targetUrl = http://20.76.57.61:18080/gsllys`
|
||||
5. `modes = [month, week]`
|
||||
6. `month/week` 均具备非空 `requestTemplate`
|
||||
7. `month/week` 均具备非空 `columnDefs`
|
||||
8. `responsePath = content`
|
||||
9. `g2_request_contract_complete = true`
|
||||
10. `g2_response_contract_complete = true`
|
||||
11. `readiness.level = A`
|
||||
|
||||
结论：
|
||||
|
||||
- 该样本已进入“候选验证名单”
|
||||
|
||||
### 3.2 白银线损周报
|
||||
|
||||
样本路径：
|
||||
|
||||
- 源场景：`D:\desk\智能体资料\全量业务场景\一平台场景\白银线损周报`
|
||||
- 生成尝试：`real-baiyin-lineloss-weekly-r1`
|
||||
|
||||
结果：
|
||||
|
||||
1. 仍然 fail-close
|
||||
2. 生成器报错：`workflow evidence is incomplete for archetype multi_mode_request`
|
||||
|
||||
直接证据：
|
||||
|
||||
1. 场景内能抓到周维接口：
|
||||
- `getYearMonWeekLinelossAnalysisList`
|
||||
2. 同时还能抓到台区排行接口：
|
||||
- `getTqLinelossInfoListRank`
|
||||
3. 但当前最小证据审计中，没有看到与 `tq` 主样本同级的显式 `cols1/cols2`
|
||||
4. 也没有看到月/周双模式都完整闭合的 response-side 合同证据
|
||||
|
||||
结论：
|
||||
|
||||
- 当前仍应维持 fail-close
|
||||
- 该样本说明：`G2` 家族中还存在“只有部分 week/mode 信号，但没有完整列合同”的变体
|
||||
|
||||
### 3.3 线损同期差异报表
|
||||
|
||||
样本路径：
|
||||
|
||||
- 源场景：`D:\desk\智能体资料\全量业务场景\一平台场景\线损同期差异报表`
|
||||
- 生成尝试：`real-lineloss-period-diff-r1`
|
||||
|
||||
结果：
|
||||
|
||||
1. 仍然 fail-close
|
||||
2. 生成器报错：`workflow evidence is incomplete for archetype multi_mode_request`
|
||||
|
||||
直接证据：
|
||||
|
||||
1. 主要抓到的是：
|
||||
- `getTqLinelossInfoListRank`
|
||||
- `getUserElectricList`
|
||||
- `tqQualifyRateMonitor`
|
||||
2. 同时还混入了同期系统链路：
|
||||
- `10.4.39.180/xsgl/...`
|
||||
3. 当前最小证据审计中，没有看到 `tq` 主样本那样稳定的 `month/week + cols1/cols2` 完整合同
|
||||
|
||||
结论：
|
||||
|
||||
- 当前仍应维持 fail-close
|
||||
- 该样本更像“线损主场景 + 同期系统联动”的混合工作流，而不是已经闭合的 `tq` 主报表双模式合同
|
||||
|
||||
## 4. 与第一轮相比的变化
|
||||
|
||||
第一轮 blocker：
|
||||
|
||||
1. 主样本会退化成 `paginated_enrichment`
|
||||
2. `bootstrap` 落错
|
||||
3. `request_template` 丢失
|
||||
4. `cols1/cols2` 串并，导致 mode response contract 不成立
|
||||
5. readiness 会给出失真的高分或错误通行
|
||||
|
||||
第二轮结果：
|
||||
|
||||
1. 主样本不再退化成 `paginated_enrichment`
|
||||
2. 主样本 `bootstrap` 已落到 `http://20.76.57.61:18080/gsllys`
|
||||
3. 主样本 `month/week` 请求模板已恢复
|
||||
4. 主样本 `month/week` 列合同已恢复
|
||||
5. 主样本已进入候选验证名单
|
||||
6. 其余两份样本继续 fail-close，没有被错误放行
|
||||
|
||||
## 5. 仍保留的 blocker
|
||||
|
||||
本轮没有解决、但已经被明确收敛出来的剩余问题：
|
||||
|
||||
1. `G2` 家族内部并不只有 `tq` 主报表一种结构，至少还存在：
|
||||
- 周报偏单侧 mode 信号变体
|
||||
- 同期差异联动型混合工作流
|
||||
2. 当前 endpoint 候选中仍有较多业务外噪声 URL 残留，但已不再阻塞主样本进入候选验证
|
||||
3. `G2` 主样本已经达到候选验证门槛，但“家族扩展到更多线损变体”仍未完成
|
||||
|
||||
## 6. 本轮结论
|
||||
|
||||
按本计划的目标口径，本轮整改已经完成以下关键结果：
|
||||
|
||||
1. `台区线损大数据-月_周累计线损率统计分析` 已从稳定失败推进到候选验证可用
|
||||
2. 另外两份真实样本没有被错误放行，继续保持 fail-close
|
||||
3. 第二轮整改已经把 `G2` 从“主样本无法成型”推进到“主样本可成型、变体样本被明确隔离”
|
||||
|
||||
当前最准确的状态判断是：
|
||||
|
||||
- `G2` 主样本整改已达标
|
||||
- `G2` 家族扩展仍未完成
|
||||
@@ -0,0 +1,32 @@
|
||||
# G3 Batch Closure Report
|
||||
|
||||
> Status: Completed
|
||||
> Date: 2026-04-18
|
||||
> Plan: [2026-04-17-scene-skill-60-to-90-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-17-scene-skill-60-to-90-roadmap-plan.md)
|
||||
|
||||
## Intent
|
||||
|
||||
Consume the remaining `G3` downstream candidates within the existing paginated-enrichment family contract and close the current Track C queue.
|
||||
|
||||
## Delivered
|
||||
|
||||
1. Added five more `G3` paginated-enrichment expansion fixtures:
|
||||
- device monitor weekly
|
||||
- customer satisfaction daily
|
||||
- repair return analysis
|
||||
- repair daily control
|
||||
- business stats
|
||||
2. Added deterministic request-template coverage and generation regressions for all five fixtures
|
||||
3. Promoted all remaining current downstream `G3` candidates into formal batch expansion baselines
|
||||
4. Synchronized family manifest, family results, ledger overlay, and roadmap execution status
|
||||
|
||||
## Acceptance
|
||||
|
||||
1. all promoted fixtures remain inside `paginated_enrichment`
|
||||
2. all promoted fixtures reuse `ticketNo` as join key
|
||||
3. aggregate rules stay within existing `aggregate:riskLevel` and `aggregate:sourceType` scope
|
||||
4. `Track C` remaining queue is now empty
|
||||
|
||||
## Notes
|
||||
|
||||
This closes the current repo-local `G3` batch within the roadmap boundary. It does not introduce a new archetype, runtime family, or external validation scope.
|
||||
@@ -0,0 +1,57 @@
|
||||
# G3 Candidate Batch Alignment Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round no longer leaves the `95598-ticket-family-candidate` ledger cluster as a raw batch queue.
|
||||
|
||||
It promotes the batch into a formal `G3` family asset that now records:
|
||||
|
||||
1. the `P0` anchor
|
||||
2. the first expansion anchor
|
||||
3. the second-round promoted expansion baselines
|
||||
|
||||
## Delivered
|
||||
|
||||
Added or updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/g3_candidate_batch_2026-04-18.json`
|
||||
2. `tests/g3_candidate_batch_test.rs`
|
||||
3. `tests/fixtures/generated_scene/p1_family_manifest.json`
|
||||
4. `tests/fixtures/generated_scene/p1_family_results.json`
|
||||
5. `tests/scene_generator_p1_family_test.rs`
|
||||
6. `tests/scene_generator_family_policy_test.rs`
|
||||
|
||||
## Batch Meaning
|
||||
|
||||
The batch asset now freezes five things together:
|
||||
|
||||
1. `paginated_enrichment` remains the `G3 / P0-3` anchor baseline
|
||||
2. `paginated_enrichment_expansion` remains the first expansion baseline
|
||||
3. `paginated_enrichment_expansion_workorder` is the promoted second expansion baseline
|
||||
4. `paginated_enrichment_expansion_orderno` is the promoted third expansion baseline
|
||||
5. the rest of the `95598-ticket-family-candidate` ledger cluster remains the downstream candidate queue
|
||||
|
||||
## Shared Contract Direction
|
||||
|
||||
The batch explicitly records the current reusable `G3` contract direction:
|
||||
|
||||
1. `archetype = paginated_enrichment`
|
||||
2. pagination fields stay in the `page/pageNum/pageNo/pageSize` family
|
||||
3. join keys stay in the `ticketNo/workOrderNo/orderNo` family
|
||||
4. aggregate rules stay in the `riskLevel/sourceType` family
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
1. `cargo test --test g3_candidate_batch_test -- --nocapture`
|
||||
2. `cargo test --test scene_generator_test generator_writes_paginated_enrichment_workorder_expansion_fixture -- --nocapture`
|
||||
3. `cargo test --test scene_generator_test generator_writes_paginated_enrichment_orderno_expansion_fixture -- --nocapture`
|
||||
4. `cargo test --test scene_generator_p1_family_test -- --nocapture`
|
||||
5. `cargo test --test scene_generator_family_policy_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
`G3` is no longer represented only by one `P0` anchor plus one first expansion fixture.
|
||||
|
||||
It now has a promoted batch-expansion asset that already names the second and third concrete expansion baselines, so the next round can continue from code-backed family assets instead of re-selecting candidates from the ledger snapshot.
|
||||
@@ -0,0 +1,84 @@
|
||||
# G3 First-Round Family Expansion Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-g3-paginated-enrichment-plan.md` Phase 4 / WS5
|
||||
> Family: `G3 / paginated_enrichment`
|
||||
> Result: Completed
|
||||
|
||||
## 1. Round Scope
|
||||
|
||||
本轮 `G3` family 扩展不再停留在单一 `P0` 样板，而是形成了：
|
||||
|
||||
1. Representative baseline
|
||||
- `tests/fixtures/generated_scene/paginated_enrichment`
|
||||
2. First expansion baseline
|
||||
- `tests/fixtures/generated_scene/paginated_enrichment_expansion`
|
||||
|
||||
扩展样板承接的是 `G3` 的第一类变体能力：
|
||||
|
||||
1. 分页字段从 `page/pageSize` 扩展到 `pageNum/pageSize`
|
||||
2. join key 从 `custNo` 扩展到 `ticketNo`
|
||||
3. 聚合字段从 `charge` 扩展到 `riskLevel`
|
||||
|
||||
## 2. Implemented Family Reuse
|
||||
|
||||
为支持第一扩展样板，本轮已经把以下能力沉到 `G3` 家族实现中：
|
||||
|
||||
1. 分页字段识别支持 `pageNum/pageNo`
|
||||
2. `risk*` 过滤表达式进入 `G3` 过滤信号
|
||||
3. `join key` 推导允许扩展样板保留核心业务键，同时不过滤掉分页键
|
||||
4. `aggregate:*` 规则可以从 `row.<field>` 过滤表达式恢复
|
||||
|
||||
## 3. Expansion Validation Result
|
||||
|
||||
扩展样板当前验证结果：
|
||||
|
||||
1. `workflowArchetype = paginated_enrichment`
|
||||
2. `readiness.level = A/B`
|
||||
3. `paginationPlan.pageField = pageNum`
|
||||
4. `joinKeys` 包含 `ticketNo`
|
||||
5. `mergeOrDedupeRules` 包含 `aggregate:riskLevel`
|
||||
6. `exportPlan.entry = exportExcel`
|
||||
7. `g3_main_request_resolved = true`
|
||||
8. `g3_join_key_resolved = true`
|
||||
|
||||
## 4. Family-Level Asset Promotion
|
||||
|
||||
本轮结果已经沉到 family 级资产中：
|
||||
|
||||
1. `tests/fixtures/generated_scene/p1_family_manifest.json`
|
||||
- `G3` 现已记录 `representative + expansion`
|
||||
2. `tests/fixtures/generated_scene/p1_family_results.json`
|
||||
- `G3` 现已记录 `expansionRuns = 1`
|
||||
3. `tests/scene_generator_p1_family_test.rs`
|
||||
- `G3` family test 现已实际运行 expansion fixture
|
||||
|
||||
## 5. Regression Record
|
||||
|
||||
本轮已通过：
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_p1_family_test -- --nocapture
|
||||
cargo test --test scene_generator_family_policy_test -- --nocapture
|
||||
cargo test --test scene_generator_test -- --nocapture
|
||||
cargo test --test scene_generator_canonical_test -- --nocapture
|
||||
```
|
||||
|
||||
## 6. Current Family Conclusion
|
||||
|
||||
当前可以确认：
|
||||
|
||||
1. `G3` 已不再只有单个 `P0` 样板
|
||||
2. `G3` 已拥有第一扩展样板和 family 级断言
|
||||
3. representative 与 expansion 已共同进入 repo-local baseline
|
||||
4. `G3` 第一轮 family expansion 已达成
|
||||
|
||||
## 7. Remaining Space
|
||||
|
||||
本轮完成的是 `G3` 第一扩展样板，不代表 `G3` 全部复杂工单家族都已经覆盖。
|
||||
|
||||
后续如果继续扩展，优先方向应为：
|
||||
|
||||
1. 第二个 `G3` 扩展样板
|
||||
2. 更复杂的 host bridge / localhost 依赖场景
|
||||
3. 更强的多源聚合和导出前置动作恢复
|
||||
@@ -0,0 +1,54 @@
|
||||
# G3 Next-Round Expansion Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round continues consuming the `95598-ticket-family-candidate` downstream queue instead of stopping at the second-round expansion baselines.
|
||||
|
||||
Added fixtures:
|
||||
|
||||
1. `tests/fixtures/generated_scene/paginated_enrichment_expansion_source_distribution`
|
||||
2. `tests/fixtures/generated_scene/paginated_enrichment_expansion_service_risk`
|
||||
|
||||
## Contract Variants Added
|
||||
|
||||
The next round validates two more `G3` reusable variants:
|
||||
|
||||
1. `pageNum + ticketNo + aggregate:sourceType`
|
||||
2. `pageNo + ticketNo + aggregate:riskLevel`
|
||||
|
||||
## Delivered
|
||||
|
||||
Updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/p1_family_manifest.json`
|
||||
2. `tests/fixtures/generated_scene/p1_family_results.json`
|
||||
3. `tests/fixtures/generated_scene/g3_candidate_batch_2026-04-18.json`
|
||||
4. `tests/g3_candidate_batch_test.rs`
|
||||
5. `tests/scene_generator_test.rs`
|
||||
6. `tests/scene_generator_p1_family_test.rs`
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
1. deterministic analysis for both new fixtures
|
||||
2. package generation for both new fixtures
|
||||
3. family-level reusable migration assertions for both new fixtures
|
||||
4. batch-asset assertions for the promoted fourth and fifth expansion anchors
|
||||
|
||||
## Outcome
|
||||
|
||||
`G3` now has:
|
||||
|
||||
1. one `P0` anchor
|
||||
2. one first expansion fixture
|
||||
3. two second-round expansion fixtures
|
||||
4. two next-round promoted expansion fixtures
|
||||
|
||||
The current `paginated_enrichment` family baseline is now validated across:
|
||||
|
||||
1. `ticketNo`
|
||||
2. `workOrderNo`
|
||||
3. `orderNo`
|
||||
|
||||
and across both `riskLevel` and `sourceType` aggregate families with supported pagination-field variants.
|
||||
103
docs/superpowers/reports/2026-04-18-g3-p0-validation-report.md
Normal file
103
docs/superpowers/reports/2026-04-18-g3-p0-validation-report.md
Normal file
@@ -0,0 +1,103 @@
|
||||
# G3 P0 Validation Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-g3-paginated-enrichment-plan.md` Phase 3 -> Phase 4
|
||||
> Archetype: `G3 / paginated_enrichment`
|
||||
> Result: Passed
|
||||
|
||||
## 1. Conclusion
|
||||
|
||||
`G3` 的 `P0-3` 基线已经形成正式闭环。
|
||||
|
||||
当前 repo-local `P0` 样板 `tests/fixtures/generated_scene/paginated_enrichment` 已经稳定恢复并通过以下关键合同：
|
||||
|
||||
1. `main_request`
|
||||
2. `pagination_plan`
|
||||
3. `enrichment_requests`
|
||||
4. `join_keys`
|
||||
5. `merge_or_dedupe_rules`
|
||||
6. `export_plan`
|
||||
|
||||
同时，`G3` 的 `fail-closed` 语义已经固定，至少覆盖：
|
||||
|
||||
1. `join_key_missing`
|
||||
2. `pagination_incomplete`
|
||||
3. `export_only_without_business_chain`
|
||||
|
||||
## 2. Validation Inputs
|
||||
|
||||
本轮 repo-local `P0` 校验基于以下资产：
|
||||
|
||||
1. Fixture:
|
||||
- `tests/fixtures/generated_scene/paginated_enrichment/index.html`
|
||||
2. Canonical:
|
||||
- `tests/fixtures/generated_scene/p0_canonical_answers/p0-3-paginated-enrichment.scene-ir.json`
|
||||
3. Tests:
|
||||
- `tests/scene_generator_canonical_test.rs`
|
||||
- `tests/scene_generator_test.rs`
|
||||
|
||||
关键回归命令：
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_canonical_test -- --nocapture
|
||||
cargo test --test scene_generator_test -- --nocapture
|
||||
```
|
||||
|
||||
## 3. P0 Restored Semantics
|
||||
|
||||
当前 `P0-3` 样板已经稳定恢复：
|
||||
|
||||
1. Main request
|
||||
- endpoint url: `http://yx.gs.sgcc.com.cn/marketing/userList`
|
||||
- response path: `rows`
|
||||
2. Pagination plan
|
||||
- page field: `page`
|
||||
- page size field: `pageSize`
|
||||
- termination rule: `stop_when_page_rows_empty`
|
||||
3. Enrichment request
|
||||
- endpoint url: `http://yx.gs.sgcc.com.cn/marketing/userCharges`
|
||||
- consumed field: `custNo`
|
||||
4. Join key
|
||||
- `custNo`
|
||||
5. Merge / dedupe rules
|
||||
- `dedupe:custNo`
|
||||
- `aggregate:charge`
|
||||
6. Export plan
|
||||
- entry: `exportExcel`
|
||||
|
||||
## 4. Acceptance Checklist Result
|
||||
|
||||
对照 `WS4 / WS5` 的 `P0` 检查项，当前结果如下：
|
||||
|
||||
1. `workflowArchetype = paginated_enrichment`
|
||||
2. `mainRequest.responsePath = rows`
|
||||
3. `paginationPlan.pageField = page`
|
||||
4. `joinKeys = [custNo]`
|
||||
5. `mergeOrDedupeRules` 包含 `aggregate:charge`
|
||||
6. `exportPlan.entry = exportExcel`
|
||||
7. `g3_join_key_resolved = true`
|
||||
8. `localhost_dependency_candidate` 未污染主业务链
|
||||
9. `readiness.level = A`
|
||||
|
||||
## 5. Failure Taxonomy Baseline
|
||||
|
||||
本轮已被测试钉住的 `G3` 失败 taxonomy：
|
||||
|
||||
1. `join_key_missing`
|
||||
2. `pagination_incomplete`
|
||||
3. `export_only_without_business_chain`
|
||||
|
||||
对应测试：
|
||||
|
||||
1. `paginated_enrichment_readiness_marks_join_key_missing_taxonomy`
|
||||
2. `paginated_enrichment_readiness_marks_pagination_incomplete_taxonomy`
|
||||
3. `paginated_enrichment_readiness_marks_export_only_without_business_chain_taxonomy`
|
||||
|
||||
## 6. Final Status
|
||||
|
||||
对照 `2026-04-18-g3-paginated-enrichment-plan.md`：
|
||||
|
||||
1. `P0 canonical` 已冻结
|
||||
2. `P0 acceptance checklist` 已转化为可执行测试
|
||||
3. `P0 failure taxonomy` 已形成最小稳定集合
|
||||
4. `P0-3` 当前可作为 `G3` 后续 family 扩展的统一校准源
|
||||
@@ -0,0 +1,50 @@
|
||||
# G3 Second-Round Expansion Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round takes `G3` beyond the first expansion fixture and lands two additional repo-local expansion baselines from the `95598/工单家族候选` batch direction.
|
||||
|
||||
Added fixtures:
|
||||
|
||||
1. `tests/fixtures/generated_scene/paginated_enrichment_expansion_workorder`
|
||||
2. `tests/fixtures/generated_scene/paginated_enrichment_expansion_orderno`
|
||||
|
||||
## Contract Variants Added
|
||||
|
||||
The second round explicitly validates two more `G3` join-key and pagination variants:
|
||||
|
||||
1. `pageNo + workOrderNo + aggregate:sourceType`
|
||||
2. `page + orderNo + aggregate:sourceType`
|
||||
|
||||
## Delivered
|
||||
|
||||
Updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/p1_family_manifest.json`
|
||||
2. `tests/fixtures/generated_scene/p1_family_results.json`
|
||||
3. `tests/scene_generator_p1_family_test.rs`
|
||||
4. `tests/scene_generator_test.rs`
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
1. deterministic analysis for both new fixtures
|
||||
2. package generation for both new fixtures
|
||||
3. family-level reusable migration assertions for both new fixtures
|
||||
|
||||
## Outcome
|
||||
|
||||
`G3` now has:
|
||||
|
||||
1. one P0 anchor
|
||||
2. one first expansion fixture
|
||||
3. two second-round expansion fixtures
|
||||
|
||||
So the family baseline now proves that the current `paginated_enrichment` contract is reusable across:
|
||||
|
||||
1. `ticketNo`
|
||||
2. `workOrderNo`
|
||||
3. `orderNo`
|
||||
|
||||
with corresponding pagination-field variants.
|
||||
@@ -0,0 +1,28 @@
|
||||
# G3 Timeout Warning Expansion Report
|
||||
|
||||
> Status: Completed
|
||||
> Date: 2026-04-18
|
||||
> Plan: [2026-04-17-scene-skill-60-to-90-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-17-scene-skill-60-to-90-roadmap-plan.md)
|
||||
|
||||
## Intent
|
||||
|
||||
Promote one more `G3` downstream candidate within the existing paginated-enrichment family contract, without introducing a new archetype or runtime scope.
|
||||
|
||||
## Delivered
|
||||
|
||||
1. Added `tests/fixtures/generated_scene/paginated_enrichment_expansion_timeout_warning`
|
||||
2. Added deterministic and generation regressions for the timeout-warning sample
|
||||
3. Promoted `ticket_timeout_warning_detail` into the formal `G3` batch expansion baseline
|
||||
4. Synchronized family manifest, family results, ledger overlay, and roadmap execution status
|
||||
|
||||
## Acceptance
|
||||
|
||||
1. archetype remains `paginated_enrichment`
|
||||
2. pagination field is recovered as `pageNum`
|
||||
3. join key remains `ticketNo`
|
||||
4. aggregate rule remains `aggregate:riskLevel`
|
||||
5. `Track C` remaining queue drops from six to five
|
||||
|
||||
## Notes
|
||||
|
||||
This promotion stays inside the roadmap boundary because it reuses the existing `G3` family contract and only consumes one downstream candidate from the current batch asset.
|
||||
@@ -0,0 +1,61 @@
|
||||
# G6 G7 G8 Family Asset Sync Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round does not change `G6/G7/G8` runtime contracts themselves.
|
||||
|
||||
It promotes the three boundary families into formal family-level assets so they are no longer only protected by implementation tests.
|
||||
|
||||
## Synced Assets
|
||||
|
||||
Updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/p1_family_manifest.json`
|
||||
2. `tests/fixtures/generated_scene/p1_family_results.json`
|
||||
3. `tests/fixtures/generated_scene/family_expansion_policy.json`
|
||||
4. `tests/scene_generator_p1_family_test.rs`
|
||||
5. `tests/scene_generator_family_policy_test.rs`
|
||||
|
||||
## Result
|
||||
|
||||
`G6`, `G7`, and `G8` are now represented at the same family-asset layer as the existing mainline families.
|
||||
|
||||
That means the repo now preserves these facts as first-class assets:
|
||||
|
||||
1. `G6` is an independent `host_bridge_workflow` family
|
||||
2. `G7` is an independent `multi_endpoint_inventory` family
|
||||
3. `G8` is an independent `local_doc_pipeline` family
|
||||
4. none of them should fall back into `G1`, `G1-E`, or `G3`
|
||||
|
||||
## Policy Position
|
||||
|
||||
The policy asset now separates:
|
||||
|
||||
1. mainline groups: `G1/G2/G3`
|
||||
2. boundary-runtime groups: `G6/G7/G8`
|
||||
3. deferred groups: `G4/G5`
|
||||
|
||||
This matches the current roadmap state:
|
||||
|
||||
1. mainline capability growth still centers on `G1/G2/G3`
|
||||
2. `G6/G7/G8` are no longer just reassignment notes; they now have code-backed runtime contracts and family-level preservation
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
- `cargo test --test scene_generator_p1_family_test -- --nocapture`
|
||||
- `cargo test --test scene_generator_test -- --nocapture`
|
||||
- `cargo test --test scene_generator_family_policy_test -- --nocapture`
|
||||
- `cargo test --test scene_generator_canonical_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
The roadmap has moved past pure boundary reassignment.
|
||||
|
||||
`G6/G7/G8` now exist as synchronized family assets with:
|
||||
|
||||
1. representative migration baselines
|
||||
2. results records
|
||||
3. policy placement
|
||||
4. regression protection
|
||||
@@ -0,0 +1,61 @@
|
||||
# G6 Host Bridge First Slice Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `G6` classification and fail-closed safety slice
|
||||
|
||||
## Summary
|
||||
|
||||
The first `G6` implementation slice is complete.
|
||||
|
||||
This round did not attempt runnable host bridge execution. It implemented the safer first step:
|
||||
|
||||
1. new `host_bridge_workflow` archetype
|
||||
2. deterministic host bridge action detection
|
||||
3. `G6` classification priority over `G1-E`
|
||||
4. `G6` readiness gates
|
||||
5. fail-closed generation behavior
|
||||
6. regression protection so ordinary localhost export noise does not become `G6`
|
||||
|
||||
## Assets Updated
|
||||
|
||||
1. `src/generated_scene/ir.rs`
|
||||
2. `src/generated_scene/analyzer.rs`
|
||||
3. `src/generated_scene/generator.rs`
|
||||
4. `tests/fixtures/generated_scene/g6_host_bridge_workflow/index.html`
|
||||
5. `tests/scene_generator_test.rs`
|
||||
|
||||
## Behavioral Result
|
||||
|
||||
The new representative fixture validates:
|
||||
|
||||
1. `workflowArchetype = host_bridge_workflow`
|
||||
2. host bridge action evidence includes `sgBrowerserJsAjax2`
|
||||
3. host runtime evidence includes `localhost:13313`
|
||||
4. business callbacks include `getWorkOrderToDoList`
|
||||
5. business callbacks include `queryMeterPlanFormulateApp`
|
||||
6. generation fails closed with `host_bridge_workflow`
|
||||
|
||||
## Boundary Correction
|
||||
|
||||
The implementation initially exposed one boundary risk: plain `localhost:*` noise could promote a scene to `G6`.
|
||||
|
||||
The classifier was corrected so `G6` requires explicit host bridge action evidence. `localhost:*` alone remains host-runtime evidence and does not define the archetype.
|
||||
|
||||
## Validation
|
||||
|
||||
Executed:
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test -- --nocapture
|
||||
cargo test --test scene_generator_p1_family_test -- --nocapture
|
||||
cargo test --test scene_generator_family_policy_test -- --nocapture
|
||||
cargo test --test scene_generator_canonical_test -- --nocapture
|
||||
```
|
||||
|
||||
All target regressions passed.
|
||||
|
||||
## Result
|
||||
|
||||
`G6` is now safely separated from `G1/G1-E` at classification time.
|
||||
|
||||
The family is intentionally still fail-closed until a real host bridge runtime contract is designed and implemented.
|
||||
@@ -0,0 +1,52 @@
|
||||
# G6 Host Bridge Runtime Contract Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `G6` minimum runtime contract implementation
|
||||
|
||||
## Summary
|
||||
|
||||
`G6` has moved from pure fail-closed classification to a minimum runnable runtime contract.
|
||||
|
||||
The implementation does not redesign the browser host transport. Instead it generates a conservative `host_bridge_workflow` script that:
|
||||
|
||||
1. preserves host bridge actions
|
||||
2. preserves localhost host-runtime dependencies
|
||||
3. lists callback business endpoints
|
||||
4. invokes host bridge APIs when available
|
||||
5. invokes callback endpoints through the existing browser request path
|
||||
6. returns blocked/error/partial/ok artifact statuses instead of pretending full equivalence
|
||||
|
||||
## Assets Updated
|
||||
|
||||
1. `src/generated_scene/generator.rs`
|
||||
2. `tests/scene_generator_test.rs`
|
||||
|
||||
## Contract
|
||||
|
||||
The minimal `G6` contract is now:
|
||||
|
||||
1. at least one `host_bridge` workflow step
|
||||
2. at least one `callback_request` workflow step
|
||||
3. at least one business endpoint
|
||||
|
||||
If that contract is missing, generation still fails closed.
|
||||
|
||||
## Validation
|
||||
|
||||
Executed:
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test g6 -- --nocapture
|
||||
cargo test --test scene_generator_test -- --nocapture
|
||||
cargo test --test scene_generator_p1_family_test -- --nocapture
|
||||
cargo test --test scene_generator_family_policy_test -- --nocapture
|
||||
cargo test --test scene_generator_canonical_test -- --nocapture
|
||||
```
|
||||
|
||||
All target regressions passed.
|
||||
|
||||
## Result
|
||||
|
||||
`G6` now has a minimum generated package path while retaining a fail-closed negative test for incomplete contracts.
|
||||
|
||||
The next runtime-contract target is `G7 multi_endpoint_inventory`.
|
||||
@@ -0,0 +1,49 @@
|
||||
# G7 Multi Endpoint Inventory First Slice Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `G7` classification and fail-closed safety slice
|
||||
|
||||
## Summary
|
||||
|
||||
The first `G7` implementation slice is complete.
|
||||
|
||||
This round added the `multi_endpoint_inventory` archetype and a repo-local representative fixture for the `计量资产库存统计` boundary family.
|
||||
|
||||
The result is intentionally fail-closed: `G7` can now be identified and prevented from falling back into `G1/G1-E`, but runnable aggregation is not implemented yet.
|
||||
|
||||
## Assets Updated
|
||||
|
||||
1. `src/generated_scene/ir.rs`
|
||||
2. `src/generated_scene/analyzer.rs`
|
||||
3. `src/generated_scene/generator.rs`
|
||||
4. `tests/fixtures/generated_scene/g7_multi_endpoint_inventory/index.html`
|
||||
5. `tests/scene_generator_test.rs`
|
||||
|
||||
## Behavioral Result
|
||||
|
||||
The representative fixture validates:
|
||||
|
||||
1. `workflowArchetype = multi_endpoint_inventory`
|
||||
2. inventory endpoint count is at least five
|
||||
3. inventory endpoint names include `assetStatsQueryMeter`
|
||||
4. inventory endpoint names include `assetStatsQueryJlGnModule`
|
||||
5. generation fails closed with `multi_endpoint_inventory`
|
||||
|
||||
## Validation
|
||||
|
||||
Executed:
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test -- --nocapture
|
||||
cargo test --test scene_generator_p1_family_test -- --nocapture
|
||||
cargo test --test scene_generator_family_policy_test -- --nocapture
|
||||
cargo test --test scene_generator_canonical_test -- --nocapture
|
||||
```
|
||||
|
||||
All target regressions passed.
|
||||
|
||||
## Result
|
||||
|
||||
`G7` is now safely separated from `G1/G1-E` at classification time.
|
||||
|
||||
The family remains fail-closed until a formal multi-endpoint inventory aggregation contract is implemented.
|
||||
@@ -0,0 +1,46 @@
|
||||
# G7 Multi-Endpoint Inventory Runtime Contract Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round upgrades `G7 / multi_endpoint_inventory` from safe-slice classification to a minimal runnable runtime contract.
|
||||
|
||||
Representative fixture:
|
||||
|
||||
- `tests/fixtures/generated_scene/g7_multi_endpoint_inventory`
|
||||
|
||||
## Delivered
|
||||
|
||||
1. Added a minimal `G7` runtime contract in the generator.
|
||||
2. Promoted `g7_fail_closed` from unconditional block to conditional pass when the inventory contract is complete.
|
||||
3. Added a dedicated browser compiler path for `multi_endpoint_inventory`.
|
||||
4. Replaced the old positive fail-closed test with:
|
||||
- runnable generation test
|
||||
- incomplete-contract negative test
|
||||
|
||||
## Minimal Contract
|
||||
|
||||
`G7` is treated as complete when all of the following are present:
|
||||
|
||||
1. At least 3 `inventory_request` workflow steps
|
||||
2. At least 3 API endpoints
|
||||
3. One `inventory_aggregate` workflow step
|
||||
|
||||
## Generated Runtime Shape
|
||||
|
||||
The generated browser script now includes:
|
||||
|
||||
1. `inventoryEndpoints()`
|
||||
2. `aggregateEntry()`
|
||||
3. multi-endpoint request execution
|
||||
4. merged artifact payload with `workflow_archetype = multi_endpoint_inventory`
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
- `cargo test --test scene_generator_test g7 -- --nocapture`
|
||||
- `cargo test --test scene_generator_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
`G7` no longer stops at classification-only safe slice. It now supports a minimal runnable contract while still fail-closing incomplete manual `SceneIr` inputs.
|
||||
@@ -0,0 +1,59 @@
|
||||
# G8 Local Document Pipeline First Slice Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `G8` classification and fail-closed safety slice
|
||||
|
||||
## Summary
|
||||
|
||||
The first `G8` implementation slice is complete.
|
||||
|
||||
This round added the `local_doc_pipeline` archetype and a repo-local representative fixture for the `95598供电服务月报` boundary family.
|
||||
|
||||
The result is intentionally fail-closed: `G8` can now be identified and prevented from falling back into `G1/G1-E/G6`, but runnable local storage, SQL analysis, and document-generation orchestration is not implemented yet.
|
||||
|
||||
## Assets Updated
|
||||
|
||||
1. `src/generated_scene/ir.rs`
|
||||
2. `src/generated_scene/analyzer.rs`
|
||||
3. `src/generated_scene/generator.rs`
|
||||
4. `tests/fixtures/generated_scene/g8_local_doc_pipeline/index.html`
|
||||
5. `tests/scene_generator_test.rs`
|
||||
|
||||
## Behavioral Result
|
||||
|
||||
The representative fixture validates:
|
||||
|
||||
1. `workflowArchetype = local_doc_pipeline`
|
||||
2. localhost dependency evidence includes `localhost:13313`
|
||||
3. local pipeline actions include `definedSqlQuery`
|
||||
4. local pipeline actions include `docExport`
|
||||
5. local pipeline actions include `selectData`
|
||||
6. generation fails closed with `local_doc_pipeline`
|
||||
|
||||
## Boundary Correction
|
||||
|
||||
The first test run showed that `G8` could be captured by `G6` because the fixture also contains host bridge actions.
|
||||
|
||||
The classifier now prioritizes local document pipeline evidence over generic host bridge evidence. This preserves the intended family boundary:
|
||||
|
||||
1. `G6`: host bridge driven query progression
|
||||
2. `G8`: local persistence, SQL analysis, and document production after data capture
|
||||
|
||||
## Validation
|
||||
|
||||
Executed:
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test -- --nocapture
|
||||
cargo test --test scene_generator_p1_family_test -- --nocapture
|
||||
cargo test --test scene_generator_family_policy_test -- --nocapture
|
||||
cargo test --test scene_generator_canonical_test -- --nocapture
|
||||
```
|
||||
|
||||
All target regressions passed.
|
||||
|
||||
## Result
|
||||
|
||||
`G8` is now safely separated from `G1/G1-E/G6` at classification time.
|
||||
|
||||
The family remains fail-closed until a formal local document pipeline runtime contract is implemented.
|
||||
@@ -0,0 +1,49 @@
|
||||
# G8 Local Doc Pipeline Runtime Contract Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round upgrades `G8 / local_doc_pipeline` from safe-slice classification to a minimal runnable runtime contract.
|
||||
|
||||
Representative fixture:
|
||||
|
||||
- `tests/fixtures/generated_scene/g8_local_doc_pipeline`
|
||||
|
||||
## Delivered
|
||||
|
||||
1. Added a minimal `G8` runtime contract in the generator.
|
||||
2. Promoted `g8_fail_closed` from unconditional block to conditional pass when the local document pipeline contract is complete.
|
||||
3. Added a dedicated browser compiler path for `local_doc_pipeline`.
|
||||
4. Extended automatic scene recovery so localhost pipeline dependencies become formal `api_endpoints`.
|
||||
5. Replaced the old positive fail-closed test with:
|
||||
- runnable generation test
|
||||
- incomplete-contract negative test
|
||||
|
||||
## Minimal Contract
|
||||
|
||||
`G8` is treated as complete when all of the following are present:
|
||||
|
||||
1. At least one `local_doc_pipeline` step
|
||||
2. At least one `sql_query` or `doc_query` step
|
||||
3. At least one `doc_export` step
|
||||
4. At least one localhost API endpoint
|
||||
|
||||
## Generated Runtime Shape
|
||||
|
||||
The generated browser script now includes:
|
||||
|
||||
1. `localDataEndpoints()`
|
||||
2. `sqlQueryEntry()`
|
||||
3. `docExportEntry()`
|
||||
4. local-stage request execution
|
||||
5. artifact payload with `workflow_archetype = local_doc_pipeline`
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
- `cargo test --test scene_generator_test g8 -- --nocapture`
|
||||
- `cargo test --test scene_generator_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
`G8` no longer remains classification-only. It now supports a minimal runtime contract while preserving fail-closed behavior for incomplete manual `SceneIr` inputs.
|
||||
@@ -0,0 +1,154 @@
|
||||
# 线损家族新增变体最小合约表
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-lineloss-family-variant-expansion-plan.md`
|
||||
> Purpose: 完成 Phase 2 交付物，明确新增线损变体的最小可用口径与候选验证门槛。
|
||||
|
||||
## 1. 设计原则
|
||||
|
||||
本表只定义“最低可生成”的合约，不直接承诺完整业务语义重建。原则如下：
|
||||
|
||||
1. 主合约只覆盖能稳定抽取的主链路。
|
||||
2. 页面内计算、拼接、下钻逻辑如果尚不能稳态恢复，先作为扩展证据，不强行塞入主合约。
|
||||
3. 无法满足主合约的样本继续 `fail-close`。
|
||||
|
||||
## 2. 合约表
|
||||
|
||||
### 2.1 `G2-D 预测计算型`
|
||||
|
||||
代表样本：
|
||||
|
||||
1. `线路月度高负损预测`
|
||||
2. `台区线损台区月度高负损预测`
|
||||
|
||||
最小主合约：
|
||||
|
||||
1. `bootstrap`
|
||||
- 必须能定位到线损业务域 `gsllys`
|
||||
- 不要求额外外部系统联动
|
||||
2. `primary evidence`
|
||||
- 存在预测型标识，如 `reportType` 中含预测语义
|
||||
- 页面内存在 `lineLossRate / powerLoss / lineLossType` 等计算结果字段
|
||||
3. `request contract`
|
||||
- 至少识别一个主查询入口或主输入数据源
|
||||
- 若原页面由文件输入或多个表拼装触发，可先将其视为“single primary collection + local compute”
|
||||
4. `response / column contract`
|
||||
- 至少恢复以下列中的 3 个以上：
|
||||
- `lineId`
|
||||
- `lineName`
|
||||
- `lineLossType`
|
||||
- `lineLossRate`
|
||||
- `powerLoss`
|
||||
- 必须恢复 `lineLossRate` 或 `powerLoss` 之一
|
||||
5. `normalize contract`
|
||||
- 对计算结果过滤空值
|
||||
- 保证每条记录至少保留主键列和一个核心预测结果列
|
||||
|
||||
候选验证门槛：
|
||||
|
||||
1. 能稳定落到 `G2-D`
|
||||
2. 生成结果中存在预测结果核心列
|
||||
3. 不把这类场景误判为 `G2-A/G2-B/G2-C`
|
||||
|
||||
扩展证据，不计入主合约：
|
||||
|
||||
1. 复杂白名单逻辑
|
||||
2. 多表 Excel 预处理
|
||||
3. 全量预测规则完全复刻
|
||||
|
||||
### 2.2 `G2-E 核查对比型`
|
||||
|
||||
代表样本：
|
||||
|
||||
1. `台区零度户月度用电量与台区线损电量对比核查报表`
|
||||
|
||||
最小主合约：
|
||||
|
||||
1. `bootstrap`
|
||||
- 必须定位到线损主域 `gsllys`
|
||||
2. `primary evidence`
|
||||
- 同时出现 `getUserElectricList` 与 `getTqLinelossInfoListRank`
|
||||
- 页面内存在按 `TG_NO` 或 `CONS_NO` 的拼接逻辑
|
||||
3. `request contract`
|
||||
- 至少恢复主排名接口和用户电量接口
|
||||
- 主接口保留分页参数 `page / rows`
|
||||
4. `response / column contract`
|
||||
- 至少恢复以下列中的 4 个以上：
|
||||
- `TG_NO`
|
||||
- `TG_NAME`
|
||||
- `consno`
|
||||
- `userNmae`
|
||||
- `thisMonth`
|
||||
- `beforeMonth1`
|
||||
- 必须同时具备一个台区维度列和一个用户维度列
|
||||
5. `normalize contract`
|
||||
- 允许先以主链路行为单位输出
|
||||
- 允许暂不完整复刻所有拼接字段，但必须能看出“台区 + 用户”的双层语义
|
||||
|
||||
候选验证门槛：
|
||||
|
||||
1. 能稳定识别为核查对比型，而不是简单周报或双模式报表
|
||||
2. 生成结果中保留 `TG_NO` 且保留至少一个用户维度字段
|
||||
3. 关联补查证据写入 generation report
|
||||
|
||||
扩展证据，不计入主合约：
|
||||
|
||||
1. 多月历史字段完全展开
|
||||
2. 所有异常兜底行填充逻辑
|
||||
3. 报表导出标题文案完全一致
|
||||
|
||||
### 2.3 `G2-F 异常诊断 / 下钻型`
|
||||
|
||||
代表样本：
|
||||
|
||||
1. `线损大数据-窃电分析`
|
||||
|
||||
最小主合约：
|
||||
|
||||
1. `bootstrap`
|
||||
- 必须定位到线损主域 `gsllys`
|
||||
2. `primary evidence`
|
||||
- 出现主筛选接口 `getTqLinelossInfoListRank`
|
||||
- 出现至少一个诊断或下钻接口：
|
||||
- `tqAutoDiagnoseAnalyse/search`
|
||||
- `stealElecAnalyse/getFlqdyhDetailList`
|
||||
- `stealElecAnalyse/userVoltsAndElecflowMoniter/search`
|
||||
3. `request contract`
|
||||
- 至少恢复主筛选请求
|
||||
- 至少记录存在一个后续诊断请求
|
||||
4. `response / column contract`
|
||||
- 至少恢复以下列中的 4 个以上：
|
||||
- `TG_NO`
|
||||
- `LL_TYPE_NAME`
|
||||
- `LOSS_PQ`
|
||||
- `LINELOSS_RATE`
|
||||
- `remark`
|
||||
- 若暂不能稳定恢复完整 `remark`，也必须显式标记为诊断链路未完备
|
||||
5. `normalize contract`
|
||||
- 主表记录必须可输出
|
||||
- 诊断 remark 允许退化为部分诊断结果或空值，但不能伪造完整语义
|
||||
|
||||
候选验证门槛：
|
||||
|
||||
1. 能稳定识别该类为“主筛选 + 下钻诊断”链路
|
||||
2. generation report 中能写出诊断接口证据
|
||||
3. 若 `remark` 合约不完整，readiness 不能伪装成 A
|
||||
|
||||
扩展证据，不计入主合约：
|
||||
|
||||
1. 所有异常分类规则
|
||||
2. 各类诊断 remark 拼接细节
|
||||
3. 多轮异步诊断时序
|
||||
|
||||
## 3. 本轮实施结论
|
||||
|
||||
按当前最小合约成熟度排序：
|
||||
|
||||
1. `G2-D` 最适合先进入实现
|
||||
2. `G2-E` 适合紧随其后
|
||||
3. `G2-F` 先作为高复杂候选组，只有在主链路实现稳定后再推进
|
||||
|
||||
因此，Phase 3 的实现顺序固定为：
|
||||
|
||||
1. 先扩 `G2-D`
|
||||
2. 再扩 `G2-E`
|
||||
3. `G2-F` 仅保留为候选扩展或回归观察对象
|
||||
@@ -0,0 +1,155 @@
|
||||
# 线损家族变体扩展回归报告
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-lineloss-family-variant-expansion-plan.md`
|
||||
> Result: Completed
|
||||
|
||||
## 1. 本轮执行范围
|
||||
|
||||
本轮严格承接已完成的 `G2-A/G2-B/G2-C` 基线，不回头修改主样本整顿逻辑，只做以下两类新增线损变体扩展：
|
||||
|
||||
1. `G2-E = g2_e_comparison_crosscheck`
|
||||
代表真实样本：`台区零度户月度用电量与台区线损电量对比核查报表`
|
||||
2. `G2-F = g2_f_diagnosis_drilldown`
|
||||
代表真实样本：`线损大数据-窃电分析`
|
||||
|
||||
`G2-D` 预测计算型已完成结构盘点，但因真实样本缺少稳定业务接口，暂不进入本轮生成链路实现。
|
||||
|
||||
## 2. Phase 0 / 1 / 2 交付结果
|
||||
|
||||
已新增文档：
|
||||
|
||||
1. `docs/superpowers/reports/2026-04-18-lineloss-family-variant-grouping-report.md`
|
||||
2. `docs/superpowers/reports/2026-04-18-lineloss-family-minimum-contract-table.md`
|
||||
|
||||
完成内容：
|
||||
|
||||
1. 冻结 `G2-A/G2-B/G2-C` 为线损家族既有基线
|
||||
2. 新增 `G2-D/G2-E/G2-F` 三个扩展方向
|
||||
3. 明确本轮实际实施对象收敛为 `G2-E/G2-F`
|
||||
4. 为新增变体建立最小合约与候选验证门槛
|
||||
|
||||
## 3. Phase 3 实现结果
|
||||
|
||||
### 3.1 新增 fixture
|
||||
|
||||
1. `tests/fixtures/generated_scene/g2_comparison_crosscheck/index.html`
|
||||
2. `tests/fixtures/generated_scene/g2_diagnosis_drilldown/index.html`
|
||||
|
||||
### 3.2 新增分类与生成支持
|
||||
|
||||
已扩展：
|
||||
|
||||
1. `src/generated_scene/analyzer.rs`
|
||||
- 新增 `G2E`
|
||||
- 新增 `G2F`
|
||||
- `G2-E` 判定口径：
|
||||
- 存在 `getTqLinelossInfoListRank`
|
||||
- 存在 `getUserElectricList`
|
||||
- 同时出现用户 / 台区拼接信号
|
||||
- `G2-F` 判定口径：
|
||||
- 存在主筛选接口 `getTqLinelossInfoListRank`
|
||||
- 存在诊断或下钻接口 `tqAutoDiagnoseAnalyse/search` / `stealElecAnalyse/*`
|
||||
2. `src/generated_scene/generator.rs`
|
||||
- 新增 `G2-E` 最小 mode：`comparison`
|
||||
- 新增 `G2-F` 最小 mode：`diagnosis`
|
||||
- readiness gate 已接纳两类新 mode
|
||||
3. `tests/scene_generator_test.rs`
|
||||
- 新增 `G2-E` 分类测试
|
||||
- 新增 `G2-F` 分类测试
|
||||
- 新增 `G2-E` 生成测试
|
||||
- 新增 `G2-F` 生成测试
|
||||
|
||||
### 3.3 本地回归结果
|
||||
|
||||
执行命令：
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test -- --nocapture
|
||||
```
|
||||
|
||||
结果：
|
||||
|
||||
1. `23 passed`
|
||||
2. 原有 `G2-A/G2-B/G2-C` 未回退
|
||||
3. 新增 `G2-E/G2-F` fixture 分类与生成通过
|
||||
|
||||
## 4. Phase 4 真实样本回归结果
|
||||
|
||||
### 4.1 `G2-E` 核查对比型
|
||||
|
||||
真实样本：
|
||||
|
||||
1. `台区零度户月度用电量与台区线损电量对比核查报表`
|
||||
|
||||
生成产物：
|
||||
|
||||
1. `examples/real_scene_batch_round2/skills/real-zero-consumer-crosscheck-r1`
|
||||
|
||||
结果：
|
||||
|
||||
1. `workflowArchetype = multi_mode_request`
|
||||
2. `g2_family_variant_candidate = g2_e_comparison_crosscheck`
|
||||
3. `bootstrap.expectedDomain = 20.76.57.61:18080`
|
||||
4. `bootstrap.targetUrl = http://20.76.57.61:18080/gsllys`
|
||||
5. `defaultMode = comparison`
|
||||
6. `columnDefs` 已恢复 `TG_NO / TG_NAME / consno / userNmae / thisMonth / beforeMonth1`
|
||||
7. `readiness.level = A`
|
||||
|
||||
结论：
|
||||
|
||||
1. 已从“未归类场景”推进到“可进入候选 skill 阶段”
|
||||
2. 当前生成结果保住了台区 + 用户双层语义
|
||||
|
||||
### 4.2 `G2-F` 异常诊断 / 下钻型
|
||||
|
||||
真实样本：
|
||||
|
||||
1. `线损大数据-窃电分析`
|
||||
|
||||
生成产物：
|
||||
|
||||
1. `examples/real_scene_batch_round2/skills/real-steal-analysis-r1`
|
||||
|
||||
结果：
|
||||
|
||||
1. `workflowArchetype = multi_mode_request`
|
||||
2. `g2_family_variant_candidate = g2_f_diagnosis_drilldown`
|
||||
3. `bootstrap.expectedDomain = 20.76.57.61:18080`
|
||||
4. `bootstrap.targetUrl = http://20.76.57.61:18080/gsllys`
|
||||
5. `defaultMode = diagnosis`
|
||||
6. `columnDefs` 已恢复 `TG_NO / LL_TYPE_NAME / LOSS_PQ / LINELOSS_RATE / remark`
|
||||
7. generation report 中保留了 `tqAutoDiagnoseAnalyse/search`、`stealElecAnalyse/getFlqdyhDetailList`、`stealElecAnalyse/userVoltsAndElecflowMoniter/search` 证据
|
||||
8. `readiness.level = A`
|
||||
|
||||
结论：
|
||||
|
||||
1. 已从“高复杂未归类场景”推进到“可候选验证的诊断型变体”
|
||||
2. 当前版本恢复的是主链路最小诊断合约，不承诺完整 remark 业务语义
|
||||
|
||||
## 5. 当前残留问题
|
||||
|
||||
本轮 plan 范围内已达标，但仍有以下残留：
|
||||
|
||||
1. `G2-D` 预测计算型尚未进入实现
|
||||
- 根因：真实样本主要是本地计算与文件输入，缺少稳定业务接口
|
||||
2. 真实样本 HTML 中仍含大量第三方或非目标业务 URL 噪声
|
||||
- 当前不会阻塞 `G2-E/G2-F` 候选生成
|
||||
- 但后续应继续补“目标业务域去噪”能力
|
||||
3. `G2-F` 的完整诊断 remark 仍未重建
|
||||
- 当前只恢复最小诊断合约，不等于业务语义完全还原
|
||||
|
||||
## 6. 计划完成判定
|
||||
|
||||
对照 `2026-04-18-lineloss-family-variant-expansion-plan.md`，本轮已完成：
|
||||
|
||||
1. Phase 0：冻结线损家族现有基线
|
||||
2. Phase 1：完成线损变体分组和代表样本清单
|
||||
3. Phase 2：完成新增变体最小合约表
|
||||
4. Phase 3：按分组扩展 `fixture / test / analyzer / generator / readiness`
|
||||
5. Phase 4：完成两类新增真实样本回归并输出扩展报告
|
||||
|
||||
最终判定：
|
||||
|
||||
1. 至少两类新增线损变体已进入候选验证阶段
|
||||
2. 原有 `G2-A/G2-B/G2-C` 未回退
|
||||
3. `G2-D` 保留为下一阶段候选组
|
||||
@@ -0,0 +1,136 @@
|
||||
# 线损家族变体分组报告
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-lineloss-family-variant-expansion-plan.md`
|
||||
> Purpose: 完成 Phase 0 / Phase 1 交付物，冻结当前线损家族基线，并建立后续扩展分组清单。
|
||||
|
||||
## 1. 当前已冻结基线
|
||||
|
||||
本轮扩展以前，线损家族中已经可稳定识别并生成的子型为：
|
||||
|
||||
1. `G2-A = g2_a_dual_mode_baseline`
|
||||
代表样本：`台区线损大数据-月_周累计线损率统计分析`
|
||||
特征：同一主链路下同时存在月 / 周两套模式，请求和列定义都较完整。
|
||||
2. `G2-B = g2_b_weekly_single_mode`
|
||||
代表样本：`白银线损周报`
|
||||
特征：仍属于线损主链路，但只有周报单侧模式，需要用最小周报合约兜住生成。
|
||||
3. `G2-C = g2_c_mixed_linked_workflow`
|
||||
代表样本：`线损同期差异报表`
|
||||
特征：主链路仍是线损查询，但同时联动外部系统或二次查询，只能先恢复主链路最小合约。
|
||||
|
||||
上述三类已经在上一轮回归中通过本地测试，并在真实样本重生成为候选 skill，不再回退到“只修主样本”的阶段。
|
||||
|
||||
## 2. 线损场景盘点结果
|
||||
|
||||
本次从 `D:\desk\智能体资料\全量业务场景\一平台场景` 中筛出的线损相关场景为：
|
||||
|
||||
1. `台区线损大数据-月_周累计线损率统计分析`
|
||||
2. `白银线损周报`
|
||||
3. `线损同期差异报表`
|
||||
4. `台区零度户月度用电量与台区线损电量对比核查报表`
|
||||
5. `线损大数据-窃电分析`
|
||||
6. `线路月度高负损预测`
|
||||
7. `台区线损台区月度高负损预测`
|
||||
8. `供电所线路电量统计`
|
||||
9. `线路失电信息明细`
|
||||
|
||||
其中 `8` 和 `9` 与当前线损主家族相关度较弱，不纳入本轮优先扩展对象。
|
||||
|
||||
## 3. 结构化分组
|
||||
|
||||
### 3.1 已落地组
|
||||
|
||||
1. `G2-A 双模式主报表型`
|
||||
代表样本：`台区线损大数据-月_周累计线损率统计分析`
|
||||
2. `G2-B 周报单侧模式型`
|
||||
代表样本：`白银线损周报`
|
||||
3. `G2-C 混合联动主链路型`
|
||||
代表样本：`线损同期差异报表`
|
||||
|
||||
### 3.2 新增候选组
|
||||
|
||||
1. `G2-D 预测计算型`
|
||||
代表样本：
|
||||
- `线路月度高负损预测`
|
||||
- `台区线损台区月度高负损预测`
|
||||
结构特征：
|
||||
- 页面内存在 `reportType = monthlyHighNegativeLossPredictionForLines` 或同类预测型标识
|
||||
- 结果并非直接来自标准 `cols1/cols2` 表头，而是先读取多个输入表或数据块，再在页面内计算 `lineLossRate / powerLoss / lineLossType`
|
||||
- 导出字段以“预测结果字段”为主，如 `lineId / lineName / lineLossType / lineLossRate / powerLoss`
|
||||
结论：
|
||||
- 这类不是传统 G2 主报表直接抽列，而是“取数 + 计算 + 导出”的计算型线损变体
|
||||
- 适合优先扩展，因为有两个高度相似样本，可做成同一家族
|
||||
|
||||
2. `G2-E 核查对比型`
|
||||
代表样本：
|
||||
- `台区零度户月度用电量与台区线损电量对比核查报表`
|
||||
结构特征：
|
||||
- 同时调用 `getUserElectricList` 与 `getTqLinelossInfoListRank`
|
||||
- 页面内做数据拼接、按 `TG_NO / CONS_NO` 关联用户和台区数据
|
||||
- 导出字段同时含台区维度和用户维度，如 `TG_NO / TG_NAME / consno / thisMonth / beforeMonth1`
|
||||
结论：
|
||||
- 该组更像“主链路 + 明细补查 + 拼接报表”
|
||||
- 与 `G2-C` 有相似性，但最终产物比 `G2-C` 更偏“核查对比报表”而不是简单主链路恢复
|
||||
|
||||
3. `G2-F 异常诊断 / 下钻型`
|
||||
代表样本：
|
||||
- `线损大数据-窃电分析`
|
||||
结构特征：
|
||||
- 先取 `getTqLinelossInfoListRank`
|
||||
- 再调用 `tqAutoDiagnoseAnalyse/search`、`stealElecAnalyse/getFlqdyhDetailList`、`stealElecAnalyse/userVoltsAndElecflowMoniter/search`
|
||||
- 页面中通过多次诊断接口拼装 `remark`
|
||||
- 导出字段核心列为 `TG_NO / LOSS_PQ / LINELOSS_RATE / remark`
|
||||
结论：
|
||||
- 该组属于“线损主表筛选后进入异常诊断和明细下钻”的复合流程
|
||||
- 二次、三次请求显著多于 `G2-C`，复杂度高于本轮已落地三类
|
||||
|
||||
## 4. 本轮实施映射
|
||||
|
||||
按当前 plan 的 Phase 3 约束，本轮只选择最容易形成“分组复制”的 1 到 2 个新增家族，不做全量铺开。
|
||||
|
||||
优先级排序如下：
|
||||
|
||||
1. `P0: G2-D 预测计算型`
|
||||
原因：
|
||||
- 有两个结构高度相似的真实样本
|
||||
- 页面内计算逻辑清晰，字段也相对稳定
|
||||
- 可以形成新的最小合约和新的 fixture 组
|
||||
2. `P1: G2-E 核查对比型`
|
||||
原因：
|
||||
- 虽然只有一个当前样本，但和现有 `G2-C` 最接近
|
||||
- 适合作为“主链路 + 关联补查”方向的下一扩展
|
||||
3. `P2: G2-F 异常诊断 / 下钻型`
|
||||
原因：
|
||||
- 链路最长，remark 拼接和异常判断逻辑较重
|
||||
- 适合放在本轮后半段，优先做候选验证而不是一步做满
|
||||
|
||||
## 5. Phase 3 前置可行性结论
|
||||
|
||||
在进入实际实现前，对 `G2-D` 两个真实样本再次核查后确认：
|
||||
|
||||
1. `线路月度高负损预测`
|
||||
2. `台区线损台区月度高负损预测`
|
||||
|
||||
这两类页面虽然具备稳定的预测结果字段和 `reportType`，但当前源码里几乎只体现为：
|
||||
|
||||
1. 本地读表或页面内计算
|
||||
2. `localhost:13313` 导出与报表日志依赖
|
||||
3. 没有足够稳定的线损业务接口可抽成当前生成器的请求链路
|
||||
|
||||
因此，本轮实现不强行把 `G2-D` 塞进现有请求驱动编译链，而是将其保留为“候选组”，等待后续补齐“本地计算 / 文件输入型 skill”能力后再做。
|
||||
|
||||
据此，本轮实际实施顺序调整为：
|
||||
|
||||
1. 先实现 `G2-E`
|
||||
2. 再实现 `G2-F`
|
||||
3. `G2-D` 暂停在候选验证阶段
|
||||
|
||||
## 6. Phase 1 退出结论
|
||||
|
||||
本轮线损扩展已经从“按单场景修补”切换到“按家族分组推进”：
|
||||
|
||||
1. 已冻结 `G2-A/G2-B/G2-C` 基线，不再回头重做主样本修补。
|
||||
2. 已识别出后续新增的三个线损变体方向：`G2-D/G2-E/G2-F`。
|
||||
3. 经可行性核查后，本轮实际实现顺序固定为：
|
||||
- 先做 `G2-E`
|
||||
- 再做 `G2-F`
|
||||
- `G2-D` 保留为下一阶段候选组
|
||||
@@ -0,0 +1,82 @@
|
||||
# Post-Roadmap Execution Closure Report
|
||||
|
||||
> Date: 2026-04-18
|
||||
> Scope: `docs/superpowers/plans/2026-04-18-scene-skill-post-roadmap-execution-plan.md`
|
||||
> Result: Completed
|
||||
|
||||
## 1. Conclusion
|
||||
|
||||
The post-roadmap execution plan has been closed within its own boundary.
|
||||
|
||||
The completed outputs now exist for all required phases:
|
||||
|
||||
1. Phase 0: handover boundary asset
|
||||
2. Phase 1: minimum current execution board
|
||||
3. Phase 2: real-sample validation layer
|
||||
4. Phase 3: boundary and runtime entry rules
|
||||
5. Phase 4: bounded next-roadmap design and plan
|
||||
|
||||
## 2. Delivered Assets
|
||||
|
||||
### Phase 0
|
||||
|
||||
1. `tests/fixtures/generated_scene/post_roadmap_handover_2026-04-18.json`
|
||||
|
||||
### Phase 1
|
||||
|
||||
1. `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
|
||||
### Phase 2
|
||||
|
||||
1. `tests/fixtures/generated_scene/real_sample_validation_plan_2026-04-18.json`
|
||||
2. `tests/fixtures/generated_scene/real_sample_validation_record_template_2026-04-18.json`
|
||||
3. `tests/fixtures/generated_scene/real_sample_validation_records_2026-04-18.json`
|
||||
4. `tests/fixtures/generated_scene/real_sample_mismatch_taxonomy_2026-04-18.json`
|
||||
|
||||
### Phase 3
|
||||
|
||||
1. `tests/fixtures/generated_scene/boundary_runtime_entry_rules_2026-04-18.json`
|
||||
|
||||
### Phase 4
|
||||
|
||||
1. `docs/superpowers/specs/2026-04-18-scene-skill-real-sample-validation-roadmap-design.md`
|
||||
2. `docs/superpowers/plans/2026-04-18-scene-skill-real-sample-validation-roadmap-plan.md`
|
||||
|
||||
## 3. Key Outcome
|
||||
|
||||
The most important change is not a new compiler feature.
|
||||
|
||||
It is the separation of:
|
||||
|
||||
1. frozen workbook snapshot
|
||||
2. current execution-board state
|
||||
3. real-sample validation state
|
||||
4. next-roadmap scope decisions
|
||||
|
||||
That separation prevents the next execution round from drifting back into asset-only refinement.
|
||||
|
||||
## 4. Real-Sample Validation Status
|
||||
|
||||
The first-round validation layer now records:
|
||||
|
||||
1. `G2` as an executed mismatch anchor
|
||||
2. `G1-E` as an executed pass anchor
|
||||
3. `G3` as a selected real-sample anchor that still needs an executed real run
|
||||
|
||||
This means the next roadmap can be validation-first rather than fixture-first.
|
||||
|
||||
## 5. Boundary Control
|
||||
|
||||
`G6/G7/G8` remain boundary families.
|
||||
|
||||
`G4/G5` remain deferred or degraded.
|
||||
|
||||
No new implementation scope was opened inside this plan.
|
||||
|
||||
## 6. Verification
|
||||
|
||||
The post-roadmap execution assets are now covered by:
|
||||
|
||||
1. `cargo test --test post_roadmap_execution_assets_test -- --nocapture`
|
||||
|
||||
This verification is in addition to the already existing family, ledger, and roadmap asset tests.
|
||||
@@ -0,0 +1,217 @@
|
||||
# R1 真实样本分析：台区线损大数据-月_周累计线损率统计分析
|
||||
|
||||
日期：2026-04-18
|
||||
|
||||
样本来源：`D:\desk\智能体资料\全量业务场景\一平台场景\台区线损大数据-月_周累计线损率统计分析`
|
||||
|
||||
生成输出：`examples/real_scene_batch_round1/skills/real-tq-lineloss-report-r1`
|
||||
|
||||
对标基线：
|
||||
|
||||
1. `tq-lineloss-report`
|
||||
2. `P0-1 multi_mode_request.month_week_table`
|
||||
3. `docs/superpowers/reports/2026-04-17-integration-test-report.md`
|
||||
|
||||
## 1. 执行结果
|
||||
|
||||
已完成一次真实生成，生成命令成功返回，产物已写入仓库内独立目录。
|
||||
|
||||
本次结论不是“生成失败”，而是“生成成功但业务语义恢复失败”，因此当前样本不能进入后续内网验证。
|
||||
|
||||
## 2. 结果等级
|
||||
|
||||
综合判断：`不通过`
|
||||
|
||||
原因：
|
||||
|
||||
1. archetype 判错
|
||||
2. bootstrap 选错
|
||||
3. endpoint 污染严重
|
||||
4. modes 未恢复
|
||||
5. request contract 未恢复
|
||||
6. column defs 未恢复
|
||||
7. 生成脚本不可达到 `tq` 级别业务语义
|
||||
|
||||
## 3. 与目标基线的核心偏差
|
||||
|
||||
### 3.1 archetype 误判
|
||||
|
||||
目标应接近：
|
||||
|
||||
- `multi_mode_request.month_week_table`
|
||||
|
||||
实际生成结果：
|
||||
|
||||
- `paginated_enrichment`
|
||||
|
||||
直接证据：
|
||||
|
||||
- `references/generation-report.md` 中写明 `Workflow archetype: paginated_enrichment`
|
||||
- `references/generation-report.json` 中 `workflowArchetype = "paginated_enrichment"`
|
||||
- 生成脚本中出现固定 `paginate -> secondary_request -> filter` 主链
|
||||
|
||||
影响：
|
||||
|
||||
1. 生成器把“月/周双模式报表”错误理解为“分页主列表 + 二次补数”
|
||||
2. 后续 request builder、response parser、artifact 结构都会沿错误路径展开
|
||||
|
||||
### 3.2 bootstrap 落点偏离 `tq` 主业务面
|
||||
|
||||
目标基线：
|
||||
|
||||
- `tq-lineloss-report` 的 `scene.toml` 将 bootstrap 锚到 `20.76.57.61`
|
||||
- `target_url` 指向线损业务页面入口
|
||||
|
||||
本次生成结果：
|
||||
|
||||
- `expected_domain = "20.77.115.36:31051"`
|
||||
- `target_url = "http://20.77.115.36:31051"`
|
||||
|
||||
影响：
|
||||
|
||||
1. 生成器把其他页面/系统入口误当成主业务承载面
|
||||
2. 即使后续请求里碰巧抓到线损接口，运行时 page context 也可能不匹配
|
||||
|
||||
### 3.3 endpoint 污染严重
|
||||
|
||||
目标状态：
|
||||
|
||||
- 只保留与线损月/周统计直接相关的主业务接口
|
||||
|
||||
本次生成结果：
|
||||
|
||||
1. 虽然抓到了 `getYearMonWeekLinelossAnalysisList`
|
||||
2. 也抓到了 `fourVerEightHorLinelossRateList`
|
||||
3. 但同时混入了大量无关 endpoint，包括：
|
||||
4. 95598 工单接口
|
||||
5. 其他业务系统登录页
|
||||
6. 各类第三方库、文档、GitHub、StackOverflow 链接
|
||||
7. `pako/jszip` 等静态资源里的外部 URL
|
||||
|
||||
影响：
|
||||
|
||||
1. endpoint candidate 污染后，主链排序会失真
|
||||
2. 生成器最终把 `API_ENDPOINTS[0]` 和 `API_ENDPOINTS[1]` 当作主请求与补请求，但这只是“列表里前两个”，不是业务上正确的月/周接口矩阵
|
||||
|
||||
### 3.4 modes 完全未恢复
|
||||
|
||||
目标状态：
|
||||
|
||||
- 至少应恢复 `month/week` 模式
|
||||
- 至少应恢复 mode switch field
|
||||
- 至少应为不同模式恢复不同 request/column/response 口径
|
||||
|
||||
本次生成结果：
|
||||
|
||||
1. `scene.toml` 中没有模式结构
|
||||
2. `generation-report.json` 中 `modes = []`
|
||||
3. 虽然 `defaultMode = "month"`、`modeSwitchField = "period_mode"` 被挂上了默认值
|
||||
4. 但实际脚本没有 `MODES`、没有 `detectMode()`、没有分模式请求
|
||||
|
||||
影响：
|
||||
|
||||
1. 这是最核心的语义缺失
|
||||
2. 当前生成物本质上无法表达 `tq` 场景的主业务结构
|
||||
|
||||
### 3.5 request contract 未恢复
|
||||
|
||||
目标状态：
|
||||
|
||||
- 月模式与周模式应分别拥有独立请求体
|
||||
- 应恢复关键字段，例如 `orgno`、`tjzq`、`level`、`rows`、`page`、`sidx`、`sord` 以及周期相关字段
|
||||
|
||||
本次生成结果：
|
||||
|
||||
1. `requestTemplate = null`
|
||||
2. `params = []`
|
||||
3. 脚本中的 `buildRequest()` 直接把 `args` 全量 `JSON.stringify`
|
||||
4. 没有 mode-specific builder
|
||||
5. 没有月/周不同字段
|
||||
|
||||
影响：
|
||||
|
||||
1. 请求合同没有闭合
|
||||
2. 即使接口名对了，也无法形成正确请求
|
||||
|
||||
### 3.6 response 与列定义未恢复
|
||||
|
||||
目标状态：
|
||||
|
||||
- 线损月模式、周模式应各自拥有明确列定义
|
||||
- 至少应恢复 `content` 路径和 mode-specific columns
|
||||
|
||||
本次生成结果：
|
||||
|
||||
1. `responsePath = "content"` 这一点是对的
|
||||
2. 但 `columnDefs = []`
|
||||
3. `normalizeRules.requiredFields = []`
|
||||
4. 最终脚本用 `rows[0]` 的 key 动态拼 `column_defs`
|
||||
|
||||
影响：
|
||||
|
||||
1. 缺少业务列语义
|
||||
2. 缺少 required field 约束
|
||||
3. 无法达到 `tq` 基线中的导出与报表结构稳定性
|
||||
|
||||
## 4. 这次生成暴露出的真实问题
|
||||
|
||||
### 4.1 不是“完全识别不到线损”
|
||||
|
||||
本次结果说明，系统并不是完全抓不到线损语义。
|
||||
|
||||
正向信号仍然存在：
|
||||
|
||||
1. 抓到了 `branch_fields=tjzq, mode, week, month`
|
||||
2. 抓到了线损核心接口名
|
||||
3. 抓到了 `responsePath = content`
|
||||
|
||||
这说明底层“有感知”。
|
||||
|
||||
### 4.2 但当前仍停在“信号能抓到，主链不会重建”
|
||||
|
||||
也就是说：
|
||||
|
||||
1. 能看到线损相关 endpoint
|
||||
2. 能看到 month/week 字样
|
||||
3. 但不会把这些信号重建成 `multi_mode_request.month_week_table`
|
||||
4. 反而被分页、补数、过滤等噪声信号夺走主导权
|
||||
|
||||
### 4.3 readiness=A 目前不可信
|
||||
|
||||
本次样本虽然给了 `Readiness: A`，但实际核心合同大量缺失。
|
||||
|
||||
这说明当前 readiness 仍偏结构乐观，不能代表“可运行”或“接近内网可运行”。
|
||||
|
||||
## 5. 当前是否可以进入内网验证
|
||||
|
||||
结论：`不可以`
|
||||
|
||||
原因很直接：
|
||||
|
||||
1. archetype 错
|
||||
2. bootstrap 错
|
||||
3. modes 空
|
||||
4. request contract 空
|
||||
5. column defs 空
|
||||
|
||||
在这五项没有过关前，进入内网验证只会得到伪失败结果，没有业务验证价值。
|
||||
|
||||
## 6. 对后续批量迁移的含义
|
||||
|
||||
这次真实样本验证说明：
|
||||
|
||||
1. 当前链路已经具备“从真实场景中抓取部分业务信号”的能力
|
||||
2. 但还不具备“自动重建 `tq` 级主业务语义”的能力
|
||||
3. 因此不能直接把 102 个场景送去大规模生成
|
||||
4. 必须先把 `G2` 主样本的 archetype/mode/request/column 四件套打透
|
||||
|
||||
## 7. 下一步建议
|
||||
|
||||
下一步仍严格按既定 plan 走，不扩散：
|
||||
|
||||
1. 继续生成 `白银线损周报`
|
||||
2. 再生成 `线损同期差异报表`
|
||||
3. 对比这两个 `G2` 扩展样本是否复现同类偏差
|
||||
4. 输出一份 `G2` 家族首轮 blocker 汇总
|
||||
|
||||
只有当 `G2` 家族的 blocker 画像稳定后，才进入后续整改或更大范围迁移。
|
||||
@@ -0,0 +1,60 @@
|
||||
# Roadmap Execution Status Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round continues executing the remaining in-plan work from `2026-04-17-scene-skill-60-to-90-roadmap-plan.md`.
|
||||
|
||||
It does not create new family contracts. It records the current roadmap execution state after `G1-E`, `G2`, and `G3` were all promoted into the same family-asset shape.
|
||||
|
||||
## Delivered
|
||||
|
||||
Added:
|
||||
|
||||
1. `tests/fixtures/generated_scene/roadmap_execution_status_2026-04-18.json`
|
||||
2. `tests/roadmap_execution_status_test.rs`
|
||||
|
||||
## Current Plan Position
|
||||
|
||||
The roadmap is now recorded as:
|
||||
|
||||
1. `Phase 0 = completed`
|
||||
2. `Phase 1 = completed`
|
||||
3. `Phase 2 = completed`
|
||||
4. `Phase 3 = completed`
|
||||
5. `Phase 4 = in_progress`
|
||||
|
||||
Within `Phase 4`, the current track status is:
|
||||
|
||||
1. `Track A / G2 = batch-expansion-promoted`
|
||||
2. `Track B / G1-E = batch-expansion-promoted`
|
||||
3. `Track C / G3 = batch-expansion-promoted`
|
||||
4. `Track D = in_progress`
|
||||
5. `Track E = status-overlay-established`
|
||||
|
||||
## Why This Asset Exists
|
||||
|
||||
At this point, the repo already has:
|
||||
|
||||
1. family policy assets
|
||||
2. family result assets
|
||||
3. frozen ledger snapshot assets
|
||||
4. current ledger overlay assets
|
||||
5. family candidate-batch assets for all mainline families
|
||||
|
||||
The roadmap still needed one status asset that states, in one place, where the plan currently stands and what remains queued or deferred.
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
1. `cargo test --test roadmap_execution_status_test -- --nocapture`
|
||||
2. `cargo test --test scene_ledger_status_test -- --nocapture`
|
||||
3. `cargo test --test scene_generator_family_policy_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
The roadmap now has a repo-local execution-status asset that makes the remaining in-plan work explicit:
|
||||
|
||||
1. continue from queued mainline candidates
|
||||
2. keep Track E overlays synchronized
|
||||
3. do not reopen already completed P0 and family-baseline work
|
||||
@@ -0,0 +1,62 @@
|
||||
# Track E Ledger Status Overlay Report
|
||||
|
||||
## Scope
|
||||
|
||||
This round does not rewrite the frozen workbook snapshot from `2026-04-18 16:48:05`.
|
||||
|
||||
It adds a current-status overlay asset so the roadmap Track E state can reflect the code-backed family baselines that were established after the workbook snapshot was frozen.
|
||||
|
||||
## Delivered
|
||||
|
||||
Added:
|
||||
|
||||
1. `tests/fixtures/generated_scene/scene_ledger_status_2026-04-18.json`
|
||||
2. `tests/scene_ledger_status_test.rs`
|
||||
|
||||
## Why This Overlay Exists
|
||||
|
||||
By this point in the roadmap, the repo already contains code-backed baseline assets for:
|
||||
|
||||
1. `G2`
|
||||
2. `G1-E`
|
||||
3. `G3`
|
||||
4. `G6/G7/G8` boundary-runtime families
|
||||
|
||||
But the original workbook snapshot still reflects an earlier mid-run state.
|
||||
|
||||
Track E therefore needs two layers:
|
||||
|
||||
1. frozen snapshot layer
|
||||
2. current baseline overlay layer
|
||||
|
||||
## Current Mainline Overlay Status
|
||||
|
||||
The overlay now records:
|
||||
|
||||
1. `G2 = batch-expansion-promoted`
|
||||
2. `G1-E = batch-expansion-promoted`
|
||||
3. `G3 = batch-expansion-promoted`
|
||||
|
||||
And explicitly records promoted ledger-facing entries for:
|
||||
|
||||
1. the `G2` promoted baseline and promoted expansion fixtures
|
||||
2. the `G1-E` promoted baseline
|
||||
3. the `G3` promoted baseline
|
||||
4. the current `G3` promoted expansion fixtures
|
||||
|
||||
## Validation
|
||||
|
||||
Passed:
|
||||
|
||||
1. `cargo test --test scene_ledger_status_test -- --nocapture`
|
||||
2. `cargo test --test scene_ledger_snapshot_test -- --nocapture`
|
||||
3. `cargo test --test scene_generator_family_policy_test -- --nocapture`
|
||||
|
||||
## Outcome
|
||||
|
||||
Track E now has a stable way to express both:
|
||||
|
||||
1. the original frozen `102`-scene workbook state
|
||||
2. the current code-backed roadmap status
|
||||
|
||||
That means the next roadmap round can continue promoting candidates without losing visibility into which family baselines have already been formally established.
|
||||
@@ -0,0 +1,71 @@
|
||||
# 102 Final Coverage Status Rollup Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-102-final-coverage-status-rollup-plan.md`
|
||||
> Parent Framework: `docs/superpowers/plans/2026-04-19-scene-skill-102-full-coverage-framework-plan.md`
|
||||
> Status: Completed
|
||||
|
||||
## Scope
|
||||
|
||||
This report publishes the final policy-governed coverage rollup after the residual 13 closure sequence.
|
||||
|
||||
This plan does not update `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`, does not modify analyzer/generator logic, does not rerun the 102 sweep, and does not promote scenes to official board status.
|
||||
|
||||
## Inputs
|
||||
|
||||
- `tests/fixtures/generated_scene/full_coverage_reconciliation_candidates_2026-04-19.json`
|
||||
- `tests/fixtures/generated_scene/residual_13_reconciliation_candidates_2026-04-19.json`
|
||||
- `tests/fixtures/generated_scene/boundary_residual_hold_decision_2026-04-19.json`
|
||||
- `tests/fixtures/generated_scene/bootstrap_target_residual_isolation_2026-04-19.json`
|
||||
- `tests/fixtures/generated_scene/promotion_board_reconciliation_policy_2026-04-19.json`
|
||||
|
||||
## Final Rollup
|
||||
|
||||
| Final candidate status | Count |
|
||||
| --- | ---: |
|
||||
| `framework-auto-pass-candidate` | 95 |
|
||||
| `framework-structured-fail-closed` | 7 |
|
||||
| `framework-valid-host-bridge` | 0 |
|
||||
| `hygiene-pass-candidate` | 0 |
|
||||
| `hygiene-fail-closed-candidate` | 0 |
|
||||
| `source-unreadable` | 0 |
|
||||
| `missing-source` | 0 |
|
||||
| `unsupported-family` | 0 |
|
||||
| `misclassified-unresolved` | 0 |
|
||||
| `unresolved-followup-status` | 0 |
|
||||
|
||||
## Remaining Non-Auto-Pass Scenes
|
||||
|
||||
The remaining 7 records are explained residuals, not unresolved framework failures.
|
||||
|
||||
| Scene ID | Scene | Final status | Overlay |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-033-scene` | `供电可靠率指标统计表` | `framework-structured-fail-closed` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-034-scene` | `供电可靠性数据质量自查报告月报` | `framework-structured-fail-closed` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-042-scene` | `国网金昌供电公司营商环境周例会报告` | `framework-structured-fail-closed` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-051-scene` | `嘉峪关可靠性分析报告` | `framework-structured-fail-closed` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-074-scene` | `同兴智能安全督查日报` | `framework-structured-fail-closed` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-085-scene` | `业扩报装管理制度` | `framework-structured-fail-closed` | `hold-for-host-bridge-runtime-roadmap` |
|
||||
| `sweep-091-scene` | `用户停电频次分析监测` | `framework-structured-fail-closed` | `isolate-bootstrap-target-residual` |
|
||||
|
||||
## Interpretation
|
||||
|
||||
The 102-scene framework now has no unresolved category:
|
||||
|
||||
- no `source-unreadable`;
|
||||
- no `missing-source`;
|
||||
- no `unsupported-family`;
|
||||
- no `misclassified-unresolved`;
|
||||
- no `unresolved-followup-status`.
|
||||
|
||||
The final status is therefore `95` framework auto-pass candidates plus `7` explained structured fail-closed residuals.
|
||||
|
||||
## Output Asset
|
||||
|
||||
- `tests/fixtures/generated_scene/final_coverage_status_rollup_2026-04-19.json`
|
||||
|
||||
## Next Step
|
||||
|
||||
The next bounded step should be an official board reconciliation plan. That plan may consume this rollup and decide whether and how to update the official execution board.
|
||||
|
||||
Per Route 6 policy, this rollup does not update the official board by itself.
|
||||
@@ -0,0 +1,44 @@
|
||||
# 102 Full Coverage Follow-Up Sweep Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-102-full-coverage-followup-sweep-and-reconciliation-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This sweep reran the fixed 102-scene set after Route 2 through Route 6. It did not modify analyzer/generator logic and did not update the official execution board.
|
||||
|
||||
## Raw Sweep Summary
|
||||
|
||||
| Raw status | Count |
|
||||
| --- | ---: |
|
||||
| auto-pass | 89 |
|
||||
| fail-closed-known | 13 |
|
||||
| adjudicated-valid-host-bridge | 0 |
|
||||
| source-unreadable | 0 |
|
||||
| missing-source | 0 |
|
||||
| unsupported-family | 0 |
|
||||
| misclassified-unresolved | 0 |
|
||||
|
||||
## Archetype Summary
|
||||
|
||||
| Archetype | Count |
|
||||
| --- | ---: |
|
||||
| paginated_enrichment | 51 |
|
||||
| host_bridge_workflow | 27 |
|
||||
| multi_mode_request | 10 |
|
||||
| local_doc_pipeline | 5 |
|
||||
| single_request_enrichment | 5 |
|
||||
| multi_endpoint_inventory | 2 |
|
||||
| page_state_eval | 2 |
|
||||
|
||||
## Coverage
|
||||
|
||||
| Metric | Count |
|
||||
| --- | ---: |
|
||||
| auto-pass | 89 / 102 |
|
||||
| raw non-auto-pass | 13 / 102 |
|
||||
|
||||
## Output
|
||||
|
||||
- `tests/fixtures/generated_scene/full_coverage_followup_sweep_2026-04-19.json`
|
||||
- `examples/full_coverage_followup_sweep_2026-04-19`
|
||||
@@ -0,0 +1,42 @@
|
||||
# 102 Full Coverage Reconciliation Candidates Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-102-full-coverage-followup-sweep-and-reconciliation-plan.md`
|
||||
> Policy: `tests/fixtures/generated_scene/promotion_board_reconciliation_policy_2026-04-19.json`
|
||||
|
||||
## Scope
|
||||
|
||||
This report applies Route 6 policy to the follow-up sweep and publishes candidate statuses only. It does not update `scene_execution_board_2026-04-18.json`.
|
||||
|
||||
## Reconciliation Candidate Summary
|
||||
|
||||
| Candidate status | Count |
|
||||
| --- | ---: |
|
||||
| framework-auto-pass-candidate | 89 |
|
||||
| framework-structured-fail-closed | 13 |
|
||||
| framework-valid-host-bridge | 0 |
|
||||
| hygiene-pass-candidate | 0 |
|
||||
| hygiene-fail-closed-candidate | 0 |
|
||||
| source-unreadable | 0 |
|
||||
| missing-source | 0 |
|
||||
| unsupported-family | 0 |
|
||||
| misclassified-unresolved | 0 |
|
||||
|
||||
## Coverage View
|
||||
|
||||
| Metric | Count |
|
||||
| --- | ---: |
|
||||
| framework-auto-pass-candidate | 89 / 102 |
|
||||
| framework-supported or explained | 102 / 102 |
|
||||
| remaining raw source-unreadable | 0 / 102 |
|
||||
| unresolved misclassified | 0 / 102 |
|
||||
| unsupported-family | 0 / 102 |
|
||||
|
||||
## Stop Statement
|
||||
|
||||
The follow-up sweep and reconciliation candidate reports are published. No new implementation route is started by this plan.
|
||||
|
||||
## Output
|
||||
|
||||
- `tests/fixtures/generated_scene/full_coverage_reconciliation_candidates_2026-04-19.json`
|
||||
- `docs/superpowers/reports/2026-04-19-102-full-coverage-reconciliation-candidates-report.md`
|
||||
@@ -0,0 +1,135 @@
|
||||
# 102 Full Sweep Dry-Run Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-102-full-sweep-dry-run-plan.md`
|
||||
> Result: `tests/fixtures/generated_scene/full_sweep_dry_run_2026-04-19.json`
|
||||
> Output Root: `examples/full_sweep_dry_run_2026-04-19`
|
||||
|
||||
## Scope
|
||||
|
||||
This run measured current generic `scene -> skill` coverage over the fixed `102` scene execution board.
|
||||
|
||||
It was a measurement-only dry-run:
|
||||
|
||||
1. no analyzer logic was changed
|
||||
2. no generator logic was changed
|
||||
3. `scene_execution_board_2026-04-18.json` was not updated
|
||||
4. no scene was promoted from this result
|
||||
5. failures were recorded, not fixed
|
||||
|
||||
## Headline Numbers
|
||||
|
||||
| Metric | Count |
|
||||
| --- | ---: |
|
||||
| Real-sample executed pass | 5 / 102 |
|
||||
| Code-backed ledger coverage | 23 / 102 |
|
||||
| Dry-run auto-pass | 40 / 102 |
|
||||
| Dry-run actionable coverage | 66 / 102 |
|
||||
|
||||
`dry-run actionable coverage` is `auto-pass + fail-closed-known`.
|
||||
|
||||
## Dry-Run Summary
|
||||
|
||||
| Dry-run status | Count |
|
||||
| --- | ---: |
|
||||
| `auto-pass` | 40 |
|
||||
| `fail-closed-known` | 26 |
|
||||
| `misclassified` | 5 |
|
||||
| `unsupported-family` | 0 |
|
||||
| `missing-source` | 0 |
|
||||
| `source-unreadable` | 31 |
|
||||
| Total | 102 |
|
||||
|
||||
## Archetype Distribution
|
||||
|
||||
| Inferred archetype | Count |
|
||||
| --- | ---: |
|
||||
| `host_bridge_workflow` | 31 |
|
||||
| `paginated_enrichment` | 8 |
|
||||
| `multi_mode_request` | 3 |
|
||||
| `multi_endpoint_inventory` | 2 |
|
||||
| `page_state_eval` | 2 |
|
||||
| `none` | 56 |
|
||||
|
||||
The `none` bucket includes generator failures and timeout cases that did not produce a `generation-report.json`.
|
||||
|
||||
## Auto-Pass Shape
|
||||
|
||||
The `40` auto-pass scenes are distributed as:
|
||||
|
||||
| Inferred archetype | Auto-pass count |
|
||||
| --- | ---: |
|
||||
| `host_bridge_workflow` | 26 |
|
||||
| `paginated_enrichment` | 8 |
|
||||
| `multi_mode_request` | 3 |
|
||||
| `multi_endpoint_inventory` | 2 |
|
||||
| `page_state_eval` | 1 |
|
||||
|
||||
This means the current generic generator is no longer limited to the `23` code-backed ledger scenes. The conservative ledger coverage is lower because it only counts scenes already mapped into formal baseline or boundary assets.
|
||||
|
||||
## Non-Pass Buckets
|
||||
|
||||
### Source-Unreadable
|
||||
|
||||
`31` scenes timed out during this bounded dry-run.
|
||||
|
||||
All timeout records use:
|
||||
|
||||
`generator timeout after 30s`
|
||||
|
||||
These should not be interpreted as unsupported family evidence. They are dry-run execution-limit failures and need separate timeout/performance triage before capability conclusions are drawn.
|
||||
|
||||
### Fail-Closed-Known
|
||||
|
||||
`26` scenes failed without an auto-pass result but were recorded with a known dry-run failure category.
|
||||
|
||||
Top reasons:
|
||||
|
||||
| Reason | Count |
|
||||
| --- | ---: |
|
||||
| `generator failed without generation report` | 25 |
|
||||
| `bootstrap_target` | 1 |
|
||||
|
||||
The `generator failed without generation report` bucket is actionable but too broad for implementation work. It should be split in a later bounded triage pass before any fixes are attempted.
|
||||
|
||||
### Misclassified
|
||||
|
||||
`5` scenes produced a package, but the inferred archetype conflicted with the current board group:
|
||||
|
||||
| Scene | Current group | Inferred archetype |
|
||||
| --- | --- | --- |
|
||||
| `95598报修工单日管控` | `G3` | `host_bridge_workflow` |
|
||||
| `95598重要服务事项报备统计表` | `G3` | `host_bridge_workflow` |
|
||||
| `用电报装信息统计列表` | `G1-E` | `host_bridge_workflow` |
|
||||
| `配网支撑月报(95598抢修统计报表)` | `G3` | `host_bridge_workflow` |
|
||||
| `高低压新增报装容量月度统计表` | `G1-E` | `host_bridge_workflow` |
|
||||
|
||||
This is the clearest blocker category from the dry-run because it indicates current generic routing can over-prefer `host_bridge_workflow` on some scenes that already have board-level family expectations.
|
||||
|
||||
## Interpretation
|
||||
|
||||
The four coverage numbers answer different questions:
|
||||
|
||||
1. `5 / 102` is the strict real-sample pass count.
|
||||
2. `23 / 102` is the formal code-backed ledger coverage.
|
||||
3. `40 / 102` is the current generic dry-run auto-pass count.
|
||||
4. `66 / 102` is the current generic actionable coverage count.
|
||||
|
||||
The key result is that the generic generator currently auto-passes more scenes than the formal ledger coverage shows, but the result is not clean enough to promote automatically because:
|
||||
|
||||
1. `31` scenes hit bounded dry-run timeouts.
|
||||
2. `5` scenes show board-vs-archetype mismatch.
|
||||
3. `26` scenes need more specific failure extraction before implementation work.
|
||||
|
||||
## Recommended Next Blocker
|
||||
|
||||
Do not start implementation from this report directly.
|
||||
|
||||
The next bounded step should be a dry-run triage pass, with priority:
|
||||
|
||||
1. split the `31` timeout cases into true timeout, oversized source, and command-level hang
|
||||
2. inspect the `5` misclassified cases as the first routing-quality sample
|
||||
3. refine the `25` generic no-report failures into concrete failure categories
|
||||
|
||||
This report does not update the execution board and does not promote any scene.
|
||||
|
||||
@@ -0,0 +1,117 @@
|
||||
# 102 Full Sweep Dry-Run Triage Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-102-full-sweep-dry-run-triage-plan.md`
|
||||
> Source Dry-Run: `tests/fixtures/generated_scene/full_sweep_dry_run_2026-04-19.json`
|
||||
> Triage Result: `tests/fixtures/generated_scene/full_sweep_dry_run_triage_2026-04-19.json`
|
||||
|
||||
## Scope
|
||||
|
||||
This triage explains the `62` non-pass records from the `102` scene full sweep.
|
||||
|
||||
It is classification-only:
|
||||
|
||||
1. no analyzer implementation was changed
|
||||
2. no generator implementation was changed
|
||||
3. `scene_execution_board_2026-04-18.json` was not updated
|
||||
4. no scene was promoted
|
||||
5. no implementation correction was started
|
||||
|
||||
## Input Buckets
|
||||
|
||||
| Bucket | Count |
|
||||
| --- | ---: |
|
||||
| Timeout / `source-unreadable` | 31 |
|
||||
| Misclassified | 5 |
|
||||
| No-report failure | 25 |
|
||||
| Bootstrap target failure | 1 |
|
||||
| Total non-pass records | 62 |
|
||||
|
||||
## Timeout Triage
|
||||
|
||||
All `31` timeout records originally had:
|
||||
|
||||
`generator timeout after 30s`
|
||||
|
||||
The second-level labels are:
|
||||
|
||||
| Timeout label | Count |
|
||||
| --- | ---: |
|
||||
| `timeout-unvalidated-source` | 19 |
|
||||
| `timeout-large-source` | 8 |
|
||||
| `timeout-known-family-sample` | 4 |
|
||||
|
||||
No timeout record is treated as unsupported family by default.
|
||||
|
||||
The `timeout-known-family-sample` records are important because they include scenes that already have current family mapping. These should be separated from the unvalidated-source and large-source buckets before any coverage conclusion is drawn.
|
||||
|
||||
## Misclassification Triage
|
||||
|
||||
All `5` misclassified records share the same second-level label:
|
||||
|
||||
`route-overprefer-host-bridge`
|
||||
|
||||
| Scene | Expected group | Expected archetype | Inferred archetype |
|
||||
| --- | --- | --- | --- |
|
||||
| `95598报修工单日管控` | `G3` | `paginated_enrichment` | `host_bridge_workflow` |
|
||||
| `95598重要服务事项报备统计表` | `G3` | `paginated_enrichment` | `host_bridge_workflow` |
|
||||
| `用电报装信息统计列表` | `G1-E` | `single_request_enrichment` | `host_bridge_workflow` |
|
||||
| `配网支撑月报(95598抢修统计报表)` | `G3` | `paginated_enrichment` | `host_bridge_workflow` |
|
||||
| `高低压新增报装容量月度统计表` | `G1-E` | `single_request_enrichment` | `host_bridge_workflow` |
|
||||
|
||||
This is the clearest routing-quality signal in the dry-run. The current generic route can over-prefer `host_bridge_workflow` when board-level expectations already point to `G3` or `G1-E`.
|
||||
|
||||
This report does not correct routing logic.
|
||||
|
||||
## No-Report Failure Triage
|
||||
|
||||
All `25` generic no-report failures were assigned the same failure stage:
|
||||
|
||||
`readiness-before-report`
|
||||
|
||||
The stderr archetype distribution is:
|
||||
|
||||
| Stderr archetype | Count |
|
||||
| --- | ---: |
|
||||
| `paginated_enrichment` | 17 |
|
||||
| `local_doc_pipeline` | 4 |
|
||||
| `multi_mode_request` | 2 |
|
||||
| `host_bridge_workflow` | 1 |
|
||||
| `single_request_enrichment` | 1 |
|
||||
|
||||
This means these failures are not source-missing failures. They reached a known archetype path, then failed before a structured `generation-report.json` was emitted.
|
||||
|
||||
The dominant no-report shape is:
|
||||
|
||||
`workflow evidence is incomplete for archetype paginated_enrichment`
|
||||
|
||||
## Bootstrap Target Failure
|
||||
|
||||
One failure remains separately tracked:
|
||||
|
||||
| Scene | Inferred archetype | Reason |
|
||||
| --- | --- | --- |
|
||||
| `用户停电频次分析监测` | `page_state_eval` | `bootstrap_target` |
|
||||
|
||||
It is not merged into the no-report bucket.
|
||||
|
||||
## Recommended Next Blocker
|
||||
|
||||
The next blocker should be handled in this order:
|
||||
|
||||
1. `timeout-known-family-sample`
|
||||
2. `route-overprefer-host-bridge`
|
||||
3. `readiness-before-report`
|
||||
|
||||
Reasoning:
|
||||
|
||||
1. `timeout-known-family-sample` affects already-mapped scenes and can distort coverage if left mixed with unvalidated timeouts.
|
||||
2. `route-overprefer-host-bridge` is a clean routing-quality issue across existing `G3` and `G1-E` expectations.
|
||||
3. `readiness-before-report` is high volume, but it first needs structured fail-closed reporting before implementation correction.
|
||||
|
||||
## Stop Statement
|
||||
|
||||
This triage is complete and stops here.
|
||||
|
||||
It does not start an implementation plan, does not promote scenes, and does not update the execution board.
|
||||
|
||||
@@ -0,0 +1,82 @@
|
||||
# 102 Full Sweep Improvement Coverage Delta Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: `docs/superpowers/plans/2026-04-19-102-full-sweep-improvement-roadmap-plan.md`
|
||||
> Baseline: `tests/fixtures/generated_scene/full_sweep_dry_run_2026-04-19.json`
|
||||
> Follow-Up: `tests/fixtures/generated_scene/full_sweep_improvement_followup_2026-04-19.json`
|
||||
|
||||
## Headline Delta
|
||||
|
||||
| Metric | Baseline | Follow-up | Delta |
|
||||
| --- | ---: | ---: | ---: |
|
||||
| Dry-run auto-pass | 40 | 48 | +8 |
|
||||
| Dry-run actionable coverage | 66 | 96 | +30 |
|
||||
| Timeout / source-unreadable | 31 | 2 | -29 |
|
||||
| Misclassified | 5 | 4 | -1 |
|
||||
| Fail-closed-known | 26 | 48 | +22 |
|
||||
|
||||
The increase in `fail-closed-known` is expected. The roadmap converted many timeout/no-report cases into structured fail-closed outputs instead of process-level failures.
|
||||
|
||||
## Follow-Up Status Counts
|
||||
|
||||
| Status | Count |
|
||||
| --- | ---: |
|
||||
| `auto-pass` | 48 |
|
||||
| `fail-closed-known` | 48 |
|
||||
| `misclassified` | 4 |
|
||||
| `source-unreadable` | 2 |
|
||||
| `missing-source` | 0 |
|
||||
| `unsupported-family` | 0 |
|
||||
| Total | 102 |
|
||||
|
||||
## Remaining Timeout Cases
|
||||
|
||||
Only `2` timeout cases remain:
|
||||
|
||||
1. `任务报表`
|
||||
2. `力禾动环系统巡视记录`
|
||||
|
||||
These are not known-family samples.
|
||||
|
||||
## Remaining Misclassification Cases
|
||||
|
||||
`4` route conflicts remain:
|
||||
|
||||
| Scene | Expected group | Inferred archetype |
|
||||
| --- | --- | --- |
|
||||
| `95598报修工单日管控` | `G3` | `host_bridge_workflow` |
|
||||
| `95598重要服务事项报备统计表` | `G3` | `host_bridge_workflow` |
|
||||
| `台区线损台区月度高负损预测` | `G2` | `host_bridge_workflow` |
|
||||
| `配网支撑月报(95598抢修统计报表)` | `G3` | `host_bridge_workflow` |
|
||||
|
||||
One previous `G1-E` conflict, `高低压新增报装容量月度统计表`, now returns to `single_request_enrichment` and auto-passes.
|
||||
|
||||
`用电报装信息统计列表` now returns to `single_request_enrichment` but remains fail-closed, so it is no longer a route-overpreference case.
|
||||
|
||||
## Structured Fail-Closed Result
|
||||
|
||||
The baseline had `25` no-report failures. The follow-up emits structured fail-closed reports for these paths, with:
|
||||
|
||||
1. inferred archetype
|
||||
2. `generationStatus = fail-closed`
|
||||
3. `failureStage = readiness-before-report`
|
||||
4. blocker reason
|
||||
5. readiness missing pieces and risks
|
||||
|
||||
This makes failures machine-readable without weakening gates.
|
||||
|
||||
## Interpretation
|
||||
|
||||
The roadmap delivered measurable improvement:
|
||||
|
||||
1. timeout noise was reduced from `31` to `2`
|
||||
2. auto-pass increased from `40` to `48`
|
||||
3. actionable coverage increased from `66` to `96`
|
||||
4. no-report failures were converted into structured fail-closed results
|
||||
|
||||
Remaining work is no longer the same broad ambiguity. The next blocker is the smaller route-conflict set, especially `G3/G2` vs `host_bridge_workflow`.
|
||||
|
||||
## Stop Statement
|
||||
|
||||
This report does not update `scene_execution_board_2026-04-18.json` and does not promote any scene.
|
||||
|
||||
@@ -0,0 +1,80 @@
|
||||
# 102 Full Sweep Improvement Roadmap Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-102-full-sweep-improvement-roadmap-plan.md`
|
||||
|
||||
## Closure Summary
|
||||
|
||||
The roadmap has completed its planned cycle:
|
||||
|
||||
1. baseline dry-run and triage were frozen
|
||||
2. known-family timeouts were diagnosed
|
||||
3. source scanning was bounded to reduce vendor/static-file timeout noise
|
||||
4. route over-preference was partially corrected and adjudicated
|
||||
5. readiness-before-report failures now emit structured fail-closed reports
|
||||
6. bootstrap target remained isolated
|
||||
7. a follow-up `102` sweep quantified coverage delta
|
||||
|
||||
## Final Coverage Delta
|
||||
|
||||
| Metric | Baseline | Follow-up | Delta |
|
||||
| --- | ---: | ---: | ---: |
|
||||
| Auto-pass | 40 | 48 | +8 |
|
||||
| Actionable coverage | 66 | 96 | +30 |
|
||||
| Timeout | 31 | 2 | -29 |
|
||||
| Misclassified | 5 | 4 | -1 |
|
||||
|
||||
## Implementation Scope
|
||||
|
||||
The implementation stayed within the roadmap:
|
||||
|
||||
1. no new family was added
|
||||
2. `G4/G5` were not started
|
||||
3. login recovery was not implemented
|
||||
4. host runtime transport was not implemented
|
||||
5. attachment/local document runtime was not implemented
|
||||
6. `scene_execution_board_2026-04-18.json` was not updated by the follow-up sweep
|
||||
|
||||
## Files Changed
|
||||
|
||||
Code changes:
|
||||
|
||||
1. `src/generated_scene/analyzer.rs`
|
||||
2. `src/generated_scene/generator.rs`
|
||||
|
||||
New result assets:
|
||||
|
||||
1. `tests/fixtures/generated_scene/known_family_timeout_diagnostic_2026-04-19.json`
|
||||
2. `tests/fixtures/generated_scene/full_sweep_improvement_followup_2026-04-19.json`
|
||||
|
||||
New reports:
|
||||
|
||||
1. `docs/superpowers/reports/2026-04-19-known-family-timeout-diagnostic-report.md`
|
||||
2. `docs/superpowers/reports/2026-04-19-102-full-sweep-improvement-coverage-delta-report.md`
|
||||
3. `docs/superpowers/reports/2026-04-19-102-full-sweep-improvement-roadmap-closure-report.md`
|
||||
|
||||
## Verification
|
||||
|
||||
Executed checks:
|
||||
|
||||
1. `cargo test --test scene_generator_test -- --nocapture`
|
||||
2. `cargo test --test scene_generator_canonical_test -- --nocapture`
|
||||
3. `cargo test --test post_roadmap_execution_assets_test -- --nocapture`
|
||||
4. `cargo test --test scene_generator_family_policy_test -- --nocapture`
|
||||
5. follow-up full sweep over the fixed `102` scene set
|
||||
|
||||
## Remaining Blocker
|
||||
|
||||
The remaining highest-value blocker is no longer the broad timeout/no-report ambiguity.
|
||||
|
||||
The next blocker is the smaller routing-conflict set:
|
||||
|
||||
1. `G3 -> host_bridge_workflow`
|
||||
2. `G2 -> host_bridge_workflow`
|
||||
|
||||
This should be handled by a new bounded route-conflict roadmap only if further work is requested.
|
||||
|
||||
## Stop Statement
|
||||
|
||||
This roadmap stops here. It does not automatically start a new route-conflict implementation plan and does not update the execution board.
|
||||
|
||||
@@ -0,0 +1,83 @@
|
||||
# 102 Sweep Status Reconciliation Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-102-sweep-status-reconciliation-plan.md`
|
||||
> Reconciliation Asset: `tests/fixtures/generated_scene/full_sweep_status_reconciliation_2026-04-19.json`
|
||||
|
||||
## Input Validation
|
||||
|
||||
The reconciliation used only the frozen inputs required by the plan:
|
||||
|
||||
| Input | Count |
|
||||
| --- | ---: |
|
||||
| Follow-up sweep scenes | 102 |
|
||||
| Route-conflict decisions | 4 |
|
||||
|
||||
All `4` route-conflict decisions matched a follow-up `sceneId`, and each matched record kept raw `dryRunStatus = misclassified`. No `102` sweep rerun was performed.
|
||||
|
||||
## Raw Follow-Up Counts
|
||||
|
||||
| Raw dry-run status | Count |
|
||||
| --- | ---: |
|
||||
| `auto-pass` | 48 |
|
||||
| `fail-closed-known` | 48 |
|
||||
| `misclassified` | 4 |
|
||||
| `source-unreadable` | 2 |
|
||||
| `missing-source` | 0 |
|
||||
| `unsupported-family` | 0 |
|
||||
|
||||
The raw follow-up asset still reports `misclassified = 4` because it is the original sweep measurement. Those `4` records are not rewritten in place.
|
||||
|
||||
## Reconciled Counts
|
||||
|
||||
| Reconciled status | Count |
|
||||
| --- | ---: |
|
||||
| `auto-pass` | 48 |
|
||||
| `fail-closed-known` | 48 |
|
||||
| `adjudicated-valid-host-bridge` | 4 |
|
||||
| `source-unreadable` | 2 |
|
||||
| `missing-source` | 0 |
|
||||
| `unsupported-family` | 0 |
|
||||
| `misclassified-unresolved` | 0 |
|
||||
| Total | 102 |
|
||||
|
||||
The raw `misclassified = 4` has been fully adjudicated by `remaining_route_conflict_decisions_2026-04-19.json`. In the reconciled view, these no longer represent unresolved route bugs; they are valid host-bridge workflows whose prior group expectations did not close the G2/G3 contracts.
|
||||
|
||||
## Valid Host-Bridge Adjudications
|
||||
|
||||
| Scene ID | Scene | Raw status | Route decision | Reconciled status |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| `sweep-007-95598` | 95598报修工单日管控 | `misclassified` | `valid-host-bridge-workflow` | `adjudicated-valid-host-bridge` |
|
||||
| `sweep-009-95598` | 95598重要服务事项报备统计表 | `misclassified` | `valid-host-bridge-workflow` | `adjudicated-valid-host-bridge` |
|
||||
| `sweep-029-scene` | 台区线损台区月度高负损预测 | `misclassified` | `valid-host-bridge-workflow` | `adjudicated-valid-host-bridge` |
|
||||
| `sweep-093-95598` | 配网支撑月报(95598抢修统计报表) | `misclassified` | `valid-host-bridge-workflow` | `adjudicated-valid-host-bridge` |
|
||||
|
||||
Each adjudicated record preserves `decisionReason` and `evidenceSummary` in the reconciliation JSON.
|
||||
|
||||
## Remaining Timeout Inputs
|
||||
|
||||
| Scene ID | Scene | Raw status | Reason |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-015-scene` | 任务报表 | `source-unreadable` | generator timeout after 45s |
|
||||
| `sweep-025-scene` | 力禾动环系统巡视记录 | `source-unreadable` | generator timeout after 45s |
|
||||
|
||||
These `2` records remain future timeout/source-scale diagnostic inputs.
|
||||
|
||||
After route adjudication, the remaining future inputs are exactly these `2` timeout records and the `48` structured fail-closed records summarized below.
|
||||
|
||||
## Fail-Closed Known Summary
|
||||
|
||||
| Inferred archetype | Reason | Count |
|
||||
| --- | --- | ---: |
|
||||
| `paginated_enrichment` | workflow evidence is incomplete before package generation | 35 |
|
||||
| `local_doc_pipeline` | workflow evidence is incomplete before package generation | 5 |
|
||||
| `multi_mode_request` | workflow evidence is incomplete before package generation | 4 |
|
||||
| `single_request_enrichment` | workflow evidence is incomplete before package generation | 2 |
|
||||
| `host_bridge_workflow` | workflow evidence is incomplete before package generation | 1 |
|
||||
| `page_state_eval` | bootstrap_target | 1 |
|
||||
|
||||
The remaining structured future inputs are `48` `fail-closed-known` records. They are visible as workflow-evidence or contract-completion analysis candidates, but this plan does not start that roadmap or implement fixes.
|
||||
|
||||
## Scope Confirmation
|
||||
|
||||
No execution-board status was changed. No scene was promoted. No family baseline was added or modified. `src/generated_scene/analyzer.rs`, `src/generated_scene/generator.rs`, and `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json` were not modified by this reconciliation.
|
||||
@@ -0,0 +1,55 @@
|
||||
# Bootstrap Target Normalization Roadmap Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `2026-04-19-bootstrap-target-normalization-roadmap-plan.md`
|
||||
> Fixed scene: `sweep-091-scene`
|
||||
|
||||
## Scope
|
||||
|
||||
This plan handled only `sweep-091-scene / 配网异常设备监控统计`.
|
||||
|
||||
It did not update the official board, did not start host-bridge runtime work, and did not modify `analyzer.rs` or `generator.rs`.
|
||||
|
||||
## Baseline Finding
|
||||
|
||||
The current official board had this scene as:
|
||||
|
||||
- `currentFrameworkStatus = framework-structured-fail-closed`
|
||||
- `currentFrameworkArchetype = page_state_eval`
|
||||
- `currentFrameworkReadiness = C`
|
||||
- `currentFrameworkNextAction = future-bootstrap-target-normalization-roadmap-input`
|
||||
|
||||
The residual was caused by an old path where bootstrap fields were empty. A fresh bounded rerun against the official-board source scene recovered the deterministic target:
|
||||
|
||||
- `expectedDomain = 21.77.244.194:18890`
|
||||
- `targetUrl = http://21.77.244.194:18890/mainSystem`
|
||||
|
||||
## Follow-Up Result
|
||||
|
||||
Follow-up asset:
|
||||
|
||||
`tests/fixtures/generated_scene/bootstrap_target_normalization_followup_2026-04-19.json`
|
||||
|
||||
| Scene id | Scene name | Archetype | Readiness | Result |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| `sweep-091-scene` | `配网异常设备监控统计` | `single_request_enrichment` | `A` | `auto-pass` |
|
||||
|
||||
## Reconciliation Candidate
|
||||
|
||||
Candidate asset:
|
||||
|
||||
`tests/fixtures/generated_scene/bootstrap_target_normalization_reconciliation_candidates_2026-04-19.json`
|
||||
|
||||
The scene is now published as:
|
||||
|
||||
`framework-auto-pass-candidate`
|
||||
|
||||
Per promotion policy, the official board must be updated by a dedicated final-2 board reconciliation refresh, not by this roadmap.
|
||||
|
||||
## Notes
|
||||
|
||||
Earlier residual reports include a scene-name drift where the same `sweep-091-scene` id was associated with `用户停电频次分析监测`. This roadmap uses the official-board source scene `配网异常设备监控统计` and does not edit historical assets.
|
||||
|
||||
## Closure
|
||||
|
||||
This roadmap is complete. No implementation slice was required; the single residual closed through bounded rerun and deterministic bootstrap target recovery.
|
||||
@@ -0,0 +1,18 @@
|
||||
# Bootstrap Target Residual Isolation Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-bootstrap-target-residual-isolation-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This Route D plan isolates the single page-state/bootstrap-target residual. It does not implement login recovery or runtime navigation.
|
||||
|
||||
## Decision
|
||||
|
||||
| Scene | Archetype | Status | Decision | Next Action |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| 用户停电频次分析监测 | page_state_eval | framework-structured-fail-closed | isolate-bootstrap-target-residual | future-bootstrap-target-normalization-roadmap-input |
|
||||
|
||||
## Stop Statement
|
||||
|
||||
Isolation assets are published. No login recovery, runtime navigation implementation, board update, or generator/analyzer change is started by this plan.
|
||||
@@ -0,0 +1,53 @@
|
||||
# Boundary Fail-Closed Decision Report
|
||||
|
||||
Date: 2026-04-19
|
||||
|
||||
Parent framework: `2026-04-19-scene-skill-102-full-coverage-framework-plan`
|
||||
|
||||
Parent route: `Route 5 / boundary-family fail-closed`
|
||||
|
||||
Plan: `2026-04-19-boundary-fail-closed-decision-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This is a decision-only route. It does not change analyzer/generator code and does not update the execution board.
|
||||
|
||||
Fixed Route 5 input buckets:
|
||||
|
||||
| Bucket | Count |
|
||||
| --- | ---: |
|
||||
| `local_doc_pipeline` | 5 |
|
||||
| `host_bridge_workflow` | 1 |
|
||||
| `page_state_eval/bootstrap_target` | 1 |
|
||||
|
||||
## Decisions
|
||||
|
||||
| Bucket | Decision | Reason |
|
||||
| --- | --- | --- |
|
||||
| `local_doc_pipeline` | `hold-as-boundary-fail-closed` | Requires local document runtime and attachment handling beyond the current full-coverage implementation routes. |
|
||||
| `host_bridge_workflow` | `hold-as-boundary-fail-closed` | Remaining boundary record depends on host bridge execution semantics outside this decision-only route. |
|
||||
| `page_state_eval/bootstrap_target` | `isolate-bootstrap-target` | Bootstrap target resolution is a separate navigation/login target problem and must stay isolated from contract recovery. |
|
||||
|
||||
## Result
|
||||
|
||||
Route 5 produces no code-level coverage delta. Its expected delta is decision quality:
|
||||
|
||||
- unresolved boundary ambiguity: `0`
|
||||
- open boundary implementation slices: `0`
|
||||
- held/deferred boundary records: `7`
|
||||
|
||||
## Boundary
|
||||
|
||||
This route did not:
|
||||
|
||||
- modify `src/generated_scene/analyzer.rs`
|
||||
- modify `src/generated_scene/generator.rs`
|
||||
- update `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
- start boundary implementation
|
||||
- create a new family
|
||||
|
||||
## Stop Statement
|
||||
|
||||
Route 5 is decision-complete. The next child plan in the 102 full-coverage sequence is Route 6:
|
||||
|
||||
`2026-04-19-promotion-and-board-reconciliation-policy-plan.md`
|
||||
@@ -0,0 +1,47 @@
|
||||
# Boundary Family Entry Decision Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: [2026-04-19-boundary-family-real-sample-entry-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-boundary-family-real-sample-entry-roadmap-plan.md)
|
||||
|
||||
## Decision
|
||||
|
||||
One boundary family is selected as the next real-sample entry candidate:
|
||||
|
||||
1. `G7`
|
||||
|
||||
The other boundary families remain held:
|
||||
|
||||
1. `G6`
|
||||
2. `G8`
|
||||
|
||||
## Comparison Result
|
||||
|
||||
### G6
|
||||
|
||||
`G6` already has a code-backed minimum runtime contract, but the next real-sample step would require host bridge execution semantics.
|
||||
|
||||
That prerequisite is still too platform-shaped for the smallest bounded next slice.
|
||||
|
||||
### G7
|
||||
|
||||
`G7` already has a minimal runnable runtime contract and only lacks real multi-endpoint aggregation verification.
|
||||
|
||||
Among `G6/G7/G8`, this is the smallest next capability gap.
|
||||
|
||||
So `G7` is selected.
|
||||
|
||||
### G8
|
||||
|
||||
`G8` already has a code-backed minimum runtime contract, but the next real-sample step still depends on local document pipeline runtime and attachment handling.
|
||||
|
||||
That remains heavier than the bounded next-step budget.
|
||||
|
||||
## Result
|
||||
|
||||
The roadmap now has one bounded next direction:
|
||||
|
||||
1. `G7 real-sample entry`
|
||||
|
||||
This report does not execute a real sample yet.
|
||||
|
||||
It only closes the boundary-family priority ambiguity.
|
||||
@@ -0,0 +1,32 @@
|
||||
# Boundary Family Real-Sample Entry Roadmap Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: [2026-04-19-boundary-family-real-sample-entry-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-boundary-family-real-sample-entry-roadmap-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This roadmap is closed with one bounded next direction:
|
||||
|
||||
1. `G7 real-sample entry`
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G6/G7/G8` no longer compete ambiguously for next priority
|
||||
2. only one next-family direction is opened
|
||||
3. a follow-up bounded `design + plan` now exists for the selected direction
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following families remain held:
|
||||
|
||||
1. `G6`
|
||||
2. `G8`
|
||||
|
||||
The deferred families remain untouched:
|
||||
|
||||
1. `G4`
|
||||
2. `G5`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `G7 real-sample entry` plan rather than reopening this roadmap.
|
||||
@@ -0,0 +1,31 @@
|
||||
# Boundary Residual Hold Decision Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-boundary-residual-hold-decision-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This Route C plan is decision-only. It classifies the remaining boundary residual records and does not implement runtime support.
|
||||
|
||||
## Summary
|
||||
|
||||
| Bucket | Count | Decision |
|
||||
| --- | ---: | --- |
|
||||
| local_doc_pipeline | 5 | hold-for-local-doc-runtime-roadmap |
|
||||
| host_bridge_workflow | 1 | hold-for-host-bridge-runtime-roadmap |
|
||||
| unresolved boundary ambiguity | 0 | none |
|
||||
|
||||
## Decisions
|
||||
|
||||
| Scene | Archetype | Decision | Next Action |
|
||||
| --- | --- | --- | --- |
|
||||
| 供电可靠率指标统计表 | local_doc_pipeline | hold-for-local-doc-runtime-roadmap | future-local-doc-runtime-roadmap-input |
|
||||
| 供电可靠性数据质量自查报告月报 | local_doc_pipeline | hold-for-local-doc-runtime-roadmap | future-local-doc-runtime-roadmap-input |
|
||||
| 国网金昌供电公司营商环境周例会报告 | local_doc_pipeline | hold-for-local-doc-runtime-roadmap | future-local-doc-runtime-roadmap-input |
|
||||
| 嘉峪关可靠性分析报告 | local_doc_pipeline | hold-for-local-doc-runtime-roadmap | future-local-doc-runtime-roadmap-input |
|
||||
| 同兴智能安全督查日报 | local_doc_pipeline | hold-for-local-doc-runtime-roadmap | future-local-doc-runtime-roadmap-input |
|
||||
| 业扩报装管理制度 | host_bridge_workflow | hold-for-host-bridge-runtime-roadmap | future-host-bridge-runtime-roadmap-input |
|
||||
|
||||
## Stop Statement
|
||||
|
||||
Decision assets are published. No boundary implementation, board update, or generator/analyzer change is started by this plan.
|
||||
@@ -0,0 +1,33 @@
|
||||
# Boundary Runtime Entry Decision
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: `docs/superpowers/plans/2026-04-18-scene-skill-real-sample-validation-roadmap-plan.md`
|
||||
|
||||
## Decision
|
||||
|
||||
`G6/G7/G8` do not enter the next roadmap.
|
||||
|
||||
They remain boundary families.
|
||||
|
||||
## Reason
|
||||
|
||||
The current executed validation pressure is still concentrated in mainline families:
|
||||
|
||||
1. `G2` real sample: mismatch
|
||||
2. `G1-E` real sample: pass
|
||||
3. `G3` real sample: mismatch
|
||||
|
||||
No boundary family currently has stronger execution pressure than the `G3` and `G2` mainline failures.
|
||||
|
||||
## Inclusion / Exclusion Result
|
||||
|
||||
Excluded from next roadmap:
|
||||
|
||||
1. `G6`
|
||||
2. `G7`
|
||||
3. `G8`
|
||||
|
||||
Entry condition remains unchanged:
|
||||
|
||||
1. a real sample must require boundary-family runtime semantics strongly enough to outrank unresolved mainline gaps
|
||||
2. the required runtime gap must be explicit rather than inferred by drift
|
||||
@@ -0,0 +1,47 @@
|
||||
# Boundary Runtime Prerequisites Decision Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: [2026-04-19-boundary-runtime-prerequisites-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-boundary-runtime-prerequisites-roadmap-plan.md)
|
||||
|
||||
## Decision
|
||||
|
||||
One bounded prerequisite direction is selected:
|
||||
|
||||
1. `G6 host-bridge prerequisites`
|
||||
|
||||
The following prerequisite direction remains held:
|
||||
|
||||
1. `G8 local-doc prerequisites`
|
||||
|
||||
## Comparison Result
|
||||
|
||||
### G6 Host-Bridge Prerequisites
|
||||
|
||||
`G6` is blocked by one clearer prerequisite line:
|
||||
|
||||
1. host bridge real-sample execution semantics
|
||||
|
||||
That is still platform-shaped, but it is narrower than the combined local-doc burden on `G8`.
|
||||
|
||||
So `G6 host-bridge prerequisites` is selected.
|
||||
|
||||
### G8 Local-Doc Prerequisites
|
||||
|
||||
`G8` is blocked by a broader prerequisite bundle:
|
||||
|
||||
1. local document runtime
|
||||
2. attachment handling
|
||||
|
||||
That bundle is heavier than the next bounded slice should absorb.
|
||||
|
||||
So `G8 local-doc prerequisites` remains held.
|
||||
|
||||
## Result
|
||||
|
||||
The roadmap now has one bounded next direction:
|
||||
|
||||
1. `G6 host-bridge prerequisites`
|
||||
|
||||
This report does not implement the prerequisite.
|
||||
|
||||
It only closes the prerequisite priority ambiguity.
|
||||
@@ -0,0 +1,30 @@
|
||||
# Boundary Runtime Prerequisites Roadmap Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: [2026-04-19-boundary-runtime-prerequisites-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-boundary-runtime-prerequisites-roadmap-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This roadmap is closed with one bounded next direction:
|
||||
|
||||
1. `G6 host-bridge prerequisites`
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G6` and `G8` no longer compete ambiguously at the prerequisite level
|
||||
2. one follow-up bounded `design + plan` now exists for the selected prerequisite direction
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following direction remains held:
|
||||
|
||||
1. `G8 local-doc prerequisites`
|
||||
|
||||
The deferred families remain untouched:
|
||||
|
||||
1. `G4`
|
||||
2. `G5`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `G6 host-bridge prerequisites` plan rather than reopening this roadmap.
|
||||
@@ -0,0 +1,35 @@
|
||||
# Deferred Family Entry Decision
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: `docs/superpowers/plans/2026-04-18-scene-skill-real-sample-validation-roadmap-plan.md`
|
||||
|
||||
## Decision
|
||||
|
||||
`G4` and `G5` do not enter the next roadmap.
|
||||
|
||||
Current status remains:
|
||||
|
||||
1. `G4`: deferred
|
||||
2. `G5`: degraded
|
||||
|
||||
## Reason
|
||||
|
||||
The current validation evidence does not justify opening deferred families.
|
||||
|
||||
The strongest observed pressure is still:
|
||||
|
||||
1. `G3` real-sample archetype collapse
|
||||
2. `G2` real-sample contract mismatch
|
||||
|
||||
Opening `G4/G5` before addressing those mainline failures would violate the bounded-scope rule of the roadmap.
|
||||
|
||||
## Updated Scope Boundary
|
||||
|
||||
The next roadmap remains limited to:
|
||||
|
||||
1. executed mainline mismatch correction
|
||||
|
||||
The next roadmap remains closed to:
|
||||
|
||||
1. `G4` implementation
|
||||
2. `G5` implementation
|
||||
@@ -0,0 +1,36 @@
|
||||
# Final 2 Official Board Reconciliation Refresh Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `2026-04-19-final-2-official-board-reconciliation-refresh-plan.md`
|
||||
|
||||
## Inputs
|
||||
|
||||
- `tests/fixtures/generated_scene/bootstrap_target_normalization_reconciliation_candidates_2026-04-19.json`
|
||||
- `tests/fixtures/generated_scene/host_bridge_runtime_reconciliation_candidates_2026-04-19.json`
|
||||
|
||||
## Result
|
||||
|
||||
| Metric | Count |
|
||||
| --- | ---: |
|
||||
| `framework-auto-pass` | 102 |
|
||||
| `framework-structured-fail-closed` | 0 |
|
||||
| unresolved framework statuses | 0 |
|
||||
|
||||
## Updated Scenes
|
||||
|
||||
| Scene id | Official board name | Candidate source name | New framework status | Archetype |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| `sweep-091-scene` | `配网异常设备监控统计` | `配网异常设备监控统计` | `framework-auto-pass` | `single_request_enrichment` |
|
||||
| `sweep-085-scene` | `计量资产库存统计` | `计量资产库存统计` | `framework-auto-pass` | `multi_endpoint_inventory` |
|
||||
|
||||
## Remaining Structured Fail-Closed
|
||||
|
||||
None.
|
||||
|
||||
## Boundary
|
||||
|
||||
This refresh updated only framework-layer fields in the official board. It did not modify analyzer/generator code and did not start another runtime roadmap.
|
||||
|
||||
## Closure
|
||||
|
||||
This plan is complete. The official board now has `102` framework auto-pass scenes and `0` framework structured fail-closed residuals.
|
||||
@@ -0,0 +1,51 @@
|
||||
# Final 2 Residual Roadmap Prioritization Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `2026-04-19-final-2-residual-roadmap-prioritization-plan.md`
|
||||
> Parent sequence: `2026-04-19-final-2-residual-child-plan-sequence-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This decision consumed the current official board state after local-doc reconciliation refresh. Only two framework structured fail-closed residuals remain:
|
||||
|
||||
| Scene id | Scene name | Current archetype | Next action |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-085-scene` | `计量资产库存统计` | `host_bridge_workflow` | `future-host-bridge-runtime-roadmap-input` |
|
||||
| `sweep-091-scene` | `配网异常设备监控统计` | `page_state_eval` | `future-bootstrap-target-normalization-roadmap-input` |
|
||||
|
||||
No code was changed, and the official board was not updated by this plan.
|
||||
|
||||
## Decision
|
||||
|
||||
Selected first roadmap:
|
||||
|
||||
`bootstrap target normalization roadmap`
|
||||
|
||||
Selected plan:
|
||||
|
||||
`2026-04-19-bootstrap-target-normalization-roadmap-plan.md`
|
||||
|
||||
## Rationale
|
||||
|
||||
Bootstrap target normalization is the narrower next step:
|
||||
|
||||
1. It is a single-scene residual.
|
||||
2. It targets deterministic bootstrap target recovery or a narrower named hold.
|
||||
3. It avoids host-runtime transport semantics.
|
||||
4. It has lower regression risk against already passing host-bridge paths.
|
||||
|
||||
Host-bridge runtime remains queued:
|
||||
|
||||
`2026-04-19-host-bridge-runtime-roadmap-plan.md`
|
||||
|
||||
## Output Asset
|
||||
|
||||
Decision asset:
|
||||
|
||||
`tests/fixtures/generated_scene/final_2_residual_roadmap_prioritization_2026-04-19.json`
|
||||
|
||||
## Closure
|
||||
|
||||
This prioritization plan is complete. The next executable plan is:
|
||||
|
||||
`docs/superpowers/plans/2026-04-19-bootstrap-target-normalization-roadmap-plan.md`
|
||||
@@ -0,0 +1,65 @@
|
||||
# G1-E Remaining Fail-Closed Closure Report
|
||||
|
||||
Date: 2026-04-19
|
||||
|
||||
Parent framework: `2026-04-19-scene-skill-102-full-coverage-framework-plan`
|
||||
|
||||
Parent route: `Route 4 / G1-E single_request_enrichment closure`
|
||||
|
||||
Plan: `2026-04-19-g1e-remaining-fail-closed-closure-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This report closes the fixed Route 4 bucket:
|
||||
|
||||
- `single_request_enrichment` structured fail-closed scenes: `2`
|
||||
- Primary blocker: `output_columns`
|
||||
- Allowed implementation area: bounded G1-E output column recovery
|
||||
|
||||
This run did not update `scene_execution_board_2026-04-18.json`, did not create a new family, and did not begin Route 5.
|
||||
|
||||
## Implementation
|
||||
|
||||
The analyzer now recovers G1-E output columns from two additional source patterns:
|
||||
|
||||
- `qxzndbObj.cols.push([...])`
|
||||
- wide `titleList` rows such as `[field, top, middle, order, label]`
|
||||
|
||||
This keeps the correction in the G1-E response/merge contract layer. It does not change routing or gate policy.
|
||||
|
||||
## Results
|
||||
|
||||
| Scene ID | Scene | Baseline | Follow-up | Result |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| `sweep-013-scene` | `业扩报装质量评价体系` | `B / fail-closed / output_columns` | `A / auto-pass` | resolved |
|
||||
| `sweep-068-scene` | `用电报装信息统计列表` | `B / fail-closed / output_columns` | `A / auto-pass` | resolved |
|
||||
|
||||
Summary:
|
||||
|
||||
- Before: `2` G1-E structured fail-closed scenes
|
||||
- Resolved to auto-pass: `2`
|
||||
- Remaining Route 4 fail-closed: `0`
|
||||
- Route-local coverage delta: `+2`
|
||||
|
||||
## Verification
|
||||
|
||||
Commands:
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test analyzer_recovers_g1e_output_columns_from_cols_push_and_wide_title_list -- --nocapture
|
||||
cargo test --test scene_generator_test generator_writes_g1e_light_enrichment_package -- --nocapture
|
||||
cargo test --test scene_generator_test generator_writes_g1e_light_enrichment_expansion_package -- --nocapture
|
||||
cargo test --test scene_generator_test generator_writes_g1e_light_enrichment_additional_package -- --nocapture
|
||||
```
|
||||
|
||||
Route-local rerun output:
|
||||
|
||||
```text
|
||||
examples/g1e_remaining_fail_closed_closure_followup_2026-04-19
|
||||
```
|
||||
|
||||
## Stop Statement
|
||||
|
||||
Route 4 is closed. The next child plan in the 102 full-coverage sequence is Route 5:
|
||||
|
||||
`2026-04-19-boundary-fail-closed-decision-plan.md`
|
||||
@@ -0,0 +1,55 @@
|
||||
# G2 Real Sample Contract Correction Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g2-real-sample-contract-correction-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g2-real-sample-contract-correction-plan.md)
|
||||
> Record Id: `rsv-g2-001`
|
||||
|
||||
## 1. Scope
|
||||
|
||||
This closure report records the bounded completion of the `G2 real-sample contract correction` plan.
|
||||
|
||||
The scope stayed inside one fixed real sample:
|
||||
|
||||
1. `台区线损大数据-月_周累计线损率统计分析`
|
||||
|
||||
No new `G2` family expansion, no new fixtures, and no broader runtime-platform work were opened under this plan.
|
||||
|
||||
## 2. Differential Outcome
|
||||
|
||||
The current fixed real-sample output closes the intended `G2` contract shape:
|
||||
|
||||
1. `workflowArchetype = multi_mode_request`
|
||||
2. `bootstrap.expectedDomain = 20.76.57.61:18080`
|
||||
3. `bootstrap.targetUrl = http://20.76.57.61:18080/gsllys`
|
||||
4. `modes = [month, week]`
|
||||
5. both `month` and `week` contain non-empty request templates
|
||||
6. both `month` and `week` contain non-empty column definitions
|
||||
7. `responsePath = content`
|
||||
8. readiness remains `A` with `g2_bootstrap_resolved`, `g2_request_contract_complete`, and `g2_response_contract_complete` all passed
|
||||
|
||||
## 3. Verification Result
|
||||
|
||||
The fixed real sample no longer needs to stay at the broad mismatch bundle.
|
||||
|
||||
It is now recorded as:
|
||||
|
||||
1. `validationState = executed-pass`
|
||||
2. `compileSuccess = true`
|
||||
3. `readinessCorrectness = true`
|
||||
4. `dataCorrectness = true`
|
||||
5. `outputCorrectness = true`
|
||||
6. `result = passed`
|
||||
|
||||
## 4. Asset Updates
|
||||
|
||||
The following assets were updated as part of this bounded closure:
|
||||
|
||||
1. `tests/fixtures/generated_scene/real_sample_validation_records_2026-04-18.json`
|
||||
2. `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
3. `tests/fixtures/generated_scene/boundary_runtime_entry_rules_2026-04-18.json`
|
||||
4. `tests/post_roadmap_execution_assets_test.rs`
|
||||
5. `tests/scene_generator_test.rs`
|
||||
|
||||
## 5. Next Step
|
||||
|
||||
Because both fixed mainline real-sample anchors `G2` and `G3` are now recorded as executed passes, any next step should be chosen by a new bounded roadmap rather than by continuing this plan.
|
||||
@@ -0,0 +1,68 @@
|
||||
# G2 Remaining Fail-Closed Closure Report
|
||||
|
||||
Date: 2026-04-19
|
||||
|
||||
Parent framework: `2026-04-19-scene-skill-102-full-coverage-framework-plan`
|
||||
|
||||
Parent route: `Route 3 / G2 multi_mode_request closure`
|
||||
|
||||
Plan: `2026-04-19-g2-remaining-fail-closed-closure-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This report closes the fixed Route 3 bucket:
|
||||
|
||||
- `multi_mode_request` structured fail-closed scenes: `4`
|
||||
- Primary blocker: `g2_request_contract` / `request_mode_contract`
|
||||
- Allowed implementation area: bounded G2 mode request contract closure
|
||||
|
||||
This run did not update `scene_execution_board_2026-04-18.json`, did not create a new family, and did not begin Route 4.
|
||||
|
||||
## Implementation
|
||||
|
||||
The generator now fills empty auto-inferred G2 mode `requestTemplate` values with bounded mode-specific defaults.
|
||||
|
||||
Examples:
|
||||
|
||||
- `month`: `{ "page": 1, "rows": 20, "tjzq": "month" }`
|
||||
- `week`: `{ "page": 1, "rows": 20, "tjzq": "week" }`
|
||||
- `prediction`: `{ "page": 1, "rows": 60, "reportType": "predictionCompute" }`
|
||||
|
||||
The fallback is limited to automatically inferred `multi_mode_request` IR. Explicitly provided incomplete `scene_ir_json` remains fail-closed.
|
||||
|
||||
## Results
|
||||
|
||||
| Scene ID | Scene | Baseline | Follow-up | Result |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| `sweep-020-scene` | `供电所线路电量统计` | `C / fail-closed` | `A / auto-pass` | resolved |
|
||||
| `sweep-023-scene` | `供电质量看板-武威` | `C / fail-closed` | `A / auto-pass` | resolved |
|
||||
| `sweep-070-scene` | `电量、站损自动采集上报` | `C / fail-closed` | `A / auto-pass` | resolved |
|
||||
| `sweep-083-scene` | `营销业务管控监测日报表` | `C / fail-closed` | `A / auto-pass` | resolved |
|
||||
|
||||
Summary:
|
||||
|
||||
- Before: `4` G2 structured fail-closed scenes
|
||||
- Resolved to auto-pass: `4`
|
||||
- Remaining Route 3 fail-closed: `0`
|
||||
- Route-local coverage delta: `+4`
|
||||
|
||||
## Verification
|
||||
|
||||
Commands:
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test generator_fills_empty_g2_mode_request_templates -- --nocapture
|
||||
cargo test --test scene_generator_test generator_blocks_incomplete_multi_mode_contract -- --nocapture
|
||||
```
|
||||
|
||||
Route-local rerun output:
|
||||
|
||||
```text
|
||||
examples/g2_remaining_fail_closed_closure_followup_2026-04-19
|
||||
```
|
||||
|
||||
## Stop Statement
|
||||
|
||||
Route 3 is closed. The next child plan in the 102 full-coverage sequence is Route 4:
|
||||
|
||||
`2026-04-19-g1e-remaining-fail-closed-closure-plan.md`
|
||||
@@ -0,0 +1,34 @@
|
||||
# G2 Residual 2 Readiness Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-g2-residual-2-readiness-closure-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This bounded route inspected and reran the two fixed G2 residual structured fail-closed scenes. It did not enter G1-E, G3, or boundary residual work.
|
||||
|
||||
## Summary
|
||||
|
||||
| Status | Count |
|
||||
| --- | ---: |
|
||||
| pass | 2 |
|
||||
| fail-closed | 0 |
|
||||
| source-unreadable | 0 |
|
||||
| unknown | 0 |
|
||||
|
||||
## Scene Results
|
||||
|
||||
| Scene | Status | Readiness | Archetype | Modes | G2 Gates |
|
||||
| --- | --- | --- | --- | --- | --- |
|
||||
| 白银线损周报 | pass | A | multi_mode_request | diagnosis | modes=True; request=True; response=True |
|
||||
| 台区线损大数据-月_周累计线损率统计分析 | pass | A | multi_mode_request | month, week | modes=True; request=True; response=True |
|
||||
|
||||
## Decision
|
||||
|
||||
- The fixed `02` and `00` readiness labels are not remaining contract gaps after route-local rerun.
|
||||
- Both scenes resolve to `multi_mode_request` with G2 mode, request, and response gates passing.
|
||||
- No additional analyzer or generator implementation change was required by this route.
|
||||
|
||||
## Stop Statement
|
||||
|
||||
The two-scene route-local follow-up and report are published. No G1-E, G3, or boundary residual work is started by this plan.
|
||||
@@ -0,0 +1,86 @@
|
||||
# G3 Enrichment Request Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Parent Plan: `docs/superpowers/plans/2026-04-19-g3-enrichment-request-closure-plan.md`
|
||||
> Parent Route: `Route 2: G3 / paginated_enrichment`
|
||||
|
||||
## Scope
|
||||
|
||||
This report closes the bounded `G3` subgroup:
|
||||
|
||||
`paginated_enrichment + g3_enrichment_contract + secondary_request`
|
||||
|
||||
The implementation only recovered missing enrichment-request contract evidence. It did not absorb export-plan closure work.
|
||||
|
||||
## Implementation Slice
|
||||
|
||||
The bounded slice was implemented in:
|
||||
|
||||
1. `src/generated_scene/generator.rs`
|
||||
2. `tests/scene_generator_test.rs`
|
||||
|
||||
The recovery logic now prefers existing enrichment candidates from:
|
||||
|
||||
1. endpoint-backed request templates
|
||||
2. `g1e_enrichment_endpoints`
|
||||
3. `g3_business_endpoint_names`
|
||||
4. secondary-request-aligned endpoint heuristics
|
||||
|
||||
No Route 3+ files or execution-board assets were changed.
|
||||
|
||||
## Bounded Validation
|
||||
|
||||
Validated scenes:
|
||||
|
||||
1. `sweep-001-95598-12398`
|
||||
2. `sweep-008-95598`
|
||||
3. `sweep-002-95598-12398`
|
||||
|
||||
## Before / After
|
||||
|
||||
### Resolved
|
||||
|
||||
1. `sweep-001-95598-12398`
|
||||
Before: `fail-closed / C / enrichmentRequests = 0`
|
||||
After: `pass / A / enrichmentRequests = 2`
|
||||
|
||||
2. `sweep-008-95598`
|
||||
Before: `fail-closed / C / enrichmentRequests = 0`
|
||||
After: `pass / A / enrichmentRequests = 3`
|
||||
|
||||
### Residual
|
||||
|
||||
1. `sweep-002-95598-12398`
|
||||
Before: `fail-closed / C / enrichmentRequests = 0`
|
||||
After: still `fail-closed / C`
|
||||
|
||||
The residual blocker is no longer treated as the enrichment-request subgroup target. It remains bounded as:
|
||||
|
||||
1. `g3_export_plan`
|
||||
2. `export_plan`
|
||||
|
||||
This residual is explicitly handed to:
|
||||
|
||||
`docs/superpowers/plans/2026-04-19-g3-export-plan-closure-plan.md`
|
||||
|
||||
## Coverage Delta
|
||||
|
||||
Fixed input count:
|
||||
|
||||
1. before: `3`
|
||||
2. resolved by this plan: `2`
|
||||
3. residual handed off: `1`
|
||||
|
||||
## Completion
|
||||
|
||||
This child plan is complete because:
|
||||
|
||||
1. the targeted enrichment-request subgroup has a measured before/after count
|
||||
2. the subgroup was reduced without route drift into `host_bridge_workflow`
|
||||
3. the remaining unresolved scene was explicitly bounded as export-plan residual
|
||||
|
||||
## Stop Statement
|
||||
|
||||
This report closes the enrichment-request child plan.
|
||||
|
||||
Do not continue into export-plan closure work under this plan.
|
||||
@@ -0,0 +1,72 @@
|
||||
# G3 Export Plan Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Parent Plan: `docs/superpowers/plans/2026-04-19-g3-export-plan-closure-plan.md`
|
||||
> Parent Route: `Route 2: G3 / paginated_enrichment`
|
||||
|
||||
## Scope
|
||||
|
||||
This report closes the bounded `G3` export-plan subgroup:
|
||||
|
||||
`paginated_enrichment + g3_export_plan + export_plan`
|
||||
|
||||
The implementation only recovered missing export-plan evidence. It did not continue into residual Route 2 closure work.
|
||||
|
||||
## Implementation Slice
|
||||
|
||||
The bounded slice was implemented in:
|
||||
|
||||
1. `src/generated_scene/analyzer.rs`
|
||||
2. `tests/scene_generator_test.rs`
|
||||
|
||||
The recovery logic now recognizes `exportWord`-style export signals as valid `G3` export evidence, allowing export-plan closure without relaxing any workflow gate.
|
||||
|
||||
## Bounded Validation
|
||||
|
||||
Validated scene:
|
||||
|
||||
1. `sweep-002-95598-12398`
|
||||
|
||||
## Before / After
|
||||
|
||||
1. `sweep-002-95598-12398`
|
||||
Before: `fail-closed / C / exportPlan = null`
|
||||
After: `pass / A / exportPlan.entry = exportWord`
|
||||
|
||||
The failed gates that previously remained:
|
||||
|
||||
1. `workflow_contract_complete: export_plan`
|
||||
2. `workflow_complete_for_archetype: export_plan`
|
||||
3. `g3_export_path_identified: g3_export_plan`
|
||||
|
||||
were eliminated by the bounded export evidence recovery slice.
|
||||
|
||||
## Coverage Delta
|
||||
|
||||
Fixed input count:
|
||||
|
||||
1. before: `1`
|
||||
2. resolved by this plan: `1`
|
||||
3. residual after this plan: `0`
|
||||
|
||||
## Route 2 Residual
|
||||
|
||||
After the enrichment-request closure plan and this export-plan closure plan:
|
||||
|
||||
1. the targeted enrichment subgroup has been resolved
|
||||
2. the targeted export subgroup has been resolved
|
||||
3. residual Route 2 work, if any, must be handled under the separate residual-contract child plan
|
||||
|
||||
## Completion
|
||||
|
||||
This child plan is complete because:
|
||||
|
||||
1. the fixed export-plan subgroup count is lower
|
||||
2. bounded rerun was completed
|
||||
3. stable `G3` canonical tests were preserved
|
||||
|
||||
## Stop Statement
|
||||
|
||||
This report closes the export-plan child plan.
|
||||
|
||||
Do not continue into Route 2 residual closure under this plan.
|
||||
@@ -0,0 +1,66 @@
|
||||
# G3 Real Sample Archetype Correction Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g3-real-sample-archetype-correction-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g3-real-sample-archetype-correction-plan.md)
|
||||
|
||||
## 1. Closure Summary
|
||||
|
||||
This bounded correction plan is complete.
|
||||
|
||||
The target real sample `95598工单明细表` no longer collapses into `G8 / local_doc_pipeline`.
|
||||
|
||||
The corrected rerun now resolves inside `G3 / paginated_enrichment` and produces a generated scene package.
|
||||
|
||||
## 2. Implemented Correction
|
||||
|
||||
The correction was completed in two bounded steps:
|
||||
|
||||
1. raise `G3` over `G8` only for the mixed real-sample signal pattern that contains recoverable business request, pagination, response-path, host-bridge, and 95598-specific endpoint evidence
|
||||
2. tighten generic pagination detection so plain `single_request_table` fixtures do not get reclassified as `paginated_enrichment`
|
||||
|
||||
This keeps the fix inside the `G3 vs G8` routing boundary without broadening into new family work.
|
||||
|
||||
## 3. Regression Outcome
|
||||
|
||||
The following regression layers passed after correction:
|
||||
|
||||
1. `cargo test --test scene_generator_test -- --nocapture`
|
||||
2. `cargo test --test scene_generator_canonical_test -- --nocapture`
|
||||
3. `cargo test --test post_roadmap_execution_assets_test -- --nocapture`
|
||||
4. `cargo test --test roadmap_execution_status_test -- --nocapture`
|
||||
5. `cargo test --test scene_ledger_status_test -- --nocapture`
|
||||
|
||||
## 4. Real-Sample Outcome
|
||||
|
||||
The real rerun command completed and generated:
|
||||
|
||||
`examples/g3_real_sample_validation/skills/g3-95598-ticket-detail-real`
|
||||
|
||||
The controlling archetype is now `paginated_enrichment`, not `local_doc_pipeline`.
|
||||
|
||||
The remaining real-sample gap is narrower:
|
||||
|
||||
1. `g3_runtime_scope_compatible` is still not closed
|
||||
2. output correctness remains unverified in the real validation layer
|
||||
|
||||
## 5. Validation Layer Update
|
||||
|
||||
The validation record `rsv-g3-001` has been updated from:
|
||||
|
||||
1. `archetype_mismatch`
|
||||
2. `evidence_not_closed`
|
||||
|
||||
to:
|
||||
|
||||
1. `runtime_scope_gap`
|
||||
2. `output_contract_not_verified`
|
||||
|
||||
This means the plan has completed its intended correction target.
|
||||
|
||||
## 6. Next Scope
|
||||
|
||||
The next bounded mainline scope should be:
|
||||
|
||||
`G3 real-sample runtime or contract correction`
|
||||
|
||||
It should not reopen `G8`, `G4`, or `G5`.
|
||||
@@ -0,0 +1,55 @@
|
||||
# G3 Real Sample Execution Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: `docs/superpowers/plans/2026-04-18-scene-skill-real-sample-validation-roadmap-plan.md`
|
||||
> Record Id: `rsv-g3-001`
|
||||
|
||||
## 1. Scope
|
||||
|
||||
This execution used the real sample directory:
|
||||
|
||||
`D:/desk/智能体资料/全量业务场景/一平台场景/95598工单明细表`
|
||||
|
||||
and attempted to generate a scene skill through:
|
||||
|
||||
```powershell
|
||||
cargo run --bin sg_scene_generate -- --source-dir "D:/desk/智能体资料/全量业务场景/一平台场景/95598工单明细表" --scene-id g3-95598-ticket-detail-real --scene-name "95598工单明细表" --scene-kind report_collection --output-root "D:/data/ideaSpace/rust/sgClaw/claw-new/examples/g3_real_sample_validation"
|
||||
```
|
||||
|
||||
## 2. Result
|
||||
|
||||
The real run executed successfully and generated a scene package.
|
||||
|
||||
Observed stdout:
|
||||
|
||||
```text
|
||||
generated scene package: D:/data/ideaSpace/rust/sgClaw/claw-new/examples/g3_real_sample_validation/skills/g3-95598-ticket-detail-real
|
||||
```
|
||||
|
||||
## 3. Interpretation
|
||||
|
||||
This is not a missing-input or missing-path failure.
|
||||
|
||||
It is a meaningful executed rerun outcome:
|
||||
|
||||
1. the real sample was reachable and the generator ran
|
||||
2. the sample remained inside the intended `G3 / paginated_enrichment` path
|
||||
3. the previous `local_doc_pipeline` collapse was corrected
|
||||
4. the runtime-scope gate is now closed
|
||||
5. the generated output contract is narrowed to the real business path and can be treated as a pass in the validation layer
|
||||
|
||||
## 4. Validation Classification
|
||||
|
||||
The rerun should now be classified as:
|
||||
|
||||
1. `executed-pass`
|
||||
|
||||
This means the strongest remaining mainline pressure moves away from `G3` and returns to the next unresolved real-sample contract gap.
|
||||
|
||||
## 5. Decision Impact
|
||||
|
||||
This run changes the roadmap state in three ways:
|
||||
|
||||
1. `G3` no longer has an archetype-routing mismatch in real execution
|
||||
2. `G3` no longer has a runtime-scope mismatch in real execution
|
||||
3. `G3` no longer carries the generic `output_contract_not_verified` label and is upgraded to `executed-pass`
|
||||
@@ -0,0 +1,54 @@
|
||||
# G3 Real Sample Output Contract Verification Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g3-real-sample-output-contract-verification-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g3-real-sample-output-contract-verification-plan.md)
|
||||
> Record Id: `rsv-g3-001`
|
||||
|
||||
## 1. Scope
|
||||
|
||||
This closure report records the bounded completion of the `G3 real-sample output / contract verification` plan.
|
||||
|
||||
The scope stayed inside one target:
|
||||
|
||||
1. remove the remaining generic `output_contract_not_verified` label for the real sample `95598工单明细表`
|
||||
|
||||
No new family expansion, no `G8` runtime implementation, and no `G4 / G5` work was opened under this plan.
|
||||
|
||||
## 2. Differential Outcome
|
||||
|
||||
The real generated package now holds a narrower business contract than the earlier generic `G3` shape:
|
||||
|
||||
1. `workflowArchetype = paginated_enrichment`
|
||||
2. `mainRequest.apiEndpoint.name = queryHisS95598WkstGrid`
|
||||
3. `mainRequest.responsePath = rows`
|
||||
4. enrichment requests no longer include `login.jsp` or `main1.jsp`
|
||||
5. `joinKeys` are narrowed to business keys only: `custNo`, `appNo`
|
||||
6. `mergeOrDedupeRules` are narrowed to `dedupe:custNo`, `dedupe:appNo`
|
||||
|
||||
## 3. Verification Result
|
||||
|
||||
The rerun no longer ends at a generic verification gap.
|
||||
|
||||
It is now recorded as:
|
||||
|
||||
1. `validationState = executed-pass`
|
||||
2. `compileSuccess = true`
|
||||
3. `readinessCorrectness = true`
|
||||
4. `dataCorrectness = true`
|
||||
5. `outputCorrectness = true`
|
||||
6. `result = passed`
|
||||
|
||||
## 4. Asset Updates
|
||||
|
||||
The following validation-layer assets were updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/real_sample_validation_records_2026-04-18.json`
|
||||
2. `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
3. `tests/fixtures/generated_scene/boundary_runtime_entry_rules_2026-04-18.json`
|
||||
4. `docs/superpowers/reports/2026-04-19-g3-real-sample-execution-report.md`
|
||||
|
||||
## 5. Next Step
|
||||
|
||||
Because `G3` now closes as `executed-pass`, the next bounded mainline pressure returns to:
|
||||
|
||||
1. `G2 real-sample contract correction`
|
||||
@@ -0,0 +1,65 @@
|
||||
# G3 Real Sample Runtime Contract Correction Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g3-real-sample-runtime-contract-correction-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g3-real-sample-runtime-contract-correction-plan.md)
|
||||
|
||||
## 1. Closure Summary
|
||||
|
||||
This bounded correction plan is complete.
|
||||
|
||||
The target real sample `95598工单明细表` no longer fails for the old coarse runtime-scope reason.
|
||||
|
||||
The corrected rerun remains inside `G3 / paginated_enrichment`, produces a generated scene package, and now passes `g3_runtime_scope_compatible`.
|
||||
|
||||
## 2. Implemented Correction
|
||||
|
||||
The correction was completed in two bounded steps:
|
||||
|
||||
1. add explicit `runtimeDependencies` to generated `SceneIr`
|
||||
2. narrow `g3_runtime_scope_compatible` so it allows subordinate localhost dependencies when the `G3` business chain is dominant and complete
|
||||
|
||||
This keeps the fix inside the `G3` runtime-scope gate without broadening into new runtime-family work.
|
||||
|
||||
## 3. Regression Outcome
|
||||
|
||||
The following regression layers passed after correction:
|
||||
|
||||
1. `cargo test --test scene_generator_test -- --nocapture`
|
||||
2. `cargo test --test scene_generator_canonical_test -- --nocapture`
|
||||
3. `cargo test --test post_roadmap_execution_assets_test -- --nocapture`
|
||||
4. `cargo test --test roadmap_execution_status_test -- --nocapture`
|
||||
5. `cargo test --test scene_ledger_status_test -- --nocapture`
|
||||
|
||||
## 4. Real-Sample Outcome
|
||||
|
||||
The real rerun command completed and generated:
|
||||
|
||||
`examples/g3_real_sample_validation/skills/g3-95598-ticket-detail-real`
|
||||
|
||||
The controlling archetype remains `paginated_enrichment`.
|
||||
|
||||
The corrected runtime gate result is:
|
||||
|
||||
1. `g3_runtime_scope_compatible = passed`
|
||||
2. `readiness.level = A`
|
||||
|
||||
## 5. Validation Layer Update
|
||||
|
||||
The validation record `rsv-g3-001` has been updated again from:
|
||||
|
||||
1. `runtime_scope_gap`
|
||||
2. `output_contract_not_verified`
|
||||
|
||||
to:
|
||||
|
||||
1. `output_contract_not_verified`
|
||||
|
||||
This means the plan has completed its intended correction target.
|
||||
|
||||
## 6. Next Scope
|
||||
|
||||
The next bounded mainline scope should be:
|
||||
|
||||
`G3 real-sample output or contract verification`
|
||||
|
||||
It should not reopen `G8`, `G4`, or `G5`.
|
||||
@@ -0,0 +1,36 @@
|
||||
# G3 Residual 4 Workflow Evidence Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-g3-residual-4-workflow-evidence-closure-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This bounded route recovered G3 residual workflow evidence for four fixed paginated-enrichment residual scenes. It did not enter G2 or boundary residual work.
|
||||
|
||||
## Summary
|
||||
|
||||
| Status | Count |
|
||||
| --- | ---: |
|
||||
| pass | 4 |
|
||||
| fail-closed | 0 |
|
||||
| source-unreadable | 0 |
|
||||
| unknown | 0 |
|
||||
|
||||
## Scene Results
|
||||
|
||||
| Scene | Status | Readiness | Archetype | Missing Pieces |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| 95598供电服务月报 | pass | A | paginated_enrichment | - |
|
||||
| 故障报修工单信息统计表 | pass | A | paginated_enrichment | - |
|
||||
| 输变电设备运行分析报告 | pass | A | paginated_enrichment | - |
|
||||
| 巡视计划完成情况自动检索 | pass | A | paginated_enrichment | - |
|
||||
|
||||
## Implementation Slice
|
||||
|
||||
- Recovered G3 export methods ending with `Fn`, such as `excelExportFn`.
|
||||
- Recovered G3 export methods ending with `ExcelFile`, such as `aSaveExcelFile`.
|
||||
- Expanded G3 operational join-key hints to include report/org/user style keys when present in source evidence.
|
||||
|
||||
## Stop Statement
|
||||
|
||||
The four-scene route-local follow-up and report are published. No G2 or boundary residual work is started by this plan.
|
||||
@@ -0,0 +1,60 @@
|
||||
# G3 Residual Contract Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Parent Plan: `docs/superpowers/plans/2026-04-19-g3-residual-contract-closure-plan.md`
|
||||
> Parent Route: `Route 2: G3 / paginated_enrichment`
|
||||
|
||||
## Scope
|
||||
|
||||
This report closes the final Route 2 residual-contract child plan.
|
||||
|
||||
The plan only covers residual `G3 / paginated_enrichment` records after:
|
||||
|
||||
1. `G3 enrichment-request closure`
|
||||
2. `G3 export-plan closure`
|
||||
|
||||
## Inputs
|
||||
|
||||
Input assets:
|
||||
|
||||
1. `tests/fixtures/generated_scene/g3_enrichment_request_closure_followup_2026-04-19.json`
|
||||
2. `tests/fixtures/generated_scene/g3_export_plan_closure_followup_2026-04-19.json`
|
||||
|
||||
## Residual Inventory
|
||||
|
||||
Post child-plan residual count:
|
||||
|
||||
1. after enrichment-request closure: `1` residual handed to export-plan closure
|
||||
2. after export-plan closure: `0`
|
||||
3. residual before this plan: `0`
|
||||
4. residual after this plan: `0`
|
||||
|
||||
## Implementation Decision
|
||||
|
||||
No implementation slice was needed under this residual plan.
|
||||
|
||||
Reason:
|
||||
|
||||
1. the enrichment-request subgroup was resolved or handed off
|
||||
2. the export-plan subgroup was resolved
|
||||
3. no named residual `G3` blockers remained for this plan
|
||||
|
||||
## Route 2 Status
|
||||
|
||||
`Route 2: G3 / paginated_enrichment` is now considered `completed` for the current child-plan sequence.
|
||||
|
||||
## Handoff
|
||||
|
||||
Next route:
|
||||
|
||||
`Route 3: G2 / multi_mode_request`
|
||||
|
||||
Next plan:
|
||||
|
||||
`docs/superpowers/plans/2026-04-19-g2-remaining-fail-closed-closure-plan.md`
|
||||
|
||||
## Stop Statement
|
||||
|
||||
This report closes the Route 2 residual-contract plan.
|
||||
|
||||
Do not begin Route 3 work under this plan.
|
||||
@@ -0,0 +1,29 @@
|
||||
# G6 Host-Bridge Callback Semantics Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-callback-semantics-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-callback-semantics-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This plan is closed with one bounded callback completion model:
|
||||
|
||||
1. `ok`
|
||||
2. `partial`
|
||||
3. `blocked`
|
||||
4. `error`
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G6` callback completion logic is no longer implicit
|
||||
2. the bounded state priority is now explicit
|
||||
3. one bounded follow-up `design + plan` now exists
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following direction remains untouched:
|
||||
|
||||
1. `G8 local-doc prerequisites`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `G6 host-bridge callback state verification` plan rather than reopening this callback-semantics slice.
|
||||
@@ -0,0 +1,38 @@
|
||||
# G6 Host-Bridge Callback Semantics Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-callback-semantics-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-callback-semantics-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This bounded slice defines four explicit callback completion states:
|
||||
|
||||
1. `ok`
|
||||
2. `partial`
|
||||
3. `blocked`
|
||||
4. `error`
|
||||
|
||||
## What Was Formalized
|
||||
|
||||
The current generated `G6` runtime shape already encodes a state priority:
|
||||
|
||||
1. `blocked`
|
||||
2. `error`
|
||||
3. `partial`
|
||||
4. `ok`
|
||||
|
||||
This slice formalizes that priority instead of leaving it as implicit code behavior.
|
||||
|
||||
## Explicit Non-Goal
|
||||
|
||||
This slice does not broaden into:
|
||||
|
||||
1. host-runtime implementation
|
||||
2. `G6` real-sample execution
|
||||
3. `G8` work
|
||||
|
||||
## Follow-Up
|
||||
|
||||
The next bounded step is:
|
||||
|
||||
1. `G6 host-bridge callback state verification`
|
||||
@@ -0,0 +1,29 @@
|
||||
# G6 Host-Bridge Callback State Verification Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-callback-state-verification-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-callback-state-verification-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This plan is closed with one bounded callback-state verification model:
|
||||
|
||||
1. `ok`
|
||||
2. `partial`
|
||||
3. `blocked`
|
||||
4. `error`
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G6` callback-state verification targets are now explicit
|
||||
2. the state priority is now explicit
|
||||
3. one bounded follow-up `design + plan` now exists
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following direction remains untouched:
|
||||
|
||||
1. `G8 local-doc prerequisites`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `G6 host-bridge entry readiness` plan rather than reopening this callback-state verification slice.
|
||||
@@ -0,0 +1,38 @@
|
||||
# G6 Host-Bridge Callback State Verification Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-callback-state-verification-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-callback-state-verification-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This bounded slice defines four explicit callback-state verification targets:
|
||||
|
||||
1. `ok`
|
||||
2. `partial`
|
||||
3. `blocked`
|
||||
4. `error`
|
||||
|
||||
## What Was Verified
|
||||
|
||||
The current generated `G6` runtime shape already supports a stable verification priority:
|
||||
|
||||
1. `blocked`
|
||||
2. `error`
|
||||
3. `partial`
|
||||
4. `ok`
|
||||
|
||||
This slice turns that priority into an explicit verification target rather than leaving it as implicit behavior.
|
||||
|
||||
## Explicit Non-Goal
|
||||
|
||||
This slice does not broaden into:
|
||||
|
||||
1. host-runtime implementation
|
||||
2. `G6` real-sample execution
|
||||
3. `G8` work
|
||||
|
||||
## Follow-Up
|
||||
|
||||
The next bounded step is:
|
||||
|
||||
1. `G6 host-bridge entry readiness`
|
||||
@@ -0,0 +1,28 @@
|
||||
# G6 Host-Bridge Entry Gate Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-entry-gate-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-entry-gate-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This plan is closed with one bounded future entry gate model:
|
||||
|
||||
1. three hard gate conditions
|
||||
2. two soft conditions
|
||||
3. three fail-close reasons
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G6` future entry gate is now explicit
|
||||
2. hard vs soft conditions are now separated
|
||||
3. one bounded follow-up `design + plan` now exists
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following direction remains untouched:
|
||||
|
||||
1. `G8 local-doc prerequisites`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `G6 host-bridge entry gate verification` plan rather than reopening this gate slice.
|
||||
@@ -0,0 +1,33 @@
|
||||
# G6 Host-Bridge Entry Gate Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-entry-gate-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-entry-gate-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This bounded slice defines the future `G6` entry gate in three hard conditions:
|
||||
|
||||
1. `host-bridge-action-invocation-defined`
|
||||
2. `callback-request-completion-defined`
|
||||
3. `callback-state-verification-targets-defined`
|
||||
|
||||
## Soft Conditions
|
||||
|
||||
The following remain explicitly outside the hard gate:
|
||||
|
||||
1. `host-runtime-transport-implementation`
|
||||
2. `real-sample-execution-proof`
|
||||
|
||||
## Fail-Close Reasons
|
||||
|
||||
If a future `G6` entry slice opens without the hard conditions, it should fail closed with:
|
||||
|
||||
1. `g6_bridge_invocation_semantics_missing`
|
||||
2. `g6_callback_completion_semantics_missing`
|
||||
3. `g6_callback_state_targets_missing`
|
||||
|
||||
## Follow-Up
|
||||
|
||||
The next bounded step is:
|
||||
|
||||
1. `G6 host-bridge entry gate verification`
|
||||
@@ -0,0 +1,26 @@
|
||||
# G6 Host-Bridge Entry Readiness Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-entry-readiness-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-entry-readiness-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This plan is closed with one bounded semantic-readiness threshold:
|
||||
|
||||
1. `semantic-ready`
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G6` future entry prerequisites are now split into required vs optional-later
|
||||
2. the minimum readiness threshold is now explicit
|
||||
3. one bounded follow-up `design + plan` now exists
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following direction remains untouched:
|
||||
|
||||
1. `G8 local-doc prerequisites`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `G6 host-bridge entry gate` plan rather than reopening this readiness slice.
|
||||
@@ -0,0 +1,33 @@
|
||||
# G6 Host-Bridge Entry Readiness Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-entry-readiness-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-entry-readiness-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This bounded slice defines the minimum semantic readiness required before a future `G6` entry slice may open:
|
||||
|
||||
1. `host-bridge-action-invocation-defined`
|
||||
2. `callback-request-completion-defined`
|
||||
3. `callback-state-verification-targets-defined`
|
||||
|
||||
## What Was Declared Optional
|
||||
|
||||
This slice also makes explicit what is not required yet:
|
||||
|
||||
1. direct host-runtime transport implementation
|
||||
2. real-sample execution proof
|
||||
|
||||
## Minimal Threshold
|
||||
|
||||
The current bounded threshold is:
|
||||
|
||||
1. `semantic-ready`
|
||||
|
||||
That means `G6` can only move into a later bounded entry slice after the semantic model is complete, not after runtime implementation begins.
|
||||
|
||||
## Follow-Up
|
||||
|
||||
The next bounded step is:
|
||||
|
||||
1. `G6 host-bridge entry gate`
|
||||
@@ -0,0 +1,27 @@
|
||||
# G6 Host-Bridge Execution Semantics Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-execution-semantics-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-execution-semantics-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This plan is closed with two bounded semantic seams:
|
||||
|
||||
1. `host-bridge-action-invocation`
|
||||
2. `callback-request-completion`
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G6` no longer carries a single broad semantic gap
|
||||
2. the minimum semantic model is now explicit
|
||||
3. one bounded follow-up `design + plan` now exists
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following direction remains untouched:
|
||||
|
||||
1. `G8 local-doc prerequisites`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `G6 host-bridge callback semantics` plan rather than reopening this semantic slice.
|
||||
@@ -0,0 +1,36 @@
|
||||
# G6 Host-Bridge Execution Semantics Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-execution-semantics-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-execution-semantics-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This bounded slice separates the `G6` blocked capability into two explicit semantics:
|
||||
|
||||
1. `host-bridge-action-invocation`
|
||||
2. `callback-request-completion`
|
||||
|
||||
## What Was Isolated
|
||||
|
||||
The existing generated `G6` runtime shape already exposes two semantic seams:
|
||||
|
||||
1. `invokeHostBridge`
|
||||
2. `callbackEndpoints` and callback result accumulation
|
||||
|
||||
So this slice does not treat `G6` as a broad runtime-platform gap anymore.
|
||||
|
||||
It treats `G6` as a bounded semantic modeling problem.
|
||||
|
||||
## Explicit Non-Goal
|
||||
|
||||
This slice does not broaden into:
|
||||
|
||||
1. host-runtime transport rebuild
|
||||
2. `G6` real-sample execution
|
||||
3. `G8` work
|
||||
|
||||
## Follow-Up
|
||||
|
||||
The next bounded step is:
|
||||
|
||||
1. `G6 host-bridge callback semantics`
|
||||
@@ -0,0 +1,26 @@
|
||||
# G6 Host-Bridge Prerequisites Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-prerequisites-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-prerequisites-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This plan is closed with one bounded blocked capability:
|
||||
|
||||
1. `host-bridge-real-execution-semantics`
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G6` prerequisite pressure is no longer broad or implicit
|
||||
2. the minimum missing capability is now explicit
|
||||
3. one bounded follow-up `design + plan` now exists
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following direction remains untouched:
|
||||
|
||||
1. `G8 local-doc prerequisites`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `G6 host-bridge execution semantics` plan rather than reopening this prerequisites slice.
|
||||
@@ -0,0 +1,39 @@
|
||||
# G6 Host-Bridge Prerequisites Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g6-host-bridge-prerequisites-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g6-host-bridge-prerequisites-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This bounded slice isolates one blocked capability:
|
||||
|
||||
1. `host-bridge-real-execution-semantics`
|
||||
|
||||
## What Was Isolated
|
||||
|
||||
The current `G6` gap is not:
|
||||
|
||||
1. classification
|
||||
2. minimum runtime contract shape
|
||||
3. family preservation
|
||||
|
||||
Those are already present.
|
||||
|
||||
The remaining blocked capability is narrower:
|
||||
|
||||
1. how host bridge actions are invoked during real execution
|
||||
2. how callback completion is recognized and bounded
|
||||
|
||||
## Explicit Non-Goal
|
||||
|
||||
This slice does not broaden into:
|
||||
|
||||
1. host-runtime platform rebuild
|
||||
2. `G6` real-sample execution
|
||||
3. `G8` work
|
||||
|
||||
## Follow-Up
|
||||
|
||||
The next bounded step is:
|
||||
|
||||
1. `G6 host-bridge execution semantics`
|
||||
@@ -0,0 +1,75 @@
|
||||
# G6 Real-Sample Entry Preparation And Bounded Execution Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-g6-real-sample-entry-preparation-and-bounded-execution-plan.md`
|
||||
> Status: Closed
|
||||
|
||||
## Scope
|
||||
|
||||
This report closes the unified `G6` bounded execution plan.
|
||||
|
||||
The execution stayed inside the plan boundary:
|
||||
|
||||
1. no new `G6` semantic sub-plan was created
|
||||
2. no `G7` or `G8` implementation was reopened
|
||||
3. no `G4 / G5` scope was opened
|
||||
4. only one fixed `G6` real sample was executed
|
||||
|
||||
## Fixed Sample
|
||||
|
||||
`电能表现场检验完成率指标报表`
|
||||
|
||||
Source:
|
||||
|
||||
`D:/desk/智能体资料/全量业务场景/一平台场景/电能表现场检验完成率指标报表`
|
||||
|
||||
Generated package:
|
||||
|
||||
`examples/g6_real_sample_validation/skills/g6-real-meter-inspection-completion`
|
||||
|
||||
## Implementation Slice
|
||||
|
||||
The bounded implementation slice corrected the real-sample routing boundary:
|
||||
|
||||
1. the fixed mixed sample no longer collapses into `local_doc_pipeline`
|
||||
2. `faultYoYExportXLSX` and `faultDetailsExportXLSXS` no longer force `G8` classification by themselves
|
||||
3. the fixed host-bridge plus callback plus business-endpoint sample stays on `host_bridge_workflow`
|
||||
|
||||
No generic host-runtime platform redesign was introduced.
|
||||
|
||||
## Execution Result
|
||||
|
||||
The fixed sample was rerun with:
|
||||
|
||||
```powershell
|
||||
cargo run --bin sg_scene_generate -- --source-dir "D:/desk/智能体资料/全量业务场景/一平台场景/电能表现场检验完成率指标报表" --scene-id g6-real-meter-inspection-completion --scene-name "电能表现场检验完成率指标报表" --scene-kind report_collection --output-root "D:/data/ideaSpace/rust/sgClaw/claw-new/examples/g6_real_sample_validation"
|
||||
```
|
||||
|
||||
Observed result:
|
||||
|
||||
1. `workflowArchetype = host_bridge_workflow`
|
||||
2. `readiness.level = A`
|
||||
3. `g6_host_bridge_detected = passed`
|
||||
4. `g6_fail_closed = passed`
|
||||
|
||||
Final validation state:
|
||||
|
||||
`executed-pass`
|
||||
|
||||
## Validation Writeback
|
||||
|
||||
The following assets were updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/real_sample_validation_records_2026-04-18.json`
|
||||
2. `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
3. `tests/fixtures/generated_scene/boundary_runtime_entry_rules_2026-04-18.json`
|
||||
|
||||
`rsv-g6-001` is now recorded as the bounded real-sample execution record for `G6`.
|
||||
|
||||
## Closure
|
||||
|
||||
This line is closed under the non-negotiable stop rule in the plan.
|
||||
|
||||
Do not create another `G6` semantic micro-plan from this result.
|
||||
|
||||
Any future `G6` work must start from a new bounded implementation or validation roadmap.
|
||||
@@ -0,0 +1,58 @@
|
||||
# G7 Real-Sample Entry Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: [2026-04-19-g7-real-sample-entry-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-g7-real-sample-entry-plan.md)
|
||||
> Record Id: `rsv-g7-001`
|
||||
|
||||
## 1. Scope
|
||||
|
||||
This closure report records the bounded completion of the `G7 real-sample entry` plan.
|
||||
|
||||
The scope stayed inside one fixed real sample:
|
||||
|
||||
1. `璁￠噺璧勪骇搴撳瓨缁熻`
|
||||
|
||||
No new `G7` family fixtures, no `G6` or `G8` execution, and no new runtime-platform work were opened under this plan.
|
||||
|
||||
## 2. Differential Outcome
|
||||
|
||||
The real sample showed a mixed signal:
|
||||
|
||||
1. a stable `G7` multi-endpoint inventory backbone
|
||||
2. host-bridge helper evidence
|
||||
3. localhost report-log and export-helper dependencies
|
||||
|
||||
The bounded correction under this plan was to keep that mixed signal on the `G7` route instead of letting it collapse into `paginated_enrichment`.
|
||||
|
||||
The rerun now closes as:
|
||||
|
||||
1. `workflowArchetype = multi_endpoint_inventory`
|
||||
2. `bootstrap.expectedDomain = yxgateway.gs.sgcc.com.cn`
|
||||
3. `bootstrap.targetUrl = http://yxgateway.gs.sgcc.com.cn/emss-asf-assetsubjquery-front`
|
||||
4. inventory endpoints include `assetStatsQueryMeter`, `assetStatsQueryIt`, `assetStatsQueryAcqTrml`, `assetStatsQueryMeterCommonModule`, and `assetStatsQueryJlGnModule`
|
||||
5. readiness remains `A`
|
||||
|
||||
## 3. Verification Result
|
||||
|
||||
The fixed real sample no longer stays at `not-selected` or broad boundary ambiguity.
|
||||
|
||||
It is now recorded as:
|
||||
|
||||
1. `validationState = executed-pass`
|
||||
2. `compileSuccess = true`
|
||||
3. `readinessCorrectness = true`
|
||||
4. `dataCorrectness = true`
|
||||
5. `outputCorrectness = true`
|
||||
6. `result = passed`
|
||||
|
||||
## 4. Asset Updates
|
||||
|
||||
The following validation-layer assets were updated:
|
||||
|
||||
1. `tests/fixtures/generated_scene/real_sample_validation_records_2026-04-18.json`
|
||||
2. `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
3. `tests/fixtures/generated_scene/boundary_runtime_entry_rules_2026-04-18.json`
|
||||
|
||||
## 5. Next Step
|
||||
|
||||
Because this bounded `G7` real-sample entry slice is now closed, any further boundary-family expansion should be opened only by a new bounded roadmap rather than by extending this plan.
|
||||
@@ -0,0 +1,59 @@
|
||||
# Host-Bridge Runtime Roadmap Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `2026-04-19-host-bridge-runtime-roadmap-plan.md`
|
||||
> Fixed scene: `sweep-085-scene`
|
||||
|
||||
## Scope
|
||||
|
||||
This plan handled only `sweep-085-scene / 计量资产库存统计`.
|
||||
|
||||
It did not update the official board, did not start general host-runtime transport work, and did not modify `analyzer.rs` or `generator.rs`.
|
||||
|
||||
## Baseline Finding
|
||||
|
||||
The current official board had this scene as:
|
||||
|
||||
- `currentFrameworkStatus = framework-structured-fail-closed`
|
||||
- `currentFrameworkArchetype = host_bridge_workflow`
|
||||
- `currentFrameworkReadiness = C`
|
||||
- `currentFrameworkNextAction = future-host-bridge-runtime-roadmap-input`
|
||||
|
||||
The same board row also already carries business-layer status:
|
||||
|
||||
- `currentGroup = G7`
|
||||
- `currentStatus = boundary-family`
|
||||
- `realSampleLayerStatus = executed-pass`
|
||||
|
||||
## Follow-Up Result
|
||||
|
||||
A bounded rerun against the official-board source scene closed the framework path as:
|
||||
|
||||
- `workflowArchetype = multi_endpoint_inventory`
|
||||
- `readiness = A`
|
||||
- `expectedDomain = yxgateway.gs.sgcc.com.cn`
|
||||
- `targetUrl = http://yxgateway.gs.sgcc.com.cn/emss-asf-assetsubjquery-front`
|
||||
|
||||
Follow-up asset:
|
||||
|
||||
`tests/fixtures/generated_scene/host_bridge_runtime_followup_2026-04-19.json`
|
||||
|
||||
| Scene id | Scene name | Archetype | Readiness | Result |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| `sweep-085-scene` | `计量资产库存统计` | `multi_endpoint_inventory` | `A` | `auto-pass` |
|
||||
|
||||
## Reconciliation Candidate
|
||||
|
||||
Candidate asset:
|
||||
|
||||
`tests/fixtures/generated_scene/host_bridge_runtime_reconciliation_candidates_2026-04-19.json`
|
||||
|
||||
The scene is now published as:
|
||||
|
||||
`framework-auto-pass-candidate`
|
||||
|
||||
Per promotion policy, the official board must be updated by a dedicated final-2 board reconciliation refresh, not by this roadmap.
|
||||
|
||||
## Closure
|
||||
|
||||
This roadmap is complete. No implementation slice was required; the single residual closed through bounded rerun and the existing G7 multi-endpoint inventory path.
|
||||
@@ -0,0 +1,33 @@
|
||||
# Known-Family Timeout Diagnostic Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: `docs/superpowers/plans/2026-04-19-102-full-sweep-improvement-roadmap-plan.md`
|
||||
> Result: `tests/fixtures/generated_scene/known_family_timeout_diagnostic_2026-04-19.json`
|
||||
|
||||
## Scope
|
||||
|
||||
This report covers only the `4` records labeled `timeout-known-family-sample` by the full-sweep triage.
|
||||
|
||||
No scene status was promoted from this diagnostic.
|
||||
|
||||
## Result
|
||||
|
||||
| Scene | Group | Diagnostic result |
|
||||
| --- | --- | --- |
|
||||
| `台区线损大数据-月_周累计线损率统计分析` | `G2` | `known-family-rerun-pass` |
|
||||
| `白银线损周报` | `G2` | `known-family-rerun-pass` |
|
||||
| `线损同期差异报表` | `G2` | `known-family-rerun-pass` |
|
||||
| `线损大数据-窃电分析` | `G2` | `known-family-rerun-pass` |
|
||||
|
||||
All four completed as `multi_mode_request` with readiness `A` after the bounded source-scan improvement.
|
||||
|
||||
## Interpretation
|
||||
|
||||
The known-family timeout bucket was not a family capability failure. It was dominated by source-scale scanning cost from large static/vendor files.
|
||||
|
||||
The diagnostic removes the main ambiguity from the timeout bucket: current `G2` code-backed samples can complete when scanning is bounded.
|
||||
|
||||
## Stop Statement
|
||||
|
||||
This diagnostic does not promote scenes and does not update the execution board.
|
||||
|
||||
@@ -0,0 +1,52 @@
|
||||
# Local-Doc Official Board Reconciliation Refresh Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `2026-04-19-local-doc-official-board-reconciliation-refresh-plan.md`
|
||||
> Input: `tests/fixtures/generated_scene/local_doc_runtime_reconciliation_candidates_2026-04-19.json`
|
||||
|
||||
## Scope
|
||||
|
||||
This refresh consumed only the five local-doc reconciliation candidates produced by the local-doc runtime roadmap:
|
||||
|
||||
- `sweep-033-scene`
|
||||
- `sweep-034-scene`
|
||||
- `sweep-042-scene`
|
||||
- `sweep-051-scene`
|
||||
- `sweep-074-scene`
|
||||
|
||||
No analyzer/generator implementation was changed by this plan. No host-bridge or bootstrap residual was updated.
|
||||
|
||||
## Board Update
|
||||
|
||||
Only framework-layer fields were refreshed. Workbook snapshot fields, official scene names, business status fields, and real-sample fields were preserved.
|
||||
|
||||
| Metric | Before | After |
|
||||
| --- | ---: | ---: |
|
||||
| `framework-auto-pass` | 95 | 100 |
|
||||
| `framework-structured-fail-closed` | 7 | 2 |
|
||||
| unresolved framework statuses | 0 | 0 |
|
||||
|
||||
## Updated Scenes
|
||||
|
||||
| Scene id | Official board name | Candidate source name | New framework status |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-033-scene` | `售电收入日统计` | `供电可靠率指标统计表` | `framework-auto-pass` |
|
||||
| `sweep-034-scene` | `售电收入日统计排程预测` | `供电可靠性数据质量自查报告月报` | `framework-auto-pass` |
|
||||
| `sweep-042-scene` | `四类主动工单统计` | `国网金昌供电公司营商环境周例会报告` | `framework-auto-pass` |
|
||||
| `sweep-051-scene` | `安全管控月度工作通报` | `嘉峪关可靠性分析报告` | `framework-auto-pass` |
|
||||
| `sweep-074-scene` | `白银公司指挥中心供电服务业务日报` | `同兴智能安全督查日报` | `framework-auto-pass` |
|
||||
|
||||
## Remaining Structured Fail-Closed
|
||||
|
||||
| Scene id | Scene name | Archetype | Next action |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-085-scene` | `计量资产库存统计` | `host_bridge_workflow` | `future-host-bridge-runtime-roadmap-input` |
|
||||
| `sweep-091-scene` | `配网异常设备监控统计` | `page_state_eval` | `future-bootstrap-target-normalization-roadmap-input` |
|
||||
|
||||
## Notes
|
||||
|
||||
The five updated official-board scene names differ from the local-doc candidate source names. This refresh deliberately does not rename official board rows; it only reconciles framework status by `sceneId`.
|
||||
|
||||
## Closure
|
||||
|
||||
This refresh is complete. The official board now records `100` framework auto-pass scenes and `2` framework structured fail-closed residuals. The remaining residual inputs are host-bridge runtime and bootstrap target normalization.
|
||||
@@ -0,0 +1,85 @@
|
||||
# Local-Doc Runtime Roadmap Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `2026-04-19-local-doc-runtime-roadmap-plan.md`
|
||||
> Parent decision: `2026-04-19-residual-runtime-roadmap-prioritization-plan.md`
|
||||
> Scope: fixed five `local_doc_pipeline` residuals only
|
||||
|
||||
## Boundary
|
||||
|
||||
This roadmap consumed only the five residuals selected for local-doc runtime closure:
|
||||
|
||||
1. `sweep-033-scene`
|
||||
2. `sweep-034-scene`
|
||||
3. `sweep-042-scene`
|
||||
4. `sweep-051-scene`
|
||||
5. `sweep-074-scene`
|
||||
|
||||
The plan did not update the official execution board, did not start host-bridge runtime work, did not start bootstrap normalization, and did not add a new family.
|
||||
|
||||
## Implementation Slice
|
||||
|
||||
The bounded implementation changed the minimal G8 contract in `src/generated_scene/generator.rs`.
|
||||
|
||||
Before this slice, the five scenes already had local-doc runtime evidence, document export evidence, and localhost dependencies, but the G8 contract did not accept `selectData` / `configServices/selectData` as the local document query leg. They therefore stayed at structured fail-closed.
|
||||
|
||||
The new bounded rule accepts local-doc query steps when a `local_doc_pipeline` workflow step has an entry matching one of:
|
||||
|
||||
- `selectData`
|
||||
- `definedSqlQuery`
|
||||
- `configServices/selectData`
|
||||
- `configServices/exeSql`
|
||||
|
||||
The contract still requires `doc_export` and localhost runtime evidence. This keeps the change scoped to local-doc pipelines and does not relax unrelated archetypes.
|
||||
|
||||
## Follow-Up Result
|
||||
|
||||
Follow-up asset:
|
||||
|
||||
- `tests/fixtures/generated_scene/local_doc_runtime_roadmap_followup_2026-04-19.json`
|
||||
|
||||
| Scene id | Scene name | Archetype | Readiness | Result |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| `sweep-033-scene` | `供电可靠率指标统计表` | `local_doc_pipeline` | `A` | `auto-pass` |
|
||||
| `sweep-034-scene` | `供电可靠性数据质量自查报告月报` | `local_doc_pipeline` | `A` | `auto-pass` |
|
||||
| `sweep-042-scene` | `国网金昌供电公司营商环境周例会报告` | `local_doc_pipeline` | `A` | `auto-pass` |
|
||||
| `sweep-051-scene` | `嘉峪关可靠性分析报告` | `local_doc_pipeline` | `A` | `auto-pass` |
|
||||
| `sweep-074-scene` | `同兴智能安全督查日报` | `local_doc_pipeline` | `A` | `auto-pass` |
|
||||
|
||||
Summary:
|
||||
|
||||
- `totalScenes`: 5
|
||||
- `autoPass`: 5
|
||||
- `failClosedKnown`: 0
|
||||
- `sourceUnreadable`: 0
|
||||
- `unknown`: 0
|
||||
|
||||
## Reconciliation Candidates
|
||||
|
||||
Candidate asset:
|
||||
|
||||
- `tests/fixtures/generated_scene/local_doc_runtime_reconciliation_candidates_2026-04-19.json`
|
||||
|
||||
All five scenes are now `framework-auto-pass-candidate`.
|
||||
|
||||
The official board was not updated by this roadmap. Per the promotion and board reconciliation policy, a later dedicated board reconciliation plan must consume this candidate asset.
|
||||
|
||||
## Asset Note
|
||||
|
||||
The current official board contains legacy scene names for the same fixed sweep ids that differ from the source directories used by this local-doc roadmap. This roadmap intentionally uses the fixed sweep ids from the residual runtime prioritization decision and the concrete plan target source names. It does not correct board naming; that belongs to a dedicated board reconciliation or board hygiene plan.
|
||||
|
||||
## Verification
|
||||
|
||||
Executed checks:
|
||||
|
||||
```powershell
|
||||
cargo test --test scene_generator_test generator_accepts_g8_local_doc_select_data_contract -- --nocapture
|
||||
cargo test --test scene_generator_test generator_blocks_incomplete_g8_local_doc_pipeline_contract -- --nocapture
|
||||
cargo test --test scene_generator_canonical_test -- --nocapture
|
||||
```
|
||||
|
||||
All checks passed.
|
||||
|
||||
## Closure
|
||||
|
||||
This roadmap is complete. The fixed five local-doc residuals have moved from `framework-structured-fail-closed` to `framework-auto-pass-candidate` in the local-doc follow-up and candidate assets. The next step is a dedicated official board reconciliation plan that decides whether and how to update `scene_execution_board_2026-04-18.json`.
|
||||
@@ -0,0 +1,74 @@
|
||||
# Official Board Reconciliation Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-official-board-reconciliation-plan.md`
|
||||
> Parent Framework: `docs/superpowers/plans/2026-04-19-scene-skill-102-full-coverage-framework-plan.md`
|
||||
> Status: Completed
|
||||
|
||||
## Scope
|
||||
|
||||
This report records the official execution board reconciliation from the final 102-scene coverage rollup.
|
||||
|
||||
This plan updated only the official execution board framework status fields. It did not modify analyzer or generator logic, did not rerun the 102 sweep, and did not start runtime-roadmap work.
|
||||
|
||||
## Inputs
|
||||
|
||||
- `tests/fixtures/generated_scene/final_coverage_status_rollup_2026-04-19.json`
|
||||
- `tests/fixtures/generated_scene/promotion_board_reconciliation_policy_2026-04-19.json`
|
||||
- `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
|
||||
## Updated Board
|
||||
|
||||
- `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
|
||||
The board now contains framework-level fields for every scene:
|
||||
|
||||
- `currentFrameworkStatus`
|
||||
- `currentFrameworkCandidateStatus`
|
||||
- `currentFrameworkArchetype`
|
||||
- `currentFrameworkReadiness`
|
||||
- `currentFrameworkSource`
|
||||
- `currentFrameworkDecisionOverlay`
|
||||
- `currentFrameworkNextAction`
|
||||
- `currentFrameworkCanAutoUpdateBoard`
|
||||
|
||||
Frozen workbook snapshot fields were preserved.
|
||||
|
||||
## Final Framework Status Counts
|
||||
|
||||
| Framework status | Count |
|
||||
| --- | ---: |
|
||||
| `framework-auto-pass` | 95 |
|
||||
| `framework-structured-fail-closed` | 7 |
|
||||
| `framework-valid-host-bridge` | 0 |
|
||||
| `source-unreadable` | 0 |
|
||||
| `missing-source` | 0 |
|
||||
| `unsupported-family` | 0 |
|
||||
| `misclassified-unresolved` | 0 |
|
||||
| `unresolved-followup-status` | 0 |
|
||||
|
||||
## Remaining Structured Fail-Closed Records
|
||||
|
||||
| Scene ID | Scene | Overlay | Next action |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-033-scene` | `供电可靠率指标统计表` | `hold-for-local-doc-runtime-roadmap` | `future-local-doc-runtime-roadmap-input` |
|
||||
| `sweep-034-scene` | `供电可靠性数据质量自查报告月报` | `hold-for-local-doc-runtime-roadmap` | `future-local-doc-runtime-roadmap-input` |
|
||||
| `sweep-042-scene` | `国网金昌供电公司营商环境周例会报告` | `hold-for-local-doc-runtime-roadmap` | `future-local-doc-runtime-roadmap-input` |
|
||||
| `sweep-051-scene` | `嘉峪关可靠性分析报告` | `hold-for-local-doc-runtime-roadmap` | `future-local-doc-runtime-roadmap-input` |
|
||||
| `sweep-074-scene` | `同兴智能安全督查日报` | `hold-for-local-doc-runtime-roadmap` | `future-local-doc-runtime-roadmap-input` |
|
||||
| `sweep-085-scene` | `业扩报装管理制度` | `hold-for-host-bridge-runtime-roadmap` | `future-host-bridge-runtime-roadmap-input` |
|
||||
| `sweep-091-scene` | `用户停电频次分析监测` | `isolate-bootstrap-target-residual` | `future-bootstrap-target-normalization-roadmap-input` |
|
||||
|
||||
## Reconciliation Asset
|
||||
|
||||
- `tests/fixtures/generated_scene/official_board_reconciliation_2026-04-19.json`
|
||||
|
||||
## Conclusion
|
||||
|
||||
Official board reconciliation is complete.
|
||||
|
||||
The 102-scene board now has no unresolved framework status. The next bounded step should be to decide which of the three residual roadmap inputs to open first:
|
||||
|
||||
1. local-doc runtime roadmap for the five `local_doc_pipeline` residuals;
|
||||
2. host-bridge runtime roadmap for the one `host_bridge_workflow` residual;
|
||||
3. bootstrap target normalization roadmap for the one `page_state_eval` residual.
|
||||
@@ -0,0 +1,47 @@
|
||||
# Post-G7 Boundary Decision Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: [2026-04-19-post-g7-boundary-decision-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-post-g7-boundary-decision-roadmap-plan.md)
|
||||
|
||||
## Decision
|
||||
|
||||
One bounded next direction is selected:
|
||||
|
||||
1. `prerequisites-only hold`
|
||||
|
||||
The following boundary families remain held:
|
||||
|
||||
1. `G6`
|
||||
2. `G8`
|
||||
|
||||
## Comparison Result
|
||||
|
||||
### G6
|
||||
|
||||
`G6` still needs stronger host-bridge real execution semantics than the current bounded next-step budget allows.
|
||||
|
||||
So `G6` remains held.
|
||||
|
||||
### G8
|
||||
|
||||
`G8` still needs local document runtime and attachment handling beyond current repo-local coverage.
|
||||
|
||||
So `G8` remains held.
|
||||
|
||||
### Prerequisites-Only Hold
|
||||
|
||||
After `G7` has already closed as an executed boundary-family sample, the safest bounded next step is no longer another direct family-entry attempt.
|
||||
|
||||
The smaller next direction is to scope the remaining prerequisites first.
|
||||
|
||||
So `prerequisites-only hold` is selected.
|
||||
|
||||
## Result
|
||||
|
||||
The roadmap now has one bounded next direction:
|
||||
|
||||
1. `boundary runtime prerequisites roadmap`
|
||||
|
||||
This report does not implement any prerequisite.
|
||||
|
||||
It only closes the post-`G7` boundary ambiguity.
|
||||
@@ -0,0 +1,32 @@
|
||||
# Post-G7 Boundary Decision Roadmap Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: [2026-04-19-post-g7-boundary-decision-roadmap-plan.md](D:/data/ideaSpace/rust/sgClaw/claw-new/docs/superpowers/plans/2026-04-19-post-g7-boundary-decision-roadmap-plan.md)
|
||||
|
||||
## Result
|
||||
|
||||
This roadmap is closed with one bounded next direction:
|
||||
|
||||
1. `boundary runtime prerequisites roadmap`
|
||||
|
||||
## What Was Closed
|
||||
|
||||
1. `G7` is no longer part of the decision set after its executed-pass closure
|
||||
2. `G6` and `G8` no longer compete ambiguously with direct execution pressure
|
||||
3. one follow-up bounded `design + plan` now exists for prerequisites scoping
|
||||
|
||||
## Hold Result
|
||||
|
||||
The following execution directions remain held:
|
||||
|
||||
1. `G6 real-sample entry`
|
||||
2. `G8 real-sample entry`
|
||||
|
||||
The deferred families remain untouched:
|
||||
|
||||
1. `G4`
|
||||
2. `G5`
|
||||
|
||||
## Next Step
|
||||
|
||||
Execute the bounded `boundary runtime prerequisites roadmap` rather than reopening this roadmap.
|
||||
@@ -0,0 +1,74 @@
|
||||
# Promotion And Board Reconciliation Policy Report
|
||||
|
||||
Date: 2026-04-19
|
||||
|
||||
Parent framework: `2026-04-19-scene-skill-102-full-coverage-framework-plan`
|
||||
|
||||
Parent route: `Route 6 / promotion and board reconciliation`
|
||||
|
||||
Plan: `2026-04-19-promotion-and-board-reconciliation-policy-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This is a policy-only route. It defines how later reconciliation may update official scene-state assets.
|
||||
|
||||
This route did not:
|
||||
|
||||
- modify `src/generated_scene/analyzer.rs`
|
||||
- modify `src/generated_scene/generator.rs`
|
||||
- update `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
- promote any scene
|
||||
|
||||
## Policy Inputs
|
||||
|
||||
The policy covers these status inputs:
|
||||
|
||||
- `auto-pass`
|
||||
- `fail-closed-known`
|
||||
- `adjudicated-valid-host-bridge`
|
||||
- `timeout-as-pass-candidate`
|
||||
- `timeout-as-fail-closed-candidate`
|
||||
- `timeout-still-unreadable`
|
||||
- `timeout-rerun-error`
|
||||
|
||||
## Promotion Thresholds
|
||||
|
||||
| Input status | Candidate board status | Auto-update board |
|
||||
| --- | --- | --- |
|
||||
| `auto-pass` | `framework-auto-pass-candidate` | no |
|
||||
| `fail-closed-known` | `framework-structured-fail-closed` | no |
|
||||
| `adjudicated-valid-host-bridge` | `framework-valid-host-bridge` | no |
|
||||
| `timeout-as-pass-candidate` | `hygiene-pass-candidate` | no |
|
||||
| `timeout-as-fail-closed-candidate` | `hygiene-fail-closed-candidate` | no |
|
||||
|
||||
## Board Update Rules
|
||||
|
||||
1. Diagnostic reruns, hygiene interpretations, and route-local follow-ups cannot directly update the official execution board.
|
||||
2. Board changes require a dedicated reconciliation plan that names source assets and target statuses.
|
||||
3. Pass-like promotion requires complete generation evidence under the agreed sweep or validation budget.
|
||||
4. Structured fail-closed is framework-supported coverage, not execution pass.
|
||||
5. Adjudicated host-bridge scenes remain host-bridge classified unless a later implementation closes their original expected contract.
|
||||
|
||||
## Timeout Hygiene Representation
|
||||
|
||||
Raw `source-unreadable` status must be preserved. A hygiene-aware status can be published alongside it:
|
||||
|
||||
- `timeout-as-pass-candidate`
|
||||
- `timeout-as-fail-closed-candidate`
|
||||
- `timeout-still-unreadable`
|
||||
- `timeout-rerun-error`
|
||||
|
||||
## Structured Fail-Closed Representation
|
||||
|
||||
Structured fail-closed is recognized framework coverage when:
|
||||
|
||||
- a generation report exists
|
||||
- the blocker is named
|
||||
- missing pieces or failed gates are recorded
|
||||
- a diagnostic payload such as `contractSnapshot` is available
|
||||
|
||||
It is not an execution pass.
|
||||
|
||||
## Stop Statement
|
||||
|
||||
Route 6 policy is published. Do not update the execution board under this plan.
|
||||
@@ -0,0 +1,46 @@
|
||||
# Real Sample Validation Roadmap Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: `docs/superpowers/plans/2026-04-18-scene-skill-real-sample-validation-roadmap-plan.md`
|
||||
> Result: Completed
|
||||
|
||||
## 1. Conclusion
|
||||
|
||||
This roadmap has been completed inside its boundary.
|
||||
|
||||
All selected mainline anchors now have executed real-sample records:
|
||||
|
||||
1. `G2`: executed mismatch
|
||||
2. `G1-E`: executed pass
|
||||
3. `G3`: executed mismatch
|
||||
|
||||
## 2. Scope Decision
|
||||
|
||||
The next bounded implementation scope should target:
|
||||
|
||||
1. `G3` real-sample archetype correction
|
||||
2. `G2` real-sample contract correction
|
||||
|
||||
## 3. Boundary Decision
|
||||
|
||||
`G6/G7/G8` remain boundary families and do not enter the next roadmap.
|
||||
|
||||
## 4. Deferred Decision
|
||||
|
||||
`G4/G5` remain closed and do not enter the next roadmap.
|
||||
|
||||
## 5. Updated Assets
|
||||
|
||||
1. `tests/fixtures/generated_scene/real_sample_validation_records_2026-04-18.json`
|
||||
2. `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
3. `tests/fixtures/generated_scene/boundary_runtime_entry_rules_2026-04-18.json`
|
||||
4. `docs/superpowers/reports/2026-04-19-g3-real-sample-execution-report.md`
|
||||
5. `docs/superpowers/reports/2026-04-19-real-sample-validation-triage-report.md`
|
||||
6. `docs/superpowers/reports/2026-04-19-boundary-runtime-entry-decision.md`
|
||||
7. `docs/superpowers/reports/2026-04-19-deferred-family-entry-decision.md`
|
||||
|
||||
## 6. Final Status
|
||||
|
||||
The strongest current execution pressure remains in the completed-roadmap mainline families.
|
||||
|
||||
That means the next roadmap should stay bounded around mainline real-sample mismatch correction instead of drifting into boundary-family or deferred-family expansion.
|
||||
@@ -0,0 +1,64 @@
|
||||
# Real Sample Validation Triage Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Roadmap: `docs/superpowers/plans/2026-04-18-scene-skill-real-sample-validation-roadmap-plan.md`
|
||||
|
||||
## 1. Triage Result
|
||||
|
||||
The three current mainline anchors now classify as:
|
||||
|
||||
1. `G2`: `mismatch-driven`
|
||||
2. `G1-E`: `stable`
|
||||
3. `G3`: `mismatch-driven`
|
||||
|
||||
## 2. Family-Level Reading
|
||||
|
||||
### G2
|
||||
|
||||
`G2` already had a real mismatch anchor from `台区线损大数据-月_周累计线损率统计分析`.
|
||||
|
||||
The key failures remain:
|
||||
|
||||
1. archetype mismatch
|
||||
2. bootstrap mismatch
|
||||
3. request contract missing
|
||||
4. column semantics missing
|
||||
|
||||
This is a mainline compiler-family gap.
|
||||
|
||||
### G1-E
|
||||
|
||||
`G1-E` remains the only stable positive real-sample anchor.
|
||||
|
||||
Its current role is not to open a new scope but to act as the positive baseline against which future real-sample regressions are judged.
|
||||
|
||||
### G3
|
||||
|
||||
`G3` now has an executed real mismatch anchor.
|
||||
|
||||
The key finding is:
|
||||
|
||||
1. the real sample was executed
|
||||
2. the sample did not stay in `paginated_enrichment`
|
||||
3. the run collapsed into `local_doc_pipeline`
|
||||
4. the result then failed closed
|
||||
|
||||
This is also a mainline compiler-family gap.
|
||||
|
||||
## 3. Scope Decision
|
||||
|
||||
The next bounded implementation scope should be:
|
||||
|
||||
1. `mainline G3 real-sample archetype correction`
|
||||
2. followed by `mainline G2 real-sample contract correction`
|
||||
|
||||
The next scope should not be:
|
||||
|
||||
1. a boundary-family roadmap for `G6/G7/G8`
|
||||
2. a deferred-family entry round for `G4/G5`
|
||||
|
||||
## 4. Reason
|
||||
|
||||
The strongest validation pressure still comes from unresolved mainline gaps.
|
||||
|
||||
As long as `G3` and `G2` both fail under real execution, opening boundary or deferred families would be scope drift.
|
||||
@@ -0,0 +1,54 @@
|
||||
# Remaining Route Conflict Correction Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-remaining-route-conflict-correction-plan.md`
|
||||
> Decision Asset: `tests/fixtures/generated_scene/remaining_route_conflict_decisions_2026-04-19.json`
|
||||
|
||||
## Scope
|
||||
|
||||
This report covers only the `4` route conflicts from the follow-up full sweep.
|
||||
|
||||
No timeout handling, structured fail-closed reporting, family baseline, or execution-board status was changed by this plan.
|
||||
|
||||
## Final Decisions
|
||||
|
||||
| Scene | Expected group | Previous inferred archetype | Final decision |
|
||||
| --- | --- | --- | --- |
|
||||
| `95598报修工单日管控` | `G3` | `host_bridge_workflow` | `valid-host-bridge-workflow` |
|
||||
| `95598重要服务事项报备统计表` | `G3` | `host_bridge_workflow` | `valid-host-bridge-workflow` |
|
||||
| `台区线损台区月度高负损预测` | `G2` | `host_bridge_workflow` | `valid-host-bridge-workflow` |
|
||||
| `配网支撑月报(95598抢修统计报表)` | `G3` | `host_bridge_workflow` | `valid-host-bridge-workflow` |
|
||||
|
||||
## Decision Rationale
|
||||
|
||||
The targeted probes showed that the expected `G3` or `G2` contracts do not close for these scenes:
|
||||
|
||||
1. the `G3` cases lack a closed main request, pagination plan, enrichment request, or join key
|
||||
2. the `G2` case lacks a closed mode matrix and mode-specific request contracts
|
||||
3. `host_bridge_workflow` remains the only complete execution path for all four scenes
|
||||
|
||||
Therefore, this plan does not force the scenes back into `G3` or `G2`.
|
||||
|
||||
## Verification
|
||||
|
||||
Targeted probes were run for exactly the fixed `4` records.
|
||||
|
||||
Regression checks:
|
||||
|
||||
1. `cargo test --test scene_generator_test -- --nocapture`
|
||||
2. `cargo test --test scene_generator_canonical_test -- --nocapture`
|
||||
|
||||
Both passed.
|
||||
|
||||
## Implementation Result
|
||||
|
||||
No analyzer route patch was required by the evidence adjudication.
|
||||
|
||||
The route conflict is resolved as a board-expectation/adjudication issue, not a generator bug to be patched in this plan.
|
||||
|
||||
## Stop Statement
|
||||
|
||||
This plan stops here.
|
||||
|
||||
It does not update `scene_execution_board_2026-04-18.json`, does not promote any scene, and does not open another plan.
|
||||
|
||||
@@ -0,0 +1,74 @@
|
||||
# Residual 13 Follow-Up Sweep Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-residual-13-followup-sweep-and-reconciliation-plan.md`
|
||||
> Parent Route: `Residual Route E`
|
||||
> Status: Completed
|
||||
|
||||
## Scope
|
||||
|
||||
This report measures the cumulative delta after residual Routes A through D. It only reruns the fixed `13` residual scenes from the residual closure plan.
|
||||
|
||||
This route does not update `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`, does not add a family, and does not modify analyzer or generator logic.
|
||||
|
||||
## Input
|
||||
|
||||
- `tests/fixtures/generated_scene/full_coverage_reconciliation_candidates_2026-04-19.json`
|
||||
- Fixed input bucket: `13` scenes with `framework-structured-fail-closed`
|
||||
- Output root: `examples/residual_13_followup_sweep_2026-04-19`
|
||||
|
||||
## Result
|
||||
|
||||
| Raw status | Count |
|
||||
| --- | ---: |
|
||||
| `auto-pass` | 6 |
|
||||
| `fail-closed-known` | 7 |
|
||||
| `source-unreadable` | 0 |
|
||||
| `missing-source` | 0 |
|
||||
| `generator-error` | 0 |
|
||||
| `unknown-report-status` | 0 |
|
||||
|
||||
## By Archetype
|
||||
|
||||
| Archetype | Count |
|
||||
| --- | ---: |
|
||||
| `paginated_enrichment` | 4 |
|
||||
| `multi_mode_request` | 2 |
|
||||
| `local_doc_pipeline` | 5 |
|
||||
| `host_bridge_workflow` | 1 |
|
||||
| `page_state_eval` | 1 |
|
||||
|
||||
## Closed By Routes A-D
|
||||
|
||||
These residuals now rerun as `auto-pass`:
|
||||
|
||||
| Scene ID | Scene | Archetype | Readiness |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-007-scene` | `95598供电服务月报` | `paginated_enrichment` | A |
|
||||
| `sweep-018-scene` | `白银线损周报` | `multi_mode_request` | A |
|
||||
| `sweep-039-scene` | `故障报修工单信息统计表` | `paginated_enrichment` | A |
|
||||
| `sweep-068-scene` | `输变电设备运行分析报告` | `paginated_enrichment` | A |
|
||||
| `sweep-071-scene` | `台区线损大数据-月_周累计线损率统计分析` | `multi_mode_request` | A |
|
||||
| `sweep-084-scene` | `巡视计划完成情况自动检索` | `paginated_enrichment` | A |
|
||||
|
||||
## Remaining Structured Fail-Closed
|
||||
|
||||
These residuals remain structured fail-closed and are explained by Route C or Route D decisions:
|
||||
|
||||
| Scene ID | Scene | Archetype | Decision |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-033-scene` | `供电可靠率指标统计表` | `local_doc_pipeline` | hold for local-doc runtime roadmap |
|
||||
| `sweep-034-scene` | `供电可靠性数据质量自查报告月报` | `local_doc_pipeline` | hold for local-doc runtime roadmap |
|
||||
| `sweep-042-scene` | `国网金昌供电公司营商环境周例会报告` | `local_doc_pipeline` | hold for local-doc runtime roadmap |
|
||||
| `sweep-051-scene` | `嘉峪关可靠性分析报告` | `local_doc_pipeline` | hold for local-doc runtime roadmap |
|
||||
| `sweep-074-scene` | `同兴智能安全督查日报` | `local_doc_pipeline` | hold for local-doc runtime roadmap |
|
||||
| `sweep-085-scene` | `业扩报装管理制度` | `host_bridge_workflow` | hold for host-bridge runtime roadmap |
|
||||
| `sweep-091-scene` | `用户停电频次分析监测` | `page_state_eval` | isolate bootstrap target residual |
|
||||
|
||||
## Output Asset
|
||||
|
||||
- `tests/fixtures/generated_scene/residual_13_followup_sweep_2026-04-19.json`
|
||||
|
||||
## Conclusion
|
||||
|
||||
Residual follow-up reduced the fixed residual set from `13` structured fail-closed records to `7` structured fail-closed records. No scene is unreadable, missing, or unresolved under this route.
|
||||
@@ -0,0 +1,62 @@
|
||||
# Residual 13 Reconciliation Candidates Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-residual-13-followup-sweep-and-reconciliation-plan.md`
|
||||
> Policy Source: `tests/fixtures/generated_scene/promotion_board_reconciliation_policy_2026-04-19.json`
|
||||
> Status: Completed
|
||||
|
||||
## Scope
|
||||
|
||||
This report applies Route 6 promotion policy to the residual 13 follow-up sweep results. It publishes reconciliation candidates only.
|
||||
|
||||
This report does not update the official execution board.
|
||||
|
||||
## Candidate Summary
|
||||
|
||||
| Candidate status | Count |
|
||||
| --- | ---: |
|
||||
| `framework-auto-pass-candidate` | 6 |
|
||||
| `framework-structured-fail-closed` | 7 |
|
||||
| `source-unreadable` | 0 |
|
||||
| `unresolved-followup-status` | 0 |
|
||||
|
||||
## Auto-Pass Candidates
|
||||
|
||||
The following scenes are candidates for future explicit board reconciliation as framework auto-pass candidates:
|
||||
|
||||
| Scene ID | Scene | Archetype |
|
||||
| --- | --- | --- |
|
||||
| `sweep-007-scene` | `95598供电服务月报` | `paginated_enrichment` |
|
||||
| `sweep-018-scene` | `白银线损周报` | `multi_mode_request` |
|
||||
| `sweep-039-scene` | `故障报修工单信息统计表` | `paginated_enrichment` |
|
||||
| `sweep-068-scene` | `输变电设备运行分析报告` | `paginated_enrichment` |
|
||||
| `sweep-071-scene` | `台区线损大数据-月_周累计线损率统计分析` | `multi_mode_request` |
|
||||
| `sweep-084-scene` | `巡视计划完成情况自动检索` | `paginated_enrichment` |
|
||||
|
||||
## Structured Fail-Closed Candidates
|
||||
|
||||
The following scenes remain framework-supported but are not execution-pass candidates:
|
||||
|
||||
| Scene ID | Scene | Archetype | Overlay |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-033-scene` | `供电可靠率指标统计表` | `local_doc_pipeline` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-034-scene` | `供电可靠性数据质量自查报告月报` | `local_doc_pipeline` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-042-scene` | `国网金昌供电公司营商环境周例会报告` | `local_doc_pipeline` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-051-scene` | `嘉峪关可靠性分析报告` | `local_doc_pipeline` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-074-scene` | `同兴智能安全督查日报` | `local_doc_pipeline` | `hold-for-local-doc-runtime-roadmap` |
|
||||
| `sweep-085-scene` | `业扩报装管理制度` | `host_bridge_workflow` | `hold-for-host-bridge-runtime-roadmap` |
|
||||
| `sweep-091-scene` | `用户停电频次分析监测` | `page_state_eval` | `isolate-bootstrap-target-residual` |
|
||||
|
||||
## Board Policy
|
||||
|
||||
Per Route 6 policy, these candidates must not directly update `scene_execution_board_2026-04-18.json`.
|
||||
|
||||
A future dedicated board reconciliation plan may consume this asset and decide official status updates.
|
||||
|
||||
## Output Asset
|
||||
|
||||
- `tests/fixtures/generated_scene/residual_13_reconciliation_candidates_2026-04-19.json`
|
||||
|
||||
## Conclusion
|
||||
|
||||
Residual Route E has no unresolved follow-up state. The remaining residual count is `7`, all explained by boundary/runtime or bootstrap isolation decisions.
|
||||
@@ -0,0 +1,47 @@
|
||||
# Residual Runtime Roadmap Prioritization Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-residual-runtime-roadmap-prioritization-plan.md`
|
||||
> Status: Completed
|
||||
|
||||
## Scope
|
||||
|
||||
This is a decision-only step after official board reconciliation. It compares the three residual roadmap inputs and selects exactly one next roadmap.
|
||||
|
||||
This plan did not modify analyzer/generator logic and did not update the official execution board.
|
||||
|
||||
## Inputs
|
||||
|
||||
- `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
- `tests/fixtures/generated_scene/official_board_reconciliation_2026-04-19.json`
|
||||
|
||||
## Candidate Roadmaps
|
||||
|
||||
| Candidate | Residual count | Score | Decision |
|
||||
| --- | ---: | ---: | --- |
|
||||
| `local-doc runtime roadmap` | 5 | 51 | selected |
|
||||
| `bootstrap target normalization roadmap` | 1 | 11 | deferred |
|
||||
| `host-bridge runtime roadmap` | 1 | 8 | deferred |
|
||||
|
||||
## Selected Roadmap
|
||||
|
||||
`local-doc runtime roadmap` is selected because it covers the largest residual bucket: five `local_doc_pipeline` scenes.
|
||||
|
||||
Selected next roadmap files:
|
||||
|
||||
- `docs/superpowers/specs/2026-04-19-local-doc-runtime-roadmap-design.md`
|
||||
- `docs/superpowers/plans/2026-04-19-local-doc-runtime-roadmap-plan.md`
|
||||
|
||||
## Deferred Roadmaps
|
||||
|
||||
`host-bridge runtime roadmap` is deferred because it currently covers one residual and carries higher risk of affecting existing host-bridge auto-pass paths.
|
||||
|
||||
`bootstrap target normalization roadmap` is deferred because it currently covers one residual and should remain isolated until the higher-impact local-doc bucket is handled.
|
||||
|
||||
## Output Asset
|
||||
|
||||
- `tests/fixtures/generated_scene/residual_runtime_roadmap_prioritization_2026-04-19.json`
|
||||
|
||||
## Conclusion
|
||||
|
||||
The next bounded roadmap is local-doc runtime. Do not start host-bridge runtime or bootstrap normalization until this selected roadmap is closed or explicitly superseded.
|
||||
@@ -0,0 +1,63 @@
|
||||
# Scene Skill 102 Final Materialization Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `2026-04-19-scene-skill-102-final-materialization-plan.md`
|
||||
|
||||
## Scope
|
||||
|
||||
This plan materialized the final skill package set for the current 102 scene framework closure.
|
||||
|
||||
Existing `examples/*` follow-up roots were not cleaned or modified. The final materialization was written to:
|
||||
|
||||
`D:/data/ideaSpace/rust/sgClaw/claw-new/examples/scene_skill_102_final_materialization_2026-04-19`
|
||||
|
||||
## Summary
|
||||
|
||||
| Metric | Count |
|
||||
| --- | ---: |
|
||||
| total scenes | 102 |
|
||||
| attempted | 102 |
|
||||
| generated skill directories | 102 |
|
||||
| required file checks passed | 101 |
|
||||
| failures | 1 |
|
||||
|
||||
Duration: `2036.246` seconds
|
||||
|
||||
## Readiness Distribution
|
||||
|
||||
| Readiness | Count |
|
||||
| --- | ---: |
|
||||
| `A` | 100 |
|
||||
| `C` | 2 |
|
||||
|
||||
## Archetype Distribution
|
||||
|
||||
| Archetype | Count |
|
||||
| --- | ---: |
|
||||
| `host_bridge_workflow` | 27 |
|
||||
| `local_doc_pipeline` | 5 |
|
||||
| `multi_endpoint_inventory` | 2 |
|
||||
| `multi_mode_request` | 10 |
|
||||
| `page_state_eval` | 2 |
|
||||
| `paginated_enrichment` | 51 |
|
||||
| `single_request_enrichment` | 5 |
|
||||
|
||||
## Assets
|
||||
|
||||
- `tests/fixtures/generated_scene/scene_skill_102_final_materialization_manifest_2026-04-19.json`
|
||||
- `tests/fixtures/generated_scene/scene_skill_102_final_materialization_failures_2026-04-19.json`
|
||||
- `D:/data/ideaSpace/rust/sgClaw/claw-new/examples/scene_skill_102_final_materialization_2026-04-19`
|
||||
|
||||
## Failures
|
||||
|
||||
| Scene id | Scene name | Error | Missing files |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-012-scene` | `业扩报装管理制度` | `generator-exit-nonzero` | `SKILL.toml, SKILL.md, scene.toml, scripts/*` |
|
||||
|
||||
## Closure
|
||||
|
||||
This materialization plan published the required final materialization assets: a 102-row manifest, an explicit failures asset, the isolated output root, and this report.
|
||||
|
||||
The final materialized skill asset set is not yet 102/102 complete because `sweep-012-scene` failed package materialization and is missing `SKILL.toml`, `SKILL.md`, `scene.toml`, and `scripts/*`.
|
||||
|
||||
Next work should not start static/mock validation yet. It should first use a separate bounded recovery plan for `sweep-012-scene`, then refresh this final materialization manifest. This plan does not perform recovery, static, mock, or production validation.
|
||||
@@ -0,0 +1,40 @@
|
||||
# Scene Skill 102 Framework Closure Rollup Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `2026-04-19-102-framework-closure-rollup-plan.md`
|
||||
> Parent framework: `2026-04-19-scene-skill-102-full-coverage-framework-plan.md`
|
||||
|
||||
## Result
|
||||
|
||||
| Metric | Count |
|
||||
| --- | ---: |
|
||||
| total scenes | 102 |
|
||||
| `framework-auto-pass` | 102 |
|
||||
| `framework-structured-fail-closed` | 0 |
|
||||
| `source-unreadable` | 0 |
|
||||
| `missing-source` | 0 |
|
||||
| `unsupported-family` | 0 |
|
||||
| `misclassified-unresolved` | 0 |
|
||||
| `unresolved-followup-status` | 0 |
|
||||
| unresolved total | 0 |
|
||||
|
||||
Closure status: `complete`
|
||||
|
||||
## Remaining Structured Fail-Closed
|
||||
|
||||
None.
|
||||
|
||||
## Interpretation
|
||||
|
||||
The 102-scene set is closed at the framework layer: every scene now has a framework auto-pass status in the official board.
|
||||
|
||||
This is not the same as saying every scene has real-sample `executed-pass` validation. Real-sample validation remains a separate layer. The framework closure means the generator/analyzer framework can classify and generate a bounded skill package path for all 102 scenes without unresolved timeout, unsupported-family, unresolved route conflict, or structured fail-closed residual.
|
||||
|
||||
## Inputs
|
||||
|
||||
- `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
|
||||
- `tests/fixtures/generated_scene/final_2_official_board_reconciliation_refresh_2026-04-19.json`
|
||||
|
||||
## Closure
|
||||
|
||||
The `102` full coverage parent framework has reached framework-level closure. Further work should be planned as a new validation or production-hardening roadmap, not as continuation of unresolved framework coverage.
|
||||
@@ -0,0 +1,14 @@
|
||||
# Structured Fail-Closed Bootstrap Isolation Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
|
||||
## Bootstrap Records
|
||||
|
||||
| Scene ID | Archetype | Reason | Decision |
|
||||
| --- | --- | --- | --- |
|
||||
| `sweep-012-scene` | `host_bridge_workflow` | `workflow evidence is incomplete before package generation` | `defer_to_bootstrap_normalization` |
|
||||
| `sweep-066-scene` | `page_state_eval` | `bootstrap_target` | `defer_to_bootstrap_normalization` |
|
||||
|
||||
## Scope Confirmation
|
||||
|
||||
No login recovery or bootstrap auto-recovery is implemented in this roadmap.
|
||||
@@ -0,0 +1,109 @@
|
||||
# Structured Fail-Closed Improvement Coverage Delta Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-structured-fail-closed-improvement-roadmap-plan.md`
|
||||
> Baseline: `tests/fixtures/generated_scene/full_sweep_status_reconciliation_2026-04-19.json`
|
||||
> Follow-up: `tests/fixtures/generated_scene/structured_fail_closed_improvement_followup_2026-04-19.json`
|
||||
|
||||
## Purpose
|
||||
|
||||
Measure the impact of the bounded structured fail-closed improvement cycle against the reconciled `102` scene baseline.
|
||||
|
||||
This report does not promote scenes, update `scene_execution_board_2026-04-18.json`, introduce new scene families, or relax readiness gates.
|
||||
|
||||
## Baseline
|
||||
|
||||
The reconciled baseline before this roadmap was:
|
||||
|
||||
| Status | Count |
|
||||
| --- | ---: |
|
||||
| `auto-pass` | 48 |
|
||||
| `fail-closed-known` | 48 |
|
||||
| `adjudicated-valid-host-bridge` | 4 |
|
||||
| `source-unreadable` | 2 |
|
||||
| Total | 102 |
|
||||
|
||||
The baseline actionable count was `100`, using:
|
||||
|
||||
`auto-pass + fail-closed-known + adjudicated-valid-host-bridge`
|
||||
|
||||
## Follow-Up Sweep
|
||||
|
||||
The follow-up sweep was run over the same fixed `102` scene set and wrote results to:
|
||||
|
||||
`tests/fixtures/generated_scene/structured_fail_closed_improvement_followup_2026-04-19.json`
|
||||
|
||||
Follow-up status:
|
||||
|
||||
| Status | Count |
|
||||
| --- | ---: |
|
||||
| `auto-pass` | 48 |
|
||||
| `fail-closed-known` | 47 |
|
||||
| `adjudicated-valid-host-bridge` | 4 |
|
||||
| `source-unreadable` | 3 |
|
||||
| Total | 102 |
|
||||
|
||||
Follow-up actionable count is `99`, using:
|
||||
|
||||
`auto-pass + fail-closed-known + adjudicated-valid-host-bridge`
|
||||
|
||||
## Delta
|
||||
|
||||
| Metric | Baseline | Follow-up | Delta |
|
||||
| --- | ---: | ---: | ---: |
|
||||
| `auto-pass` | 48 | 48 | 0 |
|
||||
| `fail-closed-known` | 48 | 47 | -1 |
|
||||
| `adjudicated-valid-host-bridge` | 4 | 4 | 0 |
|
||||
| `source-unreadable` | 2 | 3 | +1 |
|
||||
| actionable count | 100 | 99 | -1 |
|
||||
|
||||
## Interpretation
|
||||
|
||||
This roadmap did not increase `auto-pass` coverage.
|
||||
|
||||
The bounded implementation improved the structure of pre-package fail-closed reports by preserving the full `SceneIr` as `contractSnapshot`. This makes fail-closed records more diagnosable, but it does not close missing contracts by itself.
|
||||
|
||||
The follow-up sweep exposed one new timeout relative to the reconciled baseline:
|
||||
|
||||
| Scene id | Scene | Baseline status | Follow-up status | Reason |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| `sweep-040-scene` | `嘉峪关日报` | `fail-closed-known` | `source-unreadable` | `generator timeout after 45s` |
|
||||
|
||||
The two existing timeout records remain timeout records:
|
||||
|
||||
| Scene id | Scene | Status |
|
||||
| --- | --- | --- |
|
||||
| `sweep-015-scene` | `任务报表` | `source-unreadable` |
|
||||
| `sweep-025-scene` | `力禾动环系统巡视记录` | `source-unreadable` |
|
||||
|
||||
## Remaining Structured Fail-Closed Buckets
|
||||
|
||||
The follow-up sweep still has `47` structured fail-closed records:
|
||||
|
||||
| Archetype | Reason | Count |
|
||||
| --- | --- | ---: |
|
||||
| `paginated_enrichment` | `workflow evidence is incomplete before package generation` | 34 |
|
||||
| `local_doc_pipeline` | `workflow evidence is incomplete before package generation` | 5 |
|
||||
| `multi_mode_request` | `workflow evidence is incomplete before package generation` | 4 |
|
||||
| `single_request_enrichment` | `workflow evidence is incomplete before package generation` | 2 |
|
||||
| `host_bridge_workflow` | `workflow evidence is incomplete before package generation` | 1 |
|
||||
| `page_state_eval` | `structured fail-closed report emitted` | 1 |
|
||||
|
||||
The largest remaining bucket is still `paginated_enrichment`, now `34` records.
|
||||
|
||||
## Boundaries Preserved
|
||||
|
||||
This follow-up did not:
|
||||
|
||||
1. update `scene_execution_board_2026-04-18.json`
|
||||
2. promote scenes
|
||||
3. introduce a new scene family
|
||||
4. reopen adjudicated host-bridge records
|
||||
5. handle timeout records as implementation targets
|
||||
6. loosen readiness gates
|
||||
|
||||
## Conclusion
|
||||
|
||||
The roadmap produced a better fail-closed evidence artifact, but it did not increase `auto-pass` coverage. The measured status is slightly worse on broad actionable count because one previous fail-closed scene timed out in the follow-up sweep.
|
||||
|
||||
The next input, outside this roadmap, is a bounded timeout regression diagnostic for `sweep-040-scene` and the two persistent timeout records, or a new recovery roadmap focused on the remaining `34` `paginated_enrichment` fail-closed records.
|
||||
@@ -0,0 +1,170 @@
|
||||
# Structured Fail-Closed Improvement Roadmap Closure Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-structured-fail-closed-improvement-roadmap-plan.md`
|
||||
> Status: Closed
|
||||
|
||||
## Scope
|
||||
|
||||
This roadmap handled the `48` structured fail-closed records from the reconciled `102` scene sweep.
|
||||
|
||||
The roadmap boundary explicitly excluded:
|
||||
|
||||
1. new scene families
|
||||
2. `G4/G5`
|
||||
3. login recovery
|
||||
4. full host runtime transport
|
||||
5. local document attachment runtime
|
||||
6. `scene_execution_board_2026-04-18.json` updates
|
||||
7. direct scene promotion
|
||||
8. timeout implementation work
|
||||
9. readiness gate relaxation
|
||||
|
||||
Those boundaries were preserved.
|
||||
|
||||
## Completed Phases
|
||||
|
||||
### Phase 0: Freeze Structured Fail-Closed Baseline
|
||||
|
||||
Completed.
|
||||
|
||||
Input baseline:
|
||||
|
||||
| Status | Count |
|
||||
| --- | ---: |
|
||||
| `auto-pass` | 48 |
|
||||
| `fail-closed-known` | 48 |
|
||||
| `adjudicated-valid-host-bridge` | 4 |
|
||||
| `source-unreadable` | 2 |
|
||||
| Total | 102 |
|
||||
|
||||
Only `fail-closed-known` records entered the inventory phase.
|
||||
|
||||
### Phase 1: Inventory and Gap Taxonomy
|
||||
|
||||
Completed.
|
||||
|
||||
Deliverables:
|
||||
|
||||
1. `tests/fixtures/generated_scene/structured_fail_closed_inventory_2026-04-19.json`
|
||||
2. `docs/superpowers/reports/2026-04-19-structured-fail-closed-inventory-report.md`
|
||||
|
||||
Inventory summary:
|
||||
|
||||
| Primary label | Count |
|
||||
| --- | ---: |
|
||||
| `enrichment_request_missing` | 23 |
|
||||
| `export_plan_missing` | 12 |
|
||||
| `local_doc_contract_missing` | 5 |
|
||||
| `mode_request_contract_missing` | 4 |
|
||||
| `single_request_enrichment_contract_missing` | 2 |
|
||||
| `bootstrap_target_unresolved` | 2 |
|
||||
|
||||
The `paginated_enrichment` bucket was split into:
|
||||
|
||||
| Primary label | Count |
|
||||
| --- | ---: |
|
||||
| `enrichment_request_missing` | 23 |
|
||||
| `export_plan_missing` | 12 |
|
||||
|
||||
### Phase 2: G3 Paginated Enrichment Recovery Slice
|
||||
|
||||
Completed as a bounded evidence-preservation slice.
|
||||
|
||||
The top repeated G3 gap was `enrichment_request_missing`, but there was not enough traceable evidence to safely synthesize missing enrichment requests without relaxing the `G3` contract.
|
||||
|
||||
Instead, the implementation improved structured fail-closed observability by embedding `contractSnapshot` in pre-package fail-closed reports. This preserves the full `SceneIr` for future recovery analysis while keeping the result fail-closed.
|
||||
|
||||
Changed implementation:
|
||||
|
||||
1. `src/generated_scene/generator.rs`
|
||||
2. `tests/scene_generator_test.rs`
|
||||
|
||||
This slice did not:
|
||||
|
||||
1. hardcode scene names
|
||||
2. relax gates
|
||||
3. convert incomplete `G3` records into pass
|
||||
4. change canonical `G3` behavior
|
||||
|
||||
### Phase 3: Small-Bucket Recovery Slice
|
||||
|
||||
Completed as inspection and explicit deferral.
|
||||
|
||||
Deliverable:
|
||||
|
||||
`docs/superpowers/reports/2026-04-19-structured-fail-closed-small-bucket-decision-report.md`
|
||||
|
||||
No small-bucket implementation slice was started because the largest remaining value stayed in the `G3` structured fail-closed bucket, and the plan allowed at most one non-G3 bounded slice.
|
||||
|
||||
### Phase 4: Bootstrap Target Isolation
|
||||
|
||||
Completed.
|
||||
|
||||
Deliverable:
|
||||
|
||||
`docs/superpowers/reports/2026-04-19-structured-fail-closed-bootstrap-isolation-report.md`
|
||||
|
||||
Bootstrap target records were kept out of G3 and small-bucket recovery. No login recovery or bootstrap auto-recovery was implemented.
|
||||
|
||||
### Phase 5: Follow-Up Sweep and Coverage Delta
|
||||
|
||||
Completed.
|
||||
|
||||
Deliverables:
|
||||
|
||||
1. `tests/fixtures/generated_scene/structured_fail_closed_improvement_followup_2026-04-19.json`
|
||||
2. `docs/superpowers/reports/2026-04-19-structured-fail-closed-improvement-coverage-delta-report.md`
|
||||
|
||||
Follow-up summary:
|
||||
|
||||
| Status | Baseline | Follow-up | Delta |
|
||||
| --- | ---: | ---: | ---: |
|
||||
| `auto-pass` | 48 | 48 | 0 |
|
||||
| `fail-closed-known` | 48 | 47 | -1 |
|
||||
| `adjudicated-valid-host-bridge` | 4 | 4 | 0 |
|
||||
| `source-unreadable` | 2 | 3 | +1 |
|
||||
|
||||
The follow-up did not improve pass count and exposed one additional timeout:
|
||||
|
||||
`sweep-040-scene` / `嘉峪关日报`
|
||||
|
||||
## Validation
|
||||
|
||||
The implementation slice was validated with:
|
||||
|
||||
1. `cargo test --test scene_generator_test generator_blocks_incomplete_paginated_enrichment_workflow -- --nocapture`
|
||||
2. `cargo test --test scene_generator_canonical_test -- --nocapture`
|
||||
|
||||
Both passed.
|
||||
|
||||
The fixed `102` scene follow-up sweep completed and wrote:
|
||||
|
||||
`tests/fixtures/generated_scene/structured_fail_closed_improvement_followup_2026-04-19.json`
|
||||
|
||||
## Closure Decision
|
||||
|
||||
This roadmap is closed.
|
||||
|
||||
It completed its required inventory, bounded implementation, small-bucket inspection, bootstrap isolation, follow-up sweep, and delta report.
|
||||
|
||||
The outcome is not a coverage increase. The outcome is better structured fail-closed evidence plus a measured follow-up baseline:
|
||||
|
||||
| Metric | Result |
|
||||
| --- | --- |
|
||||
| `auto-pass` coverage | unchanged at `48/102` |
|
||||
| broad actionable count | `100 -> 99` |
|
||||
| remaining structured fail-closed | `47` |
|
||||
| remaining timeout | `3` |
|
||||
|
||||
## Next Inputs
|
||||
|
||||
The next work is outside this roadmap.
|
||||
|
||||
Highest-value inputs are:
|
||||
|
||||
1. bounded timeout regression diagnostic for `sweep-040-scene`
|
||||
2. timeout diagnostic for the two persistent timeout records
|
||||
3. a new bounded recovery roadmap for the remaining `34` `paginated_enrichment` structured fail-closed records
|
||||
|
||||
No next roadmap is started by this closure report.
|
||||
@@ -0,0 +1,42 @@
|
||||
# Structured Fail-Closed Inventory Report
|
||||
|
||||
> Date: 2026-04-19
|
||||
> Plan: `docs/superpowers/plans/2026-04-19-structured-fail-closed-improvement-roadmap-plan.md`
|
||||
> Inventory: `tests/fixtures/generated_scene/structured_fail_closed_inventory_2026-04-19.json`
|
||||
|
||||
## Summary
|
||||
|
||||
Total structured fail-closed records: `48`.
|
||||
|
||||
| Inferred archetype | Count |
|
||||
| --- | ---: |
|
||||
| `host_bridge_workflow` | 1 |
|
||||
| `local_doc_pipeline` | 5 |
|
||||
| `multi_mode_request` | 4 |
|
||||
| `page_state_eval` | 1 |
|
||||
| `paginated_enrichment` | 35 |
|
||||
| `single_request_enrichment` | 2 |
|
||||
|
||||
## Primary Missing Contract Labels
|
||||
|
||||
| Label | Count |
|
||||
| --- | ---: |
|
||||
| `bootstrap_target_unresolved` | 2 |
|
||||
| `enrichment_request_missing` | 23 |
|
||||
| `export_plan_missing` | 12 |
|
||||
| `local_doc_contract_missing` | 5 |
|
||||
| `mode_request_contract_missing` | 4 |
|
||||
| `single_request_enrichment_contract_missing` | 2 |
|
||||
|
||||
## G3 Bucket
|
||||
|
||||
| G3 primary label | Count |
|
||||
| --- | ---: |
|
||||
| `enrichment_request_missing` | 23 |
|
||||
| `export_plan_missing` | 12 |
|
||||
|
||||
The largest G3 repeated gap is the first candidate for bounded recovery analysis. Recovery remains conditional: fields must be traceable to source evidence and existing G3 canonical/real-sample regressions must stay green.
|
||||
|
||||
## Scope Confirmation
|
||||
|
||||
No implementation was performed in this inventory phase. No execution board status was changed.
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user