|
30906
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"6fb129d40 {"Commits":[{"Sha1":"6fb129d40fb9e160af6f70871eb8881ff5eef743","Message":"chore(reports): 门禁消费回绑 @ 69784be\n\nR9 落地后第一条真实命中：69784be 改了 scripts/ 下两个文件，而 scripts/ 正是这份报告\n声明的输入之一（它拿 门禁消费登记.json 对盘受控脚本，读的是脚本内容不只是文件名）。\n干净树重跑，内容逐字未变，只有绑定前移——这正是「绑定干净不等于结论还成立」的反面：\n输入动了就该重跑，哪怕结论恰好没变。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:55:46-07:00"},{"Sha1":"69784be6b04c850ae1f2ea0b948344f9627738cf","Message":"feat(governance): 报告登记门禁补 R9 / R10——绑定干净不等于结论还成立\n\nR7 / R8 守的是绑定本身（不脏、在历史上）。这一轮补另一半：**一份绑得干干净净的报告，\n结论仍可能早已被推翻**。2026-09-18 实测 `门禁消费.latest.json` 相对它自己的输入\n`门禁消费登记.json` 已经陈旧——登记表改了并提交了，报告停在旧文本，R1—R8 一条都不响。\n\n R9 LATEST_STALE 登记声明了 inputs，绑定 SHA → HEAD 之间这些输入有变更\n R10 INPUTS_UNACCOUNTED latest 既没声明 inputs、也没写明为什么判不了\n\n判据与真源仓 check:evidence 同口径：不看「落后几个提交」，看**该报告自己覆盖的输入**\n动没动。同目录的 latest 报告不算输入，它们是产物。\n\n四份 latest 逐条给出口径，禁止留空白：\n 门禁消费 inputs：门禁消费登记.json + scripts/（它正是拿登记表对盘受控脚本）\n 基础目录统一审计 判不了——读其它仓的 HEAD / 工作树 / 验收报告，全在治理仓历史之外，\n 且 ageHours 每跑一次都变，本仓任何提交都决定不了它是否仍成立\n 延后op 判不了一半——决议材料在本仓，活跃 Catalog 在嵌套独立仓（治理仓忽略该\n 目录，git diff 看不见）。只声明能判的一半会给出假绿，比不判更坏\n 门禁运行 判不了——编排产物，写 inputs 只能写成 scripts/ 这种筐，真源仓\n evidence-scopes 的同一条纪律禁止为覆盖率造模糊作用域\n\n集成用例改为只在真实登记上断言**结构类**规则。仓纪律 5 要求两次提交（实现，然后在它\n的干净树上回绑），必然存在一个提交报告是陈旧的——那是纪律规定的流程，不是缺陷；在\n单测里断言它为零等于让 pnpm test 在自家流程的中间一步必红。「待回绑」四条由阻断性\n门禁 check:reports 把关。该用例原本就已按同一理由排除 LATEST_WITHOUT_PROVENANCE，\n本轮只是把这一类补全。\n\n登记门禁测试 13 → 19，治理层全量 389（383 过 / 6 跳 / 0 失败）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:55:24-07:00"}],"HeadCommit":{"Sha1":"6fb129d40fb9e160af6f70871eb8881ff5eef743","Message":"chore(reports): 门禁消费回绑 @ 69784be\n\nR9 落地后第一条真实命中：69784be 改了 scripts/ 下两个文件，而 scripts/ 正是这份报告\n声明的输入之一（它拿 门禁消费登记.json 对盘受控脚本，读的是脚本内容不只是文件名）。\n干净树重跑，内容逐字未变，只有绑定前移——这正是「绑定干净不等于结论还成立」的反面：\n输入动了就该重跑，哪怕结论恰好没变。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:55:46-07:00"},"CompareURL":"luoanwu/platform-governance/compare/f4540e69b1656232b677571584d3bd6ad57217be...6fb129d40fb9e160af6f70871eb8881ff5eef743","Len":2}...
|
1789718149
|
Edit
Delete
|
|
30905
|
5
|
5
|
5
|
76
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"ea965ed90 {"Commits":[{"Sha1":"ea965ed902a9807d07d7d89fc173668804c6294c","Message":"chore(reports): 内核准入门禁首轮 8 处诚实红 回绑 @ ee6d5a9\n\nclean 绑定（worktreeDirty:false）——工作区里三份 reports/ui-*.latest.json 的改动\n早于本批次，且 report-provenance 按设计忽略 latest 报告，不影响本次绑定。\n\nK4 七处（声明为 true，evidenceSources 无机器回执）：\n- kernel.digital-employee-core.industryIndependent ← 只有 scripts/check-kernel-boundaries.mjs\n- kernel.digital-employee-core.crossProductRequired ← 只有下游夹具 .ts\n- kernel.digital-employee-core.dualBackendGoverned ← 一份 ADR + 两个脚本\n- pack.generative-ai 的 industryIndependent / dualBackendGoverned ← 源码文件与目录\n- pack.enterprise-fact-consumer 的两项同上\n\nK3 一处：tests/kernel-downstream-fixture/src/index.ts 只有命名空间导入。\n\n未一并修 evidenceSources。改指某份已存在的报告能立刻转绿，但只有当那份报告确实判过\n该条时才是真修：dual-backend-parity 是全仓级判定、不按组件出结论，改指它等于把自证\n换个地方做。哪份报告承接哪一条属 Owner 裁定，不随本次实现一并决定。\n\nK5 观测（不计红）：跨产品具名消费面 5 个符号——ProductManifestInput、\nProductManifestComposition、ProductRuntimeContractCatalog、\nProductDataContractImplementation、ProductEventContractImplementation，\n全部来自 product-manifest.ts；33/38 个模块零跨产品消费，含 tenant.ts 的 65 个导出与\ntool.ts 的 99 个。该数字只作边界收缩决策的输入，不构成「应移出 CORE」的判定。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:52:16-07:00"},{"Sha1":"ee6d5a98aa6d372d2fc2ea81f328d55cc15511f9","Message":"feat(governance): 内核准入证据门禁 K1—K5，判据从声明下沉到导出清单与证据形态\n\nkernel.boundaries.json 的 admissionEvidence 是四个手写 true，evidenceSources\n只列了路径，没有「该判据确实被机器判过」的回执。实测两处后果：\n\n- CORE 声明 crossProductRequired: true，其证据源 tests/kernel-downstream-fixture/\n src/index.ts 是 10 个 import * as + 判非空 + 一个 typeof === \"function\" 的\n 命名空间冒烟测试——28 个契约模块全空它也照样通过。它能证明包可安装可加载，\n 不能证明任何一个模块跨产品必需\n- 归属粒度是文件：owns.contractModules 是文件名清单，于是\n packages/contracts/src/tenant.ts 一个 1075 行 / 65 导出的文件，把租户策略、\n JWT 验签、自建权限、work 域业务判定四类一起带进了 CORE\n\n手法同 check:platform-provenance（那一条把判据从 package.json 的声明下沉到\npnpm-lock.yaml 的解析结果）。五条判据：\n\n K1 EXPORT_LOCK_DRIFT 组件所辖契约模块的直接导出集合须与\n kernel.exports.lock.json 一致。只认直接声明，\n barrel 的再导出不算拥有——否则 index.ts 会「拥有」全仓\n K2 EXPORT_OWNERSHIP_AMBIGUOUS 同一导出名不得被两个组件同时拥有。拆分契约模块时\n 最容易「复制一份留在原处」，这条守住它\n K3 EVIDENCE_NAMESPACE_ONLY crossProductRequired 的证据源若只有命名空间导入、\n 无任何具名导入，按构造无法证明符号级必需性\n K4 EVIDENCE_NOT_BACKED 声明为 true 的每项须至少有一份机器回执\n （reports/*.json 或 *.lock.json）。脚本路径与文档路径\n 不是回执：只证明文件存在，不证明该判据被判过。\n 声明为 false 的项不要求回执\n K5 CROSS_PRODUCT_UNCONSUMED observed，不计红。CORE 模块未被产品消费不等于不该在\n CORE（OS 自己的两个后端在用）；把观测项当红盘会逼出\n 「为了绿而乱改归属」的反向激励\n\n自检 pnpm check:kernel-admission:self-test 共 10 条，含 1 条基线正例与 2 条防过度\n判红（声明 false 不因无回执计红、零跨产品消费只出 observed）——对照组是为了防\n「只会红的假门禁」。\n\n锁基线 38 个模块 / 779 个导出（CORE 30 + generative-ai 7 + fact-consumer 1）。\n首次签发只固化「现在是什么」，不代表归属正确。\n\ncheck-all 末位注册，理由同 check:platform-provenance：本门禁自引入起即为已知诚实红，\nfail-fast 语义下排前面会让其余门禁全部不执行。pnpm check 因此从 27 步变 28 步且末步红，\n前 27 步照常产出证据。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:52:01-07:00"}],"HeadCommit":{"Sha1":"ea965ed902a9807d07d7d89fc173668804c6294c","Message":"chore(reports): 内核准入门禁首轮 8 处诚实红 回绑 @ ee6d5a9\n\nclean 绑定（worktreeDirty:false）——工作区里三份 reports/ui-*.latest.json 的改动\n早于本批次，且 report-provenance 按设计忽略 latest 报告，不影响本次绑定。\n\nK4 七处（声明为 true，evidenceSources 无机器回执）：\n- kernel.digital-employee-core.industryIndependent ← 只有 scripts/check-kernel-boundaries.mjs\n- kernel.digital-employee-core.crossProductRequired ← 只有下游夹具 .ts\n- kernel.digital-employee-core.dualBackendGoverned ← 一份 ADR + 两个脚本\n- pack.generative-ai 的 industryIndependent / dualBackendGoverned ← 源码文件与目录\n- pack.enterprise-fact-consumer 的两项同上\n\nK3 一处：tests/kernel-downstream-fixture/src/index.ts 只有命名空间导入。\n\n未一并修 evidenceSources。改指某份已存在的报告能立刻转绿，但只有当那份报告确实判过\n该条时才是真修：dual-backend-parity 是全仓级判定、不按组件出结论，改指它等于把自证\n换个地方做。哪份报告承接哪一条属 Owner 裁定，不随本次实现一并决定。\n\nK5 观测（不计红）：跨产品具名消费面 5 个符号——ProductManifestInput、\nProductManifestComposition、ProductRuntimeContractCatalog、\nProductDataContractImplementation、ProductEventContractImplementation，\n全部来自 product-manifest.ts；33/38 个模块零跨产品消费，含 tenant.ts 的 65 个导出与\ntool.ts 的 99 个。该数字只作边界收缩决策的输入，不构成「应移出 CORE」的判定。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:52:16-07:00"},"CompareURL":"luoanwu/digital-employee-os/compare/c433d57b4aae2249bd19a085d1c520bf6baedf55...ea965ed902a9807d07d7d89fc173668804c6294c","Len":2}...
|
1789718112
|
Edit
Delete
|
|
30904
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"f4540e69b {"Commits":[{"Sha1":"f4540e69b1656232b677571584d3bd6ad57217be","Message":"docs(架构): 总架构增 §3.8——基座的装载 / 调用两个方向与接入能力归属\n\n起因是一次实际误读：`c433d57` 被当成「另一个 OS 产品」。它是数字员工OS 仓的提交\nSHA（2026-09-14），由 base-framework-os-product/os.upstream.json 连同 protocol\n1.3.0 钉住，用途是让「上游 OS 走动了」以 O0-SHA 显式变红。与基座并列的不是它，\n是 base-framework-os-product 这个仓——示例产品扩展加两道 OS 门禁，2026-09-10\n随 G26 自框架仓移出以保框架零依赖。误读本身说明这几件事在文档里没写清，本节补上。\n\n**两个方向。** PNG 给基座的定语「理解、编排、调用、执行」对应两个方向相反的接口：\n装载（OS → 应用）已有协议（KERNEL_EXTENSION_PROTOCOL_VERSION = 1.3.0，定义在 OS\n仓 packages/contracts/src/product-manifest.ts），真装 1 个；调用（应用 → OS）基本\n空白——协议草案 §4 已给出 ProductActivationIntent / ToolIntentReceipt 并写明「OS\n应提供、产品不再自建」，2026-09-18 复核 OS 仓 apps/ 下仍无 product-activations\n路由，智服已在自己仓内造了一套。两个方向归属不同，混谈会把落点找错。\n\n**接入形态实测已分叉成五种：** product.manifest.json（示例 / 设备云 / 知识云）、\nmanifest.json（HR，另有 ddl-authority / distribution / readiness 三个独有文件且无\npackage.json）、manifest.ts（工单，另有 kernel.products.entry.json）。只有示例与\n设备云带 os-host 测试配置。\n\n**为什么「每个应用继承一份」今天做不到**（两条方向相反的硬事实）：OS 的包是\n@repo/* 仓内作用域名（@repo/contracts@1.16.0，与工单仓同名包撞名），外仓依赖不了，\n只能照抄；而协议 §2.1 又反向禁止产品导入宿主私有 @repo/*，宿主能力一律由组合根\n注入、产品契约要自包含。由此划边界：可继承的是契约 / Schema / 校验器 / 门禁，不可\n继承的是宿主运行时——搞混会废掉 §2.1 的 fail-closed import 扫描。\n\n**四个落点按依赖方向逐个判定：** 协议与 SDK 由基座以 @juhai/ 发布（合法，阻碍是\n@repo/* 命名）；分发通道归基础设施（合法，DEC-031 已批准，缺口是契约包 pin 无门禁\n——HR 与 OS 停在 rc.1、源仓已 rc.3，不像 kernel 有 L12 守着）；脚手架归框架（合法\n但受零依赖限制，只能给目录形状）；**把装载协议搬进基础设施不合法**——基础设施要\n编码基座契约即反向依赖，撞图 1 的唯一硬约束。本节只判合法性，不选落点。\n\n另记一条门禁抓不到的过期陈述：协议草案 §4 写「DEC-006 / 007 pending 期间」，两条\n实际都已 approved；L10 只抓「等 / 待 / 尚未裁决」措辞，这句不在模式内。该文是\nDEC-009 材料，处置留给 OS Owner，本次未动。\n\n需走 DEC / CHG 的两件已列入 §9 不裁决清单：装载协议是否切出改名发布、走哪条通道、\n谁做 Owner；草案 §4 两个契约由基座实现的排期。\n\n门禁复跑：L3 / L4 / L10 / L13 均为 0，仍只剩既有的 L9；新增的两条跨层链接\n（os.upstream.json、装载与对账协议草案）解析通过。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:54:12-07:00"}],"HeadCommit":{"Sha1":"f4540e69b1656232b677571584d3bd6ad57217be","Message":"docs(架构): 总架构增 §3.8——基座的装载 / 调用两个方向与接入能力归属\n\n起因是一次实际误读：`c433d57` 被当成「另一个 OS 产品」。它是数字员工OS 仓的提交\nSHA（2026-09-14），由 base-framework-os-product/os.upstream.json 连同 protocol\n1.3.0 钉住，用途是让「上游 OS 走动了」以 O0-SHA 显式变红。与基座并列的不是它，\n是 base-framework-os-product 这个仓——示例产品扩展加两道 OS 门禁，2026-09-10\n随 G26 自框架仓移出以保框架零依赖。误读本身说明这几件事在文档里没写清，本节补上。\n\n**两个方向。** PNG 给基座的定语「理解、编排、调用、执行」对应两个方向相反的接口：\n装载（OS → 应用）已有协议（KERNEL_EXTENSION_PROTOCOL_VERSION = 1.3.0，定义在 OS\n仓 packages/contracts/src/product-manifest.ts），真装 1 个；调用（应用 → OS）基本\n空白——协议草案 §4 已给出 ProductActivationIntent / ToolIntentReceipt 并写明「OS\n应提供、产品不再自建」，2026-09-18 复核 OS 仓 apps/ 下仍无 product-activations\n路由，智服已在自己仓内造了一套。两个方向归属不同，混谈会把落点找错。\n\n**接入形态实测已分叉成五种：** product.manifest.json（示例 / 设备云 / 知识云）、\nmanifest.json（HR，另有 ddl-authority / distribution / readiness 三个独有文件且无\npackage.json）、manifest.ts（工单，另有 kernel.products.entry.json）。只有示例与\n设备云带 os-host 测试配置。\n\n**为什么「每个应用继承一份」今天做不到**（两条方向相反的硬事实）：OS 的包是\n@repo/* 仓内作用域名（@repo/contracts@1.16.0，与工单仓同名包撞名），外仓依赖不了，\n只能照抄；而协议 §2.1 又反向禁止产品导入宿主私有 @repo/*，宿主能力一律由组合根\n注入、产品契约要自包含。由此划边界：可继承的是契约 / Schema / 校验器 / 门禁，不可\n继承的是宿主运行时——搞混会废掉 §2.1 的 fail-closed import 扫描。\n\n**四个落点按依赖方向逐个判定：** 协议与 SDK 由基座以 @juhai/ 发布（合法，阻碍是\n@repo/* 命名）；分发通道归基础设施（合法，DEC-031 已批准，缺口是契约包 pin 无门禁\n——HR 与 OS 停在 rc.1、源仓已 rc.3，不像 kernel 有 L12 守着）；脚手架归框架（合法\n但受零依赖限制，只能给目录形状）；**把装载协议搬进基础设施不合法**——基础设施要\n编码基座契约即反向依赖，撞图 1 的唯一硬约束。本节只判合法性，不选落点。\n\n另记一条门禁抓不到的过期陈述：协议草案 §4 写「DEC-006 / 007 pending 期间」，两条\n实际都已 approved；L10 只抓「等 / 待 / 尚未裁决」措辞，这句不在模式内。该文是\nDEC-009 材料，处置留给 OS Owner，本次未动。\n\n需走 DEC / CHG 的两件已列入 §9 不裁决清单：装载协议是否切出改名发布、走哪条通道、\n谁做 Owner；草案 §4 两个契约由基座实现的排期。\n\n门禁复跑：L3 / L4 / L10 / L13 均为 0，仍只剩既有的 L9；新增的两条跨层链接\n（os.upstream.json、装载与对账协议草案）解析通过。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:54:12-07:00"},"CompareURL":"luoanwu/platform-governance/compare/34c5fecec1074d5a01acd8b184e70c63c5ed947b...f4540e69b1656232b677571584d3bd6ad57217be","Len":1}...
|
1789718060
|
Edit
Delete
|
|
30902
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"34c5fecec {"Commits":[{"Sha1":"34c5fecec1074d5a01acd8b184e70c63c5ed947b","Message":"feat(governance): 报告登记门禁补 R7 / R8——让它去读登记里早就写着的 binding 口径\n\n治理层是给全工作区定证据纪律的那一层，却没对自己的报告用同一条。登记表里每条 latest\n都写着 `\"binding\": \"clean-head\"`，note 也写着「只在 clean HEAD 上提交」，但 R1—R6 里\n没有任何一条去读它——R4 只查 provenance.gitSha 存不存在。\n\n于是 `基础目录统一审计.latest.json` 绑在 589467e 且 worktreeDirty=true（当时 9 个脏\n文件）被提交了，一路绿灯。一份绑在脏树上的报告证明不了任何一个提交态，它既不对应\n589467e 的树，也不对应此后任何提交。已于 ad86507 在干净树上重跑回绑。\n\n R7 BINDING_NOT_CLEAN 声明 clean-head 却自称绑在脏树上，且该报告本身已提交\n R8 BINDING_OFF_HISTORY 绑定的 SHA 不在当前 HEAD 的历史上（换基 / 重写历史后遗留）\n\n两条都只对**登记已声明绑定口径**的条目判——不替没声明的条目另立规矩。R7 跳过报告\n自身未提交的情况：那是在途产物，也可能是并行会话正在跑的中间态（本工作区常有 2—3\n个会话同时写 reports/），判红只会误伤对方的回绑，与真源仓偏差 #34-B 同一口径。\nR8 取不到祖先关系时整条不判，非 git 环境退回原判据。缺 gitSha 的报告由 R4 单独报，\n不叠加成三条。\n\n登记门禁测试 8 → 13，治理层全量 383（377 过 / 6 跳 / 0 失败）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:49:24-07:00"},{"Sha1":"ad86507edf7f52241c3918e8c634b804e8b81b58","Message":"chore(reports): 治理层四份 latest 回绑 @ 44686cc\n\n`基础目录统一审计.latest.json` 此前绑在 589467e 且 worktreeDirty=true（当时 9 个脏\n文件）并被提交——登记里写着 binding=clean-head、note 也写着「只在 clean HEAD 上提交」，\n但没有任何检查去读，于是这份证明不了任何提交态的报告一直挂在顶层。\n\n`pnpm run audit` 一次编排写三份，加上它触发的 check-gate-consumption 共四份，本轮一并\n在干净树上重跑回绑（governanceProvenance 忽略 latest 报告自身，故并行会话在途的报告\n不影响绑定判定）。四份现均为 44686cc / dirty=false。\n\n结论未变：审计仍 BLOCKED（FACT_DECISIONS_PENDING_OR_UNKNOWN、\nASSET_EVIDENCE_REVIEW_REQUIRED），assets 去掉 ageHours 后逐字相同。\n\n另记一笔：`门禁消费.latest.json` 相对它自己的输入 `门禁消费登记.json` 已经陈旧——\n登记表里 C-13 / C-14 那段 2026-09-18 的文字早已提交，报告却停在旧文本，无人察觉。\n这是新鲜度缺口，不是绑定缺口，本轮不处置。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:47:30-07:00"}],"HeadCommit":{"Sha1":"34c5fecec1074d5a01acd8b184e70c63c5ed947b","Message":"feat(governance): 报告登记门禁补 R7 / R8——让它去读登记里早就写着的 binding 口径\n\n治理层是给全工作区定证据纪律的那一层，却没对自己的报告用同一条。登记表里每条 latest\n都写着 `\"binding\": \"clean-head\"`，note 也写着「只在 clean HEAD 上提交」，但 R1—R6 里\n没有任何一条去读它——R4 只查 provenance.gitSha 存不存在。\n\n于是 `基础目录统一审计.latest.json` 绑在 589467e 且 worktreeDirty=true（当时 9 个脏\n文件）被提交了，一路绿灯。一份绑在脏树上的报告证明不了任何一个提交态，它既不对应\n589467e 的树，也不对应此后任何提交。已于 ad86507 在干净树上重跑回绑。\n\n R7 BINDING_NOT_CLEAN 声明 clean-head 却自称绑在脏树上，且该报告本身已提交\n R8 BINDING_OFF_HISTORY 绑定的 SHA 不在当前 HEAD 的历史上（换基 / 重写历史后遗留）\n\n两条都只对**登记已声明绑定口径**的条目判——不替没声明的条目另立规矩。R7 跳过报告\n自身未提交的情况：那是在途产物，也可能是并行会话正在跑的中间态（本工作区常有 2—3\n个会话同时写 reports/），判红只会误伤对方的回绑，与真源仓偏差 #34-B 同一口径。\nR8 取不到祖先关系时整条不判，非 git 环境退回原判据。缺 gitSha 的报告由 R4 单独报，\n不叠加成三条。\n\n登记门禁测试 8 → 13，治理层全量 383（377 过 / 6 跳 / 0 失败）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:49:24-07:00"},"CompareURL":"luoanwu/platform-governance/compare/44686cc39752ceb9b94fbfed9169b6fe0a2d6ea3...34c5fecec1074d5a01acd8b184e70c63c5ed947b","Len":2}...
|
1789717791
|
Edit
Delete
|
|
30884
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"44686cc39 {"Commits":[{"Sha1":"44686cc39752ceb9b94fbfed9169b6fe0a2d6ea3","Message":"docs(计划): 登记两轮治理改动——未登记报告纳入机器记账、验收 runner 前置失败不得销毁证据\n\n§8 增两条当日记录：\n\n一、`check:evidence` 的未登记类此前只由 info 级 `SCOPE_UNDECLARED` 记着，处置口径\n写在 `evidence-scopes.json` 的散文里、无人机器核对，48 份报告里 13 份在此（27%）。\n已按例外机制的同一纪律搬成 `undeclared` 逐条登记，三条新规则判红两端。顺带订正\n「聚合报告」类的两处误归类，`runtime-governance` 与 `conformance-differential`\n转入正式登记（35 → 37）。\n\n二、为重跑被注释提交打过期的验收报告，照 runbook §7 逐字执行，复现两个缺陷：\n配方缺七个 `DATABASE_URL_\u003cMODULE\u003e`；三个验收 runner 在前置守卫失败时仍无条件写\n报告，把 775/775 覆盖成 failed。后者是证据完整性问题，三处已统一修并各自实测。\n\n三份需真实容器的验收报告仍未重跑，理由写在条目里：其设计中的保鲜路径是 CI。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:29:15-07:00"}],"HeadCommit":{"Sha1":"44686cc39752ceb9b94fbfed9169b6fe0a2d6ea3","Message":"docs(计划): 登记两轮治理改动——未登记报告纳入机器记账、验收 runner 前置失败不得销毁证据\n\n§8 增两条当日记录：\n\n一、`check:evidence` 的未登记类此前只由 info 级 `SCOPE_UNDECLARED` 记着，处置口径\n写在 `evidence-scopes.json` 的散文里、无人机器核对，48 份报告里 13 份在此（27%）。\n已按例外机制的同一纪律搬成 `undeclared` 逐条登记，三条新规则判红两端。顺带订正\n「聚合报告」类的两处误归类，`runtime-governance` 与 `conformance-differential`\n转入正式登记（35 → 37）。\n\n二、为重跑被注释提交打过期的验收报告，照 runbook §7 逐字执行，复现两个缺陷：\n配方缺七个 `DATABASE_URL_\u003cMODULE\u003e`；三个验收 runner 在前置守卫失败时仍无条件写\n报告，把 775/775 覆盖成 failed。后者是证据完整性问题，三处已统一修并各自实测。\n\n三份需真实容器的验收报告仍未重跑，理由写在条目里：其设计中的保鲜路径是 CI。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:29:15-07:00"},"CompareURL":"luoanwu/platform-governance/compare/6bf1e7d524a70a3ecd012c29ea6c627866c01c5c...44686cc39752ceb9b94fbfed9169b6fe0a2d6ea3","Len":1}...
|
1789716558
|
Edit
Delete
|
|
30882
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"0abd15c28 {"Commits":[{"Sha1":"0abd15c28dd2b92c583289934db660363b202fd4","Message":"chore(reports): check:evidence 回绑 @ 34fresh\n\n37 份已登记、34 份新鲜、0 error；余 3 份过期为需真实容器重跑的验收报告\n（runtime-acceptance / mainline-acceptance / revocation-sla，warn 按设计）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:28:32-07:00"},{"Sha1":"a4b3d356f9d2f7a831198787a5b19194f4f763c8","Message":"chore(reports): 双后端差分验收重跑回绑 @ 746f323\n\n本轮把 conformance-differential 从「未登记的聚合报告」订正为登记条目后，它第一次\n被新鲜度门禁判为过期——触发者正是同轮改到的 check-conformance-differential.mjs，\n说明登记的作用域是对的。真实 DB + Redis 重跑：7 checkpoints × 2 backends，\n逐字段 0 differences。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:28:10-07:00"},{"Sha1":"746f323c203531c4ced5f3ca60ea29a4c53c7c4d","Message":"chore(reports): 工作台两份快照回绑 @ adda76a\n\nrunbook §7 的订正落在运维快照作用域内（docs/runbook.md），经 reports:rebind 在\nHEAD 的干净检出里重跑，两份同源快照一并带回。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:27:33-07:00"},{"Sha1":"adda76acfd8fa046e1f2a320c15b5384b4ddef73","Message":"fix(runtime): 前置守卫失败不得覆盖验收报告，并补「一模块一库」检查\n\n按 docs/runbook.md §7 的配方逐字重跑 runtime 验收，当场复现两个缺陷。\n\n一、配方本身是错的。§7 只写了 DATABASE_URL 与 REDIS_URL，而七个模块的\nprisma/schema.prisma 各自写死 url = env(\"DATABASE_URL_\u003cMODULE\u003e\")。照它执行时，\n前两步 nestjs/fastify-migrate-deploy 会通过，第三步 migrate-status 才以\nPrisma P1012 \"Environment variable not found\" 失败——报错指向 wasm 校验，\n完全看不出是环境没配齐。§7 已补齐七个变量，并写明 prepare-mainline.mjs\n只供给主线验收要的 5 个（fact/permission/audit/owner/consumer），差的四个要自己建。\n\n二、更要紧的：那次失败把 runtime-acceptance.latest.json 覆盖成了 status:failed。\n775/775 那份真实结论就此丢失，靠 git checkout 才救回——差一点被当作真实回归提交上去。\n根因是三个 runner 都无条件 writeFileSync：前置守卫失败意味着**验收根本没跑**，\n此时写一份 failed 报告不是诚实，是用操作者的环境问题顶替代码的结论，\n并且顺手销毁了上一次真的跑过的证据。三个 runner 统一改为前置失败即退出、不碰报告：\n check-runtime-acceptance 实测：报告 sha256 前后一致\n check-ui-acceptance 实测：缺 REDIS_URL 时报告 sha256 前后一致\n check-conformance-differential 实测：同上\n\n新增 preflightModuleDatabases + requiredModuleDatabaseEnvNames，把缺失的七个\n变量在动任何东西之前一次列清。变量名**现取自各模块 prisma/schema.prisma**，\n不写死清单：模块增减自动跟上，也避免抄错 ai-gateway 那个没有下划线的\nDATABASE_URL_AIGATEWAY。该检查只接进 runtime 验收——UI 与差分验收跑的是\n两个 app，不走一模块一库。\n\npnpm check 全链通过。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:26:48-07:00"}],"HeadCommit":{"Sha1":"0abd15c28dd2b92c583289934db660363b202fd4","Message":"chore(reports): check:evidence 回绑 @ 34fresh\n\n37 份已登记、34 份新鲜、0 error；余 3 份过期为需真实容器重跑的验收报告\n（runtime-acceptance / mainline-acceptance / revocation-sla，warn 按设计）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:28:32-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/df678ac14f88628a7bc48ce3b53e248f1b257e8b...0abd15c28dd2b92c583289934db660363b202fd4","Len":4}...
|
1789716516
|
Edit
Delete
|
|
30881
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"6bf1e7d52 {"Commits":[{"Sha1":"6bf1e7d524a70a3ecd012c29ea6c627866c01c5c","Message":"docs(架构): 总架构配套图与口径订正——重画图 1 / 图 2，新增图 4 关系全景\n\n企业应用群总架构-2026-09-17.md 与图 3 已随 5b7a1c0 提交（暂存区被并行会话的\n提交扫走，不是有意合并进那个范围），本提交补齐其余部分。\n\n**图 1 / 图 2 重画。** 图 1 原写「基础设施（23 项）」、画着「四候选」框，应用名\n还是改名前的「嗨设首席」；图 2 把 OS 标成「Projection 未实现」、干线标成「持久\n投递待裁决」。四处都已被后续裁决推翻：四候选 2026-09-16 撤销（CHG-015）、Catalog\n现为 21 条，OS 投影 2026-09-14 已在 main，三条前置 DEC 全部批准。图 1 另加一条\n跨层治理带，把「平台治理不是第六层」画进图面。\n\n**新增图 4 现存关系全景。** 逐条查实全组的边：左半五类有边的关系，右半七条负\n结论。查证推翻了总架构初稿 §3.1 的一处判断——初稿用 grep '\"@juhai/…\"' 取依赖，\n同时命中了 name 字段，于是把报价的 @juhai/public-file 与智服的\n@juhai/service-contracts 当成「对下层消费」，实际两者都是该仓自己的 workspace 包。\n按带版本号的依赖项重取：四个业务应用对下层的跨仓包依赖为零，全组跨仓包依赖只有\nHR、OS、enterprise-platform 三个消费仓。\n\n由此改口径的另有两条：\n\n- 三项 module_e2（文件 / 通知 / 配置）在工单仓里是 @repo/* private 包，仓外零\n 消费者，只被本仓 api-fastify / api-nestjs 依赖——「已被业务应用消费」不成立，\n 三消费者规则的计数应从 0 算起\n- relationships.json 的 relationships 为 {}，关系在机器层面没有登记；图 4 的边\n 全部由 package.json / kernel.products.json / facts.json / runtime/modules.json\n 推导，今天没有门禁能发现某条关系消失\n\n**四份下游文档按活跃 Catalog 改齐：** 分层与装配模型（23→21、派生仓 3→0、1/23\n的分母；§3.1 是带日期快照，只加订正行不改写正文）、基础/CLAUDE.md 分层词表段、\n基础/README 头段（「已退役 15 个、现存 3 个」是 09-16 白天的中间态）、企业领域\n应用层 README（引的 platform-contracts 已于 2026-09-16 物理移除，事实数 1→11）。\n平台治理 README 收纳表随新增文件改计数（基础/ 根文件 63→65，根散文件 9→10）。\n\n复跑：工作区一致性门禁 L1—L13 只剩既有的 L9（8 个仓共用 compose project 名\napi-nestjs），L4 / L13 均为 0；治理层 npm test 378 项 0 失败；check-domain-layer\n13/13 条全在基线内、无新增。未动任何冻结物、审批记录与机器 Catalog。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:20:28-07:00"}],"HeadCommit":{"Sha1":"6bf1e7d524a70a3ecd012c29ea6c627866c01c5c","Message":"docs(架构): 总架构配套图与口径订正——重画图 1 / 图 2，新增图 4 关系全景\n\n企业应用群总架构-2026-09-17.md 与图 3 已随 5b7a1c0 提交（暂存区被并行会话的\n提交扫走，不是有意合并进那个范围），本提交补齐其余部分。\n\n**图 1 / 图 2 重画。** 图 1 原写「基础设施（23 项）」、画着「四候选」框，应用名\n还是改名前的「嗨设首席」；图 2 把 OS 标成「Projection 未实现」、干线标成「持久\n投递待裁决」。四处都已被后续裁决推翻：四候选 2026-09-16 撤销（CHG-015）、Catalog\n现为 21 条，OS 投影 2026-09-14 已在 main，三条前置 DEC 全部批准。图 1 另加一条\n跨层治理带，把「平台治理不是第六层」画进图面。\n\n**新增图 4 现存关系全景。** 逐条查实全组的边：左半五类有边的关系，右半七条负\n结论。查证推翻了总架构初稿 §3.1 的一处判断——初稿用 grep '\"@juhai/…\"' 取依赖，\n同时命中了 name 字段，于是把报价的 @juhai/public-file 与智服的\n@juhai/service-contracts 当成「对下层消费」，实际两者都是该仓自己的 workspace 包。\n按带版本号的依赖项重取：四个业务应用对下层的跨仓包依赖为零，全组跨仓包依赖只有\nHR、OS、enterprise-platform 三个消费仓。\n\n由此改口径的另有两条：\n\n- 三项 module_e2（文件 / 通知 / 配置）在工单仓里是 @repo/* private 包，仓外零\n 消费者，只被本仓 api-fastify / api-nestjs 依赖——「已被业务应用消费」不成立，\n 三消费者规则的计数应从 0 算起\n- relationships.json 的 relationships 为 {}，关系在机器层面没有登记；图 4 的边\n 全部由 package.json / kernel.products.json / facts.json / runtime/modules.json\n 推导，今天没有门禁能发现某条关系消失\n\n**四份下游文档按活跃 Catalog 改齐：** 分层与装配模型（23→21、派生仓 3→0、1/23\n的分母；§3.1 是带日期快照，只加订正行不改写正文）、基础/CLAUDE.md 分层词表段、\n基础/README 头段（「已退役 15 个、现存 3 个」是 09-16 白天的中间态）、企业领域\n应用层 README（引的 platform-contracts 已于 2026-09-16 物理移除，事实数 1→11）。\n平台治理 README 收纳表随新增文件改计数（基础/ 根文件 63→65，根散文件 9→10）。\n\n复跑：工作区一致性门禁 L1—L13 只剩既有的 L9（8 个仓共用 compose project 名\napi-nestjs），L4 / L13 均为 0；治理层 npm test 378 项 0 失败；check-domain-layer\n13/13 条全在基线内、无新增。未动任何冻结物、审批记录与机器 Catalog。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:20:28-07:00"},"CompareURL":"luoanwu/platform-governance/compare/650711d6cb4d48b6d77a3edea998d26d60f87567...6bf1e7d524a70a3ecd012c29ea6c627866c01c5c","Len":1}...
|
1789716065
|
Edit
Delete
|
|
30872
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"df678ac14 {"Commits":[{"Sha1":"df678ac14f88628a7bc48ce3b53e248f1b257e8b","Message":"chore(reports): check:evidence 回绑 @ e9d0779\n\n37 份已登记（+2：runtime-governance error、conformance-differential warn）、34 份新鲜、\n11 份未登记且全部有处置说明、evidenceUnaccounted=0。3 条 warn 为待重跑的验收报告。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:14:15-07:00"},{"Sha1":"e9d07790cee2a6780ecc755cb6b6efe70890a727","Message":"feat(governance): 未登记的报告也要逐条记账——把散文口径搬成机器判据，并订正两处误归类\n\ncheck:evidence 的 SCOPE_UNDECLARED 是 info，本身不拦任何东西。于是「未登记作用域」\n这一类可以无声增长，而它恰恰是新鲜度门禁完全看不见的那部分：实测 48 份报告里 13 份\n在此，占 27%。这 13 份为什么不登，此前只写在 evidence-scopes.json 的 note 散文里，\n没有任何机器核对——明天多出一份新报告，它会自动落进这个黑箱，全链门禁一片绿。\n\n按例外机制的同一纪律（G-13：自带出口）把口径搬成逐条登记的 undeclared：\n UNDECLARED_UNACCOUNTED 报告存在，既无作用域也无账（新洞的唯一入口，判红）\n UNDECLARED_STALE 账指向的报告已登记作用域，或文件已不存在（判红）\n UNDECLARED_INCOMPLETE 缺 report/class/reason/recordedAt/resolveBy，或 class 非法\nclass 只允许 self / aggregate / provenance-missing / one-off-snapshot / no-generator。\n记账不是豁免：后两类的出口只有「补生成脚本后转正式登记」或「退役该报告」二选一。\n\n逐份走读生成脚本时发现 note 的 ③「聚合报告」类有两处归错，两份因此一直不受监控：\n runtime-governance 不聚合任何报告，直接走读 apps/ 源码与两份 prisma schema\n 判租户接线、outbox 事务、直写发布；纯静态零容器 → error\n conformance-differential 在 apps/api-nestjs 真跑 vitest，用例同时 import 了\n NestJS 与 Fastify 两个后端的装配；要真 DB → warn\n两份均已转入 entries（35 → 37 条），③ 实际只剩 governance 与 governance-status。\n\nclass=self 不查文件存在性：本门禁的报告是判定跑完之后才写的，判定当时它可以合法地\n不存在。这是集成用例抓出来的——起初一视同仁地查，临时仓第一次运行就被自己判红。\n\n门禁测试 277 → 283（新增 6 条负向用例，含「新报告悄悄落进 reports/ 必须判红」的\n集成证明）。pnpm check 全链通过；check:evidence 37 份已登记、34 份新鲜、11 份未登记\n且全部有处置说明、0 份无说明。余 3 份过期是注释提交 c9e0f59 落进验收作用域所致，\n须重跑重绑，不在本提交。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:13:43-07:00"}],"HeadCommit":{"Sha1":"df678ac14f88628a7bc48ce3b53e248f1b257e8b","Message":"chore(reports): check:evidence 回绑 @ e9d0779\n\n37 份已登记（+2：runtime-governance error、conformance-differential warn）、34 份新鲜、\n11 份未登记且全部有处置说明、evidenceUnaccounted=0。3 条 warn 为待重跑的验收报告。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-18T00:14:15-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/c9e0f59d11a533e7e5e3386f80eff8fb6d6ba6c8...df678ac14f88628a7bc48ce3b53e248f1b257e8b","Len":2}...
|
1789715664
|
Edit
Delete
|
|
30854
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"650711d6c {"Commits":[{"Sha1":"650711d6cb4d48b6d77a3edea998d26d60f87567","Message":"docs(框架): FR-2 / FR-3 一并提交框架 Owner，附 2026-09-18 复核与新证据\n\n两项 2026-09-12 提出后一直未受理。复核后确认**仍成立**，而不是直接重提一份\n可能已过期的请求——内核 0.22.0 实测：\n\n- FR-2：createLocalPlatformPorts 的 auditLog 依旧是同步回调，实现只能表达\n 「记了 / 没记」，Q04 要求的「与业务写同事务落盘」在框架层没有接缝\n- FR-3：RealtimeEnvelope 仍只有 seq 与 context，无 ordering / aggregateKey /\n aggregateVersion，replay-buffer 只能按全局 seq 检缺口\n\n六天里本仓积累的新证据让两项更有说服力：\n\n- FR-2：已落地 audit-commit-protocol.ts（2026-09-16），把 Q04 两模式与六类\n 故障注入点写成机器可读的 AUDIT_FAULT_MATRIX 并以夹具 + 集成用例覆盖 M6 侧；\n 其中 F1 / F2 / F4 与 F6 的 Owner 侧义务只能由框架写链提供接缝，模块侧补不了。\n 受理后该矩阵即可两端对账\n- FR-3：DEC-013—016 Approval gate 材料 ④ 已备 Envelope v2 候选 schema\n （aggregate_version + previous_aggregate_version 必填，检查器判\n REQUIRED_PROPERTY_ADDED ⇒ 需发新主版本）。那是契约侧的轨，与本请求的内核侧\n 可选字段互不冲突，两轨可并行——这一点此前没写清，容易被误读为重复请求\n\n两项都向后兼容、默认关闭，不改现有派生仓行为。README 送达入口与原始 FR 正文\n两处同步补记，避免说法不一；并明写请求给一个受理或不受理的结论——长期悬空会\n让本仓的 Q04 / Q06 落地一直只能做半边。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:59:35-07:00"}],"HeadCommit":{"Sha1":"650711d6cb4d48b6d77a3edea998d26d60f87567","Message":"docs(框架): FR-2 / FR-3 一并提交框架 Owner，附 2026-09-18 复核与新证据\n\n两项 2026-09-12 提出后一直未受理。复核后确认**仍成立**，而不是直接重提一份\n可能已过期的请求——内核 0.22.0 实测：\n\n- FR-2：createLocalPlatformPorts 的 auditLog 依旧是同步回调，实现只能表达\n 「记了 / 没记」，Q04 要求的「与业务写同事务落盘」在框架层没有接缝\n- FR-3：RealtimeEnvelope 仍只有 seq 与 context，无 ordering / aggregateKey /\n aggregateVersion，replay-buffer 只能按全局 seq 检缺口\n\n六天里本仓积累的新证据让两项更有说服力：\n\n- FR-2：已落地 audit-commit-protocol.ts（2026-09-16），把 Q04 两模式与六类\n 故障注入点写成机器可读的 AUDIT_FAULT_MATRIX 并以夹具 + 集成用例覆盖 M6 侧；\n 其中 F1 / F2 / F4 与 F6 的 Owner 侧义务只能由框架写链提供接缝，模块侧补不了。\n 受理后该矩阵即可两端对账\n- FR-3：DEC-013—016 Approval gate 材料 ④ 已备 Envelope v2 候选 schema\n （aggregate_version + previous_aggregate_version 必填，检查器判\n REQUIRED_PROPERTY_ADDED ⇒ 需发新主版本）。那是契约侧的轨，与本请求的内核侧\n 可选字段互不冲突，两轨可并行——这一点此前没写清，容易被误读为重复请求\n\n两项都向后兼容、默认关闭，不改现有派生仓行为。README 送达入口与原始 FR 正文\n两处同步补记，避免说法不一；并明写请求给一个受理或不受理的结论——长期悬空会\n让本仓的 Q04 / Q06 落地一直只能做半边。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:59:35-07:00"},"CompareURL":"luoanwu/platform-governance/compare/5b7a1c033ef290f7fa48f58bf0eae5eaa38150aa...650711d6cb4d48b6d77a3edea998d26d60f87567","Len":1}...
|
1789714779
|
Edit
Delete
|
|
30853
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"5b7a1c033 {"Commits":[{"Sha1":"5b7a1c033ef290f7fa48f58bf0eae5eaa38150aa","Message":"docs(框架): FR-6 提交框架 Owner——在送达入口 README 登记待受理\n\nFR-6 的正文 2026-09-18 已写入 基础设施仓框架变更请求-2026-09-12.md（48c5a9c），\n但那份文档本身不是送达入口：框架 Owner 是从 工程基础框架/README.md 的变更请求\n条目看待办的（FR-1 / FR-4 / FR-5 与另两份平台治理的请求都是这么送达并受理的）。\n此前只写了正文没更新入口，等于没提。\n\nREADME 该条更新为：FR-1 已落地 0.19.0、FR-4 / FR-5 已落地 0.20.0、**FR-6 新提\n待受理**、FR-2 / FR-3 仍未受理，并就地摘出 FR-6 的请求与依据：\n\ncheck:docs-truth 的四级证据新鲜度须按「该级别是否已接线」判定。起因是基础设施仓\n同步到 0.22.0 时实测，新增的制品部署 / 生产运维两级依赖 apps/api-nestjs/Dockerfile，\n而那份里有 COPY packages/kernel/package.json——该仓按 DEC-010 / DEC-031 只 exact\npin 已发布的 @juhai/kernel，仓内不存在 packages/kernel，构建必然失败，这道门禁\n在该仓只能恒红。恒红门禁不携带信息，正是 0.22.0 自己要治的失效模式的反面。\n\n请求对应 pnpm 脚本存在则断言照旧、不存在则 SKIPPED，不接受占位报告；接线了却让\n报告陈旧的仓仍照旧判红。影响面是框架仓单点改动加一条负向用例。基础设施仓已本地\n修正并在改动处注释指向 FR-6，受理后删除该段回到框架原文。\n\n未触碰框架仓 base-framework（其工作区有 12 项他人在途改动）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:57:57-07:00"}],"HeadCommit":{"Sha1":"5b7a1c033ef290f7fa48f58bf0eae5eaa38150aa","Message":"docs(框架): FR-6 提交框架 Owner——在送达入口 README 登记待受理\n\nFR-6 的正文 2026-09-18 已写入 基础设施仓框架变更请求-2026-09-12.md（48c5a9c），\n但那份文档本身不是送达入口：框架 Owner 是从 工程基础框架/README.md 的变更请求\n条目看待办的（FR-1 / FR-4 / FR-5 与另两份平台治理的请求都是这么送达并受理的）。\n此前只写了正文没更新入口，等于没提。\n\nREADME 该条更新为：FR-1 已落地 0.19.0、FR-4 / FR-5 已落地 0.20.0、**FR-6 新提\n待受理**、FR-2 / FR-3 仍未受理，并就地摘出 FR-6 的请求与依据：\n\ncheck:docs-truth 的四级证据新鲜度须按「该级别是否已接线」判定。起因是基础设施仓\n同步到 0.22.0 时实测，新增的制品部署 / 生产运维两级依赖 apps/api-nestjs/Dockerfile，\n而那份里有 COPY packages/kernel/package.json——该仓按 DEC-010 / DEC-031 只 exact\npin 已发布的 @juhai/kernel，仓内不存在 packages/kernel，构建必然失败，这道门禁\n在该仓只能恒红。恒红门禁不携带信息，正是 0.22.0 自己要治的失效模式的反面。\n\n请求对应 pnpm 脚本存在则断言照旧、不存在则 SKIPPED，不接受占位报告；接线了却让\n报告陈旧的仓仍照旧判红。影响面是框架仓单点改动加一条负向用例。基础设施仓已本地\n修正并在改动处注释指向 FR-6，受理后删除该段回到框架原文。\n\n未触碰框架仓 base-framework（其工作区有 12 项他人在途改动）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:57:57-07:00"},"CompareURL":"luoanwu/platform-governance/compare/e393d2db524e2c16a517866d412b1d039c01640a...5b7a1c033ef290f7fa48f58bf0eae5eaa38150aa","Len":1}...
|
1789714680
|
Edit
Delete
|
|
30852
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"e393d2db5 {"Commits":[{"Sha1":"e393d2db524e2c16a517866d412b1d039c01640a","Message":"docs(计划): 偏差 #36 裁决「维持分离」，S-5 由过渡方案改判为终态\n\n目录负责人裁定选项 ①：模块依赖故障 ≠ 宿主运行时不健康。模块健康只在\nGET /api/platform/modules[/:id/health] 可见，/api/health 只反映宿主依赖。\n\n两处改判：\n- S-5 不再标「过渡方案」而是终态。这条分离由 e2e/identity-profile.test.ts\n 第 188—189 行的成对断言守着，是设计不是欠账；真源仓已给该用例补注释说明\n 裁决出处与那次 775 → 561 的实锤\n- 跨目录待办更新：FR-1—5 已随框架 0.22.0 全部到位（FR-1 健康子检查 0.19.0、\n FR-4 租户 claim 读序 0.20.0、FR-5 runner 三值 0.20.0），余 FR-6 待受理\n\n偏差 #36 移入已关闭，并在其中记下「FR 被受理 ≠ 它解锁的工作就该做」。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:49:17-07:00"}],"HeadCommit":{"Sha1":"e393d2db524e2c16a517866d412b1d039c01640a","Message":"docs(计划): 偏差 #36 裁决「维持分离」，S-5 由过渡方案改判为终态\n\n目录负责人裁定选项 ①：模块依赖故障 ≠ 宿主运行时不健康。模块健康只在\nGET /api/platform/modules[/:id/health] 可见，/api/health 只反映宿主依赖。\n\n两处改判：\n- S-5 不再标「过渡方案」而是终态。这条分离由 e2e/identity-profile.test.ts\n 第 188—189 行的成对断言守着，是设计不是欠账；真源仓已给该用例补注释说明\n 裁决出处与那次 775 → 561 的实锤\n- 跨目录待办更新：FR-1—5 已随框架 0.22.0 全部到位（FR-1 健康子检查 0.19.0、\n FR-4 租户 claim 读序 0.20.0、FR-5 runner 三值 0.20.0），余 FR-6 待受理\n\n偏差 #36 移入已关闭，并在其中记下「FR 被受理 ≠ 它解锁的工作就该做」。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:49:17-07:00"},"CompareURL":"luoanwu/platform-governance/compare/4be34a8411ad6ac48125865937ecad8ec32abc1f...e393d2db524e2c16a517866d412b1d039c01640a","Len":1}...
|
1789714159
|
Edit
Delete
|
|
30851
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"c9e0f59d1 {"Commits":[{"Sha1":"c9e0f59d11a533e7e5e3386f80eff8fb6d6ba6c8","Message":"test(identity): 给「模块故障不传导到宿主健康」的成对断言补设计说明\n\n2026-09-18 目录负责人裁决「维持分离」（开发计划 §6 偏差 #36）：模块依赖故障\n≠ 宿主运行时不健康。模块健康只在 GET /api/platform/modules[/:id/health] 可见，\n/api/health 只反映宿主自己的依赖。\n\n这条用例第 188—189 行的 degraded 与 200 是成对断言，此前没有注释说明它守的是\n一个有意的设计选择，读起来像实现欠账——于是框架 0.19.0 受理本仓 FR-1 后，我\n想当然地把模块 health() 汇入 /api/health，真实 DB 验收当场由 775 跌到 561，\n正是被这两行拦下的。\n\n补注释写明裁决出处与那次实锤，并点出「FR 被受理 ≠ 它解锁的工作就该做」。\n只加注释，断言与逻辑一字未动；typecheck 14/14。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:49:14-07:00"}],"HeadCommit":{"Sha1":"c9e0f59d11a533e7e5e3386f80eff8fb6d6ba6c8","Message":"test(identity): 给「模块故障不传导到宿主健康」的成对断言补设计说明\n\n2026-09-18 目录负责人裁决「维持分离」（开发计划 §6 偏差 #36）：模块依赖故障\n≠ 宿主运行时不健康。模块健康只在 GET /api/platform/modules[/:id/health] 可见，\n/api/health 只反映宿主自己的依赖。\n\n这条用例第 188—189 行的 degraded 与 200 是成对断言，此前没有注释说明它守的是\n一个有意的设计选择，读起来像实现欠账——于是框架 0.19.0 受理本仓 FR-1 后，我\n想当然地把模块 health() 汇入 /api/health，真实 DB 验收当场由 775 跌到 561，\n正是被这两行拦下的。\n\n补注释写明裁决出处与那次实锤，并点出「FR 被受理 ≠ 它解锁的工作就该做」。\n只加注释，断言与逻辑一字未动；typecheck 14/14。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:49:14-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/a63414ab085268e9570c6f51816a92718f2effeb...c9e0f59d11a533e7e5e3386f80eff8fb6d6ba6c8","Len":1}...
|
1789714158
|
Edit
Delete
|
|
30850
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"4be34a841 {"Commits":[{"Sha1":"4be34a8411ad6ac48125865937ecad8ec32abc1f","Message":"docs(计划): 登记偏差 #36——S-5 转正的前提不成立，试做后撤回\n\n框架同步带来 FR-1（HealthResponse.checks 可扩展，0.19.0 受理），本仓提该 FR\n的动机正是让 S-5 模块健康子检查从「过渡方案」转正、汇入 /api/health。\n\n试做后撤回：接线后真实 DB 验收由 775 跌到 561，e2e/identity-profile.test.ts\n两条失败。读用例才看清——其中一条的名字就写着「identity 依赖故障时宿主数据库\n仍健康」，第 188—189 行是成对断言：模块视图须报 degraded，而宿主 /api/health\n须仍为 200。本仓早已有意分离「模块依赖故障」与「宿主运行时不健康」，我的改动\n等于把模块故障提升成宿主故障，破坏既有设计。\n\n已完整撤回，真源仓回到 a63414a / dirty 0，未提交任何接线改动。\n\n§6 增偏差 #36，给三个处置选项（维持分离 / 汇入但不影响状态码，需另提 FR 支持\nobservational 子检查 / 按模块 state 分级传导，属设计问题）；裁决前 S-5 保持\n过渡方案，不因 FR-1 已受理就当作可转正。\n\n方法上的教训：FR 被受理 ≠ 它解锁的工作就该做；转正前要先读既有用例在守什么。\n这条与当日另一次假阴性（验收红先读错误码、别先怀疑代码）互补——那次是错怪了\n代码，这次是代码真错了，分辨靠的都是读断言而不是看数字。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:36:51-07:00"}],"HeadCommit":{"Sha1":"4be34a8411ad6ac48125865937ecad8ec32abc1f","Message":"docs(计划): 登记偏差 #36——S-5 转正的前提不成立，试做后撤回\n\n框架同步带来 FR-1（HealthResponse.checks 可扩展，0.19.0 受理），本仓提该 FR\n的动机正是让 S-5 模块健康子检查从「过渡方案」转正、汇入 /api/health。\n\n试做后撤回：接线后真实 DB 验收由 775 跌到 561，e2e/identity-profile.test.ts\n两条失败。读用例才看清——其中一条的名字就写着「identity 依赖故障时宿主数据库\n仍健康」，第 188—189 行是成对断言：模块视图须报 degraded，而宿主 /api/health\n须仍为 200。本仓早已有意分离「模块依赖故障」与「宿主运行时不健康」，我的改动\n等于把模块故障提升成宿主故障，破坏既有设计。\n\n已完整撤回，真源仓回到 a63414a / dirty 0，未提交任何接线改动。\n\n§6 增偏差 #36，给三个处置选项（维持分离 / 汇入但不影响状态码，需另提 FR 支持\nobservational 子检查 / 按模块 state 分级传导，属设计问题）；裁决前 S-5 保持\n过渡方案，不因 FR-1 已受理就当作可转正。\n\n方法上的教训：FR 被受理 ≠ 它解锁的工作就该做；转正前要先读既有用例在守什么。\n这条与当日另一次假阴性（验收红先读错误码、别先怀疑代码）互补——那次是错怪了\n代码，这次是代码真错了，分辨靠的都是读断言而不是看数字。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T23:36:51-07:00"},"CompareURL":"luoanwu/platform-governance/compare/9a959dc86d101a8abfcd42db52f679d5dbb8de8c...4be34a8411ad6ac48125865937ecad8ec32abc1f","Len":1}...
|
1789713415
|
Edit
Delete
|
|
30825
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"9a959dc86 {"Commits":[{"Sha1":"9a959dc86d101a8abfcd42db52f679d5dbb8de8c","Message":"docs(计划): 框架同步 0.16.1 → 0.22.0 完成，一致性门禁 L12 清零\n\n七批里 F-2a / F-2b / F-3 / F-4 / F-7 完成，F-1 证伪取消，F-5 撤回并提 FR-6，\nF-6 并入。真源仓 a63414a。\n\n验收（clean 765300b，与升级前逐项持平）：runtime 775/775 且租户 RLS enforce、\nUI 5/5、主线四套件 11/17/19/18、恢复演练 passed、端口一致性 41/41，\ncheck:evidence passed 35 份全新鲜 / 0 过期 / 0 例外。\nmigration:record --mode sync 出凭证 frameworkVersion=0.22.0。\n\n立项书补两节：§5.2 执行记录（含同步过程中撞到的三件事：git show 重定向会造\n空文件、并入上游 CHANGELOG 带来跨仓断链、治理测试锚定旧 pin），§5.3 三级验收\n结果，并记下两次教训——\n- 一次假阴性：首轮报 682/775 失败 5 条，读错误码才发现是环境缺 MAINLINE_CHAOS_\n CONTAINER，不是代码回归。验收红了先读错误码，别先怀疑代码\n- 一次自伤：git checkout -- runtime/reports 连带把真跑出的六份验收证据还原成\n 旧版，只能重跑找回；还原前须按 provenance.gitSha 分辨本轮产出与脏绑\n\n常驻问题 L9 / L12 两条降为 L9 一条（8 个上层应用仓的 compose 身份，待 N-11）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T22:40:49-07:00"},{"Sha1":"48c5a9c821798dbd7b2aae8355a55aba95d68c20","Message":"docs(框架): 提 FR-6——四级新鲜度须按「该级别是否已接线」断言\n\n框架 0.22.0 把 check:docs-truth 的证据新鲜度断言由两级扩到四级，新增的部署 /\n生产两级在本仓无法成立：其证据生产者依赖 apps/api-nestjs/Dockerfile，而那份\n里写着 COPY packages/kernel/package.json——本仓按 DEC-010 / DEC-031 的分发口径\n只 exact pin 已发布的 @juhai/kernel，仓内不存在 packages/kernel，2026-09-18\n实测该镜像构建必然失败。本仓自己的制品镜像是另一套架构（runtime/Dockerfile，\nD-1），并存会变成两条镜像路径。\n\n结果是这道门禁在本仓只能恒红，而恒红门禁不携带信息——正是框架自己在 0.22.0\n里要治的失效模式的反面。\n\n请求：四级断言按该级别是否已接线判定——对应 pnpm 脚本在 package.json 里存在\n则断言照旧，不存在则打印 SKIPPED 并跳过。不接受占位报告，与 doctrine「诚实的\n空，不得用占位报告转绿」一致；接线了却让报告陈旧的仓仍照旧判红，门禁的牙一\n颗不少。\n\n临时处置已在本仓 runtime/scripts/check-docs-truth.mjs 落地，改动处带注释指向\n本 FR，框架受理后删除该段回到框架原文（同 FR-4 / FR-5 的先例）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T21:53:48-07:00"}],"HeadCommit":{"Sha1":"9a959dc86d101a8abfcd42db52f679d5dbb8de8c","Message":"docs(计划): 框架同步 0.16.1 → 0.22.0 完成，一致性门禁 L12 清零\n\n七批里 F-2a / F-2b / F-3 / F-4 / F-7 完成，F-1 证伪取消，F-5 撤回并提 FR-6，\nF-6 并入。真源仓 a63414a。\n\n验收（clean 765300b，与升级前逐项持平）：runtime 775/775 且租户 RLS enforce、\nUI 5/5、主线四套件 11/17/19/18、恢复演练 passed、端口一致性 41/41，\ncheck:evidence passed 35 份全新鲜 / 0 过期 / 0 例外。\nmigration:record --mode sync 出凭证 frameworkVersion=0.22.0。\n\n立项书补两节：§5.2 执行记录（含同步过程中撞到的三件事：git show 重定向会造\n空文件、并入上游 CHANGELOG 带来跨仓断链、治理测试锚定旧 pin），§5.3 三级验收\n结果，并记下两次教训——\n- 一次假阴性：首轮报 682/775 失败 5 条，读错误码才发现是环境缺 MAINLINE_CHAOS_\n CONTAINER，不是代码回归。验收红了先读错误码，别先怀疑代码\n- 一次自伤：git checkout -- runtime/reports 连带把真跑出的六份验收证据还原成\n 旧版，只能重跑找回；还原前须按 provenance.gitSha 分辨本轮产出与脏绑\n\n常驻问题 L9 / L12 两条降为 L9 一条（8 个上层应用仓的 compose 身份，待 N-11）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T22:40:49-07:00"},"CompareURL":"luoanwu/platform-governance/compare/d70b1e5b6f3ed902a64adac86d0a4d953cadddb8...9a959dc86d101a8abfcd42db52f679d5dbb8de8c","Len":2}...
|
1789710053
|
Edit
Delete
|
|
30824
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"a63414ab0 {"Commits":[{"Sha1":"a63414ab085268e9570c6f51816a92718f2effeb","Message":"chore(reports): F-7 收口——框架迁移凭证 frameworkVersion 0.22.0\n\npnpm migration:record --mode sync：source=base-framework 0.22.0 @ df53867 →\ntarget=enterprise-platform 0.22.0，runId cb50e485。\n\n**一致性门禁 L12 清零**（内核 pin 落后六个 minor 的常驻问题）。常驻问题由\nL9 / L12 两条降为 L9 一条——后者是 8 个上层应用仓的 compose 身份，待 N-11\n裁决扩散路径，与本仓无关。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T22:39:42-07:00"}],"HeadCommit":{"Sha1":"a63414ab085268e9570c6f51816a92718f2effeb","Message":"chore(reports): F-7 收口——框架迁移凭证 frameworkVersion 0.22.0\n\npnpm migration:record --mode sync：source=base-framework 0.22.0 @ df53867 →\ntarget=enterprise-platform 0.22.0，runId cb50e485。\n\n**一致性门禁 L12 清零**（内核 pin 落后六个 minor 的常驻问题）。常驻问题由\nL9 / L12 两条降为 L9 一条——后者是 8 个上层应用仓的 compose 身份，待 N-11\n裁决扩散路径，与本仓无关。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T22:39:42-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/d24b762cd6740f0bce850ae3fffcc84e8a279c28...a63414ab085268e9570c6f51816a92718f2effeb","Len":1}...
|
1789709986
|
Edit
Delete
|
|
30823
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"d24b762cd {"Commits":[{"Sha1":"d24b762cd6740f0bce850ae3fffcc84e8a279c28","Message":"chore(reports): 框架同步 0.16.1 → 0.22.0 后全套验收回绑 @ 765300b\n\n三级验收全部在干净树重跑，与升级前逐项持平：\n\n- runtime-acceptance **775 / 775 passed**，租户 RLS enforce（地板 775 未降）\n- ui-acceptance 5 / 5，1 组合\n- mainline-acceptance 四套件 11 / 17 / 19 / 18\n- mainline-restore passed，port-conformance 41 / 41\n- revocation-sla partial（invalidationPush 仍是 harness 传输，生产未接线）\n- release-manifest / sbom / identity-ui 随后重跑\n\ncheck:evidence **passed：35 份全新鲜 / 0 过期 / 0 例外**。\n\n过程中一处自伤已修复：曾用 git checkout -- runtime/reports 还原「跑门禁弄脏\n的报告」，连带把真跑出的六份验收证据还原成旧版，已在干净树重跑全套找回。\n教训：还原前必须按 provenance.gitSha 分辨本轮产出与脏绑，不可整目录 checkout。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T22:38:56-07:00"},{"Sha1":"d66f63659bc07fa65ac84685335a0f3c5e15af5d","Message":"docs(runtime): 动态区 SHA 回灌至 765300b（重跑后的验收绑定）\n\n上一轮误用 git checkout -- runtime/reports 把真跑出的六份验收报告还原成了\n旧版，已在干净树重跑全套（runtime 775/775、UI 5/5、主线 11/17/19/18、恢复\n演练、端口一致性 41/41），六份统一绑 765300b / dirty=false。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T22:36:52-07:00"},{"Sha1":"765300b962ca658d112ac6064f925d82d0579dfd","Message":"docs(runtime): 动态区回灌框架同步后的三级验收 @ 56e78df\n\ncheck:docs-truth 再次拦下「报告已刷新而动态区未跟上」。\n\n- 真实 DB：775 tests / 0 failures clean 56e78df——框架同步 0.16.1 → 0.22.0 后\n 首跑，测试数与升级前持平、租户 RLS 仍 enforce\n- 浏览器：5 用例 / 1 组合 clean 56e78df\n- 环境如实记：隔离栈 enterprise-platform-ms23-ci（PG :55485 / Redis :56395 /\n Redpanda :59105）+ 验收库 enterprise_platform_acceptance_fwsync\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T22:29:47-07:00"},{"Sha1":"56e78dfa4fb1001fad92e6d93b0e3bb2f7569594","Message":"fix(runtime): 修正 CHANGELOG 并入六段后的跨仓断链（一致性门禁 L4）\n\n框架 CHANGELOG 0.17.0—0.22.0 六段里的 ../\u003c框架变更请求\u003e.md 指的是框架仓同级，\n并入本仓后全部失效，L4 报 5 条。改为 ../../../工程基础框架/\u003c文件\u003e 跨仓路径。\n\n这是框架同步的真实副作用：并入上游文档时，其相对链接的锚点也跟着换了仓。\nL4 / L10 / L13 恢复 CLEAN。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T21:38:33-07:00"},{"Sha1":"b30ddd94965db4c6686a249d280ce579a879fc65","Message":"feat(runtime): 框架同步 0.16.1 → 0.22.0 主体（F-2b / F-3 / F-4）\n\nC-14 的实际形态。本轮把版本轴与两后端行为接线一次做完，部署两级按目录负责人\n裁决走「改判据 + 提 FR-6」。\n\nF-3 版本轴\n- 12 处 @juhai/kernel pin、runtime/package.json.version、packages/contracts\n version 全部 0.16.1 → 0.22.0；**F11 转绿**（这是抬 pin 一直卡住的门禁）\n- CHANGELOG 并入框架 0.17.0—0.22.0 六段，版本序连续；COMPATIBILITY.md 取 0.22.0\n\nF-4 两后端接线（C34 / C35 / C39 绊网全绿）\n- 直取六个未改过的文件：health.controller / health.module / health-checks.registry\n / request-log.middleware / logger / routes/health，另 types.ts、四个 C39 文件\n- 两处真三方合并：\n · api-fastify/src/app.ts：取 0.22.0 后按 sample-domain 标记剥离三个示例域块\n （本仓以 --strip-sample 建仓，本仓对该文件的唯一定制就是这三处剥离）\n · api-nestjs/src/create-app.ts：两侧改动不重叠——本仓加平台模块装载、框架加\n 访问日志中间件，合并保留两者\n\nF-2b 门禁脚本\n- check-dual-backend-parity / check-docs-truth / check-runtime-acceptance 取 0.22.0\n\ndocs-truth 四级新鲜度的本地修正（已提 FR-6）\n框架 0.22.0 把新鲜度断言由两级扩到四级，新增的部署 / 生产两级在本仓无法成立：\n其验收链要求 apps/api-nestjs/Dockerfile，而那份里写着 COPY packages/kernel/\npackage.json——本仓**不 vendor 内核**（pin 已发布包，packages/kernel 不存在），\n实测构建必然失败。本仓自己的制品镜像是 D-1 的 runtime/Dockerfile，与框架那套\n并存会变成两条镜像路径。判据改为「只对本仓真正接线了的级别断言」，未接线的\n打印 SKIPPED 而非伪装通过——不接受占位报告，与框架 doctrine「诚实的空」一致。\n已撤回全部 F-5 部署物与失败留下的半截报告。\n\n治理测试锚点 0.16.1 → 0.22.0。\n\n验收：turbo build 17/17、typecheck 31/31、runtime check 全链通过（两级 SKIPPED\n如实打印）、governance 测试 277/277。**真实三级验收随后在 clean 树重跑并回绑**\n——check:evidence 已如实报出 20 份证据的输入被本轮改动，不得只回绑静态报告。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T21:23:29-07:00"}],"HeadCommit":{"Sha1":"d24b762cd6740f0bce850ae3fffcc84e8a279c28","Message":"chore(reports): 框架同步 0.16.1 → 0.22.0 后全套验收回绑 @ 765300b\n\n三级验收全部在干净树重跑，与升级前逐项持平：\n\n- runtime-acceptance **775 / 775 passed**，租户 RLS enforce（地板 775 未降）\n- ui-acceptance 5 / 5，1 组合\n- mainline-acceptance 四套件 11 / 17 / 19 / 18\n- mainline-restore passed，port-conformance 41 / 41\n- revocation-sla partial（invalidationPush 仍是 harness 传输，生产未接线）\n- release-manifest / sbom / identity-ui 随后重跑\n\ncheck:evidence **passed：35 份全新鲜 / 0 过期 / 0 例外**。\n\n过程中一处自伤已修复：曾用 git checkout -- runtime/reports 还原「跑门禁弄脏\n的报告」，连带把真跑出的六份验收证据还原成旧版，已在干净树重跑全套找回。\n教训：还原前必须按 provenance.gitSha 分辨本轮产出与脏绑，不可整目录 checkout。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T22:38:56-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/96bfdde508464b4b0e1e3fb05f77a9f488492fc0...d24b762cd6740f0bce850ae3fffcc84e8a279c28","Len":6}...
|
1789709941
|
Edit
Delete
|
|
30789
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"d70b1e5b6 {"Commits":[{"Sha1":"d70b1e5b6f3ed902a64adac86d0a4d953cadddb8","Message":"docs(计划): F-1 开工即证伪——「零判读快进」实收 0 个文件，本批取消\n\n按立项书从 F-1 起做，逐项审定 75 个候选（55 safe-copy + 20 个本仓从未改过的\n冲突文件）后实收 0 个，未改动仓内任何文件。\n\n拒收 44：\n- 30 个 packages/kernel/*：本仓不 vendor 内核源码，pin 的是已发布的\n @juhai/kernel；而 packages/* 在本仓 workspace 内，收进来会与 pin 直接冲突\n- 14 个示例 Order 域（orders.* / order.ts / orderMachine.ts / OrdersPanel.tsx）：\n 本仓以 --strip-sample 建仓，实测 0 个同名文件，收回即让剥离回潮；仓内\n docs/strip-sample-domain.md 正是该机制的说明\n\n其余 31 个归位：F-2 门禁脚本 11、F-4 的 C34/C35/C39 接线面 12、F-5 部署物 4、\nF-3 COMPATIBILITY 1、F-6 文档 1，另 2 个保护路径延后判读。\n\n核心教训：migration:plan 的「safe-copy」标签指的是「目标没有这个文件」，不是\n「拿过来安全」。对一个 --strip-sample 且不 vendor 内核的派生仓，「目标没有」\n恰恰常常意味着「本仓刻意不要」。照单全收会同时干两件坏事：把剥离掉的示例域\n装回来、把 pin 的内核变成 vendor 的源码。\n\n立项书 §5.1 记下证伪与批次修正：不存在零判读快进层；余下六批内容不变，但每批\n开工须先逐文件审定，不得直接采信 migration:plan 的分类；F-2 成为实际起点。\n\n顺带修 L13：新增立项书使 平台治理/README.md 收纳表计数 157 → 158。\n全量一致性门禁回到 L9 / L12 两条常驻基线；测试 35/35。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T21:01:00-07:00"}],"HeadCommit":{"Sha1":"d70b1e5b6f3ed902a64adac86d0a4d953cadddb8","Message":"docs(计划): F-1 开工即证伪——「零判读快进」实收 0 个文件，本批取消\n\n按立项书从 F-1 起做，逐项审定 75 个候选（55 safe-copy + 20 个本仓从未改过的\n冲突文件）后实收 0 个，未改动仓内任何文件。\n\n拒收 44：\n- 30 个 packages/kernel/*：本仓不 vendor 内核源码，pin 的是已发布的\n @juhai/kernel；而 packages/* 在本仓 workspace 内，收进来会与 pin 直接冲突\n- 14 个示例 Order 域（orders.* / order.ts / orderMachine.ts / OrdersPanel.tsx）：\n 本仓以 --strip-sample 建仓，实测 0 个同名文件，收回即让剥离回潮；仓内\n docs/strip-sample-domain.md 正是该机制的说明\n\n其余 31 个归位：F-2 门禁脚本 11、F-4 的 C34/C35/C39 接线面 12、F-5 部署物 4、\nF-3 COMPATIBILITY 1、F-6 文档 1，另 2 个保护路径延后判读。\n\n核心教训：migration:plan 的「safe-copy」标签指的是「目标没有这个文件」，不是\n「拿过来安全」。对一个 --strip-sample 且不 vendor 内核的派生仓，「目标没有」\n恰恰常常意味着「本仓刻意不要」。照单全收会同时干两件坏事：把剥离掉的示例域\n装回来、把 pin 的内核变成 vendor 的源码。\n\n立项书 §5.1 记下证伪与批次修正：不存在零判读快进层；余下六批内容不变，但每批\n开工须先逐文件审定，不得直接采信 migration:plan 的分类；F-2 成为实际起点。\n\n顺带修 L13：新增立项书使 平台治理/README.md 收纳表计数 157 → 158。\n全量一致性门禁回到 L9 / L12 两条常驻基线；测试 35/35。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T21:01:00-07:00"},"CompareURL":"luoanwu/platform-governance/compare/3cbdf7409759ef6652fe7811dee6e30fb84baaa5...d70b1e5b6f3ed902a64adac86d0a4d953cadddb8","Len":1}...
|
1789704064
|
Edit
Delete
|
|
30788
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"3cbdf7409 {"Commits":[{"Sha1":"3cbdf7409759ef6652fe7811dee6e30fb84baaa5","Message":"docs(计划): 框架同步 0.16.1 → 0.22.0 立项，分 F-1—F-7 七批\n\n目录负责人指令：偏差 #35 选③「框架同步单独立项」。新增立项书\ndocs/框架同步立项.md，文件名刻意不带日期——否则会被一致性门禁的\nHISTORY_NAME_PATTERN 当历史快照而跳过 L3/L4/L10 扫描（教训出自\n开发计划梳理-2026-09-14 的同名坑）。\n\n立项前逐文件核实，三条比首轮更准的事实：\n\n1. 74 个冲突里 20 个本仓从未改过（仍逐字节等于框架 v0.16.1），可直接快进；\n 真需三方合并的是 53 个，另 1 个是两侧各自新增的 .dockerignore\n2. apps/api-nestjs 的 17 个真冲突里只有 1 个在 src/platform/——首轮按目录\n 粗计说「22 个正是平台装配点」不成立。平台自有的七模块 / clients /\n platform-tests 共 531 文件全在 target-only，框架同步根本不碰\n3. 两个前置都比裁决时乐观：C-13 的卷那一半已执行；F14 在本仓基本已合规\n （runtime 的 compose 已是 ${COMPOSE_PROJECT_NAME:-…} 形式，只剩\n stack/compose.yaml 改变量形式与 act-runner 补 name:）\n\n内核 API 不是风险源：六个 minor 只有 364 增 / 3 删，纯新增八个导出，唯一\n语义变更是 tenantIdFromClaims 改读序——本仓 FR-4 被受理，对 DEC-032 是白赚。\n风险全在框架源码的行为接线（C34 / C35 / C39 / F14）。\n\n七批各带机器退出条件，F-4 两后端接线标为高风险（唯一会动 platform-ports.module.ts\n的批次）。立项书显式禁止「只改 pin + 版本 + CHANGELOG 而不取框架源码」——\nF11 会过，但会伪造迁移凭证，与本仓当日清理的病同源。\n\n§3A 增领取行，§6 偏差 #35 结论改为已选③并立项，README 增索引。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T20:57:38-07:00"}],"HeadCommit":{"Sha1":"3cbdf7409759ef6652fe7811dee6e30fb84baaa5","Message":"docs(计划): 框架同步 0.16.1 → 0.22.0 立项，分 F-1—F-7 七批\n\n目录负责人指令：偏差 #35 选③「框架同步单独立项」。新增立项书\ndocs/框架同步立项.md，文件名刻意不带日期——否则会被一致性门禁的\nHISTORY_NAME_PATTERN 当历史快照而跳过 L3/L4/L10 扫描（教训出自\n开发计划梳理-2026-09-14 的同名坑）。\n\n立项前逐文件核实，三条比首轮更准的事实：\n\n1. 74 个冲突里 20 个本仓从未改过（仍逐字节等于框架 v0.16.1），可直接快进；\n 真需三方合并的是 53 个，另 1 个是两侧各自新增的 .dockerignore\n2. apps/api-nestjs 的 17 个真冲突里只有 1 个在 src/platform/——首轮按目录\n 粗计说「22 个正是平台装配点」不成立。平台自有的七模块 / clients /\n platform-tests 共 531 文件全在 target-only，框架同步根本不碰\n3. 两个前置都比裁决时乐观：C-13 的卷那一半已执行；F14 在本仓基本已合规\n （runtime 的 compose 已是 ${COMPOSE_PROJECT_NAME:-…} 形式，只剩\n stack/compose.yaml 改变量形式与 act-runner 补 name:）\n\n内核 API 不是风险源：六个 minor 只有 364 增 / 3 删，纯新增八个导出，唯一\n语义变更是 tenantIdFromClaims 改读序——本仓 FR-4 被受理，对 DEC-032 是白赚。\n风险全在框架源码的行为接线（C34 / C35 / C39 / F14）。\n\n七批各带机器退出条件，F-4 两后端接线标为高风险（唯一会动 platform-ports.module.ts\n的批次）。立项书显式禁止「只改 pin + 版本 + CHANGELOG 而不取框架源码」——\nF11 会过，但会伪造迁移凭证，与本仓当日清理的病同源。\n\n§3A 增领取行，§6 偏差 #35 结论改为已选③并立项，README 增索引。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T20:57:38-07:00"},"CompareURL":"luoanwu/platform-governance/compare/bc014de63a169082ff81ff54e77012c5fda7e50f...3cbdf7409759ef6652fe7811dee6e30fb84baaa5","Len":1}...
|
1789703862
|
Edit
Delete
|
|
30787
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"bc014de63 {"Commits":[{"Sha1":"bc014de63a169082ff81ff54e77012c5fda7e50f","Message":"docs(计划): 登记偏差 #35——C-14「抬内核 pin」的实际范围是框架同步 74 冲突\n\n按目录负责人指令动工后实测：裁决文本说的是抬 pin，机制成本是一次完整框架\n同步。已完整撤回，未提交任何 pin 改动、未做框架合并。\n\n做到哪一步：12 处 pin 全改 0.22.0，pnpm install 成功、锁文件更新、\ncheck:pins 通过、构建 17/17、类型检查 31/31 全过。\n\n卡在哪：F11（check:fork-readiness）要求 @juhai/kernel pin 必须等于本仓\npackage.json.version——框架版本轴 = 内核版本轴；仓版本又被 check:docs-truth\n的 changelog-version-matches-package 绑到 runtime/CHANGELOG.md。\n\n真实范围（migration:plan，以框架仓为源，只读）：冲突 74（apps/api-nestjs 22、\napps/api-fastify 14、apps/web 8，另含 check-dual-backend-parity.mjs、\ncheck-docs-truth.mjs、CLAUDE.md、package.json、pnpm-lock.yaml）、safe-copy 55、\nmerge-sensitive 12。六版累加含 F14 compose 身份、C34 健康经内核聚合、C35 统一\n访问日志与 req 序列化变更、新增 Dockerfile 与生产编排、三个门禁脚本迁入。\napi-nestjs 那 22 个正是平台装配点（端口合成 / HTTP 适配 / 运维命令 API）。\n\n反向的好消息：内核包本身六个 minor 只有 364 增 / 3 删，纯新增八个导出，且\ntenantIdFromClaims 改为 tenant_id → tenantId → tid 读序——本仓 FR-4 已被框架\n受理，对 DEC-032 canonical 租户是白赚。风险全在框架源码同步，不在内核 API。\n\n给三个处置选项待裁，并写明不可接受的第四条：只改 pin + 仓版本 + 补 CHANGELOG\n而不取框架源码——F11 会过，但仓就在声称自己是 0.22.0 却没有 0.22.0 的门禁与\n中间件，迁移凭证会记下一个假事实。\n\n门禁消费登记同步：L12 的接入条件应读作「待框架同步立项并完成」，不是「待抬\n一次 pin」。check:gates 通过，相关测试 46/46。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T20:54:37-07:00"}],"HeadCommit":{"Sha1":"bc014de63a169082ff81ff54e77012c5fda7e50f","Message":"docs(计划): 登记偏差 #35——C-14「抬内核 pin」的实际范围是框架同步 74 冲突\n\n按目录负责人指令动工后实测：裁决文本说的是抬 pin，机制成本是一次完整框架\n同步。已完整撤回，未提交任何 pin 改动、未做框架合并。\n\n做到哪一步：12 处 pin 全改 0.22.0，pnpm install 成功、锁文件更新、\ncheck:pins 通过、构建 17/17、类型检查 31/31 全过。\n\n卡在哪：F11（check:fork-readiness）要求 @juhai/kernel pin 必须等于本仓\npackage.json.version——框架版本轴 = 内核版本轴；仓版本又被 check:docs-truth\n的 changelog-version-matches-package 绑到 runtime/CHANGELOG.md。\n\n真实范围（migration:plan，以框架仓为源，只读）：冲突 74（apps/api-nestjs 22、\napps/api-fastify 14、apps/web 8，另含 check-dual-backend-parity.mjs、\ncheck-docs-truth.mjs、CLAUDE.md、package.json、pnpm-lock.yaml）、safe-copy 55、\nmerge-sensitive 12。六版累加含 F14 compose 身份、C34 健康经内核聚合、C35 统一\n访问日志与 req 序列化变更、新增 Dockerfile 与生产编排、三个门禁脚本迁入。\napi-nestjs 那 22 个正是平台装配点（端口合成 / HTTP 适配 / 运维命令 API）。\n\n反向的好消息：内核包本身六个 minor 只有 364 增 / 3 删，纯新增八个导出，且\ntenantIdFromClaims 改为 tenant_id → tenantId → tid 读序——本仓 FR-4 已被框架\n受理，对 DEC-032 canonical 租户是白赚。风险全在框架源码同步，不在内核 API。\n\n给三个处置选项待裁，并写明不可接受的第四条：只改 pin + 仓版本 + 补 CHANGELOG\n而不取框架源码——F11 会过，但仓就在声称自己是 0.22.0 却没有 0.22.0 的门禁与\n中间件，迁移凭证会记下一个假事实。\n\n门禁消费登记同步：L12 的接入条件应读作「待框架同步立项并完成」，不是「待抬\n一次 pin」。check:gates 通过，相关测试 46/46。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T20:54:37-07:00"},"CompareURL":"luoanwu/platform-governance/compare/1b940277848578801e16f969b874470ff90adce7...bc014de63a169082ff81ff54e77012c5fda7e50f","Len":1}...
|
1789703681
|
Edit
Delete
|
|
30776
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"1b9402778 {"Commits":[{"Sha1":"1b940277848578801e16f969b874470ff90adce7","Message":"docs(治理): C-13 卷归属执行记录——备份 + 删卷已做，compose 半边未做\n\n决策批次 C-13 的附条件是「删卷前 dump 必须复验」，按字面执行：\n\n备份（/Users/hillao/AI产品/仓备份-2026-09-16/碰撞卷-C13/，24 MB）\n- 五个碰撞卷文件级 tar + 两个 PG 簇的 pg_dumpall，附 SHA256SUMS 与 README\n- 原卷全程 :ro，快照解到临时副本卷后在副本上起 PG 导出——在原卷上起 PG 会\n 触发 WAL 与恢复写入，而删不删当时尚未决定，不能先把它改了\n\n复验（不是「文件存在」）\n- 逻辑备份灌入全新空实例：4 库表数逐库一致，74 张表逐表 count(*) 零差异\n\n删除（核实确属历史遗留后）\n- 六个容器的工作目录全部指向迁移前旧布局，三条路径均已不存在，创建于\n 08-28—09-05；顺带清除一个漏网的同工程容器 api-nestjs-minio-init-1\n- 五卷 + 六容器清零，dev 常驻栈 13 容器未动、健康探针 200\n\n实测更正：api-nestjs_pgdata 占 47.8 MB 却是空簇（唯一库 0 张表），有数据的\n只有 enterprise-idp_pgdata（4 库 / 74 表 / ≈2088 行）。\n\n未做的半边：8 个上层应用仓的 compose 补顶层 name:，跨仓且待 N-11 裁决。\n因此 L9 仍报 1 条（它查 compose 声明，不是卷是否存在），C-14 抬内核 pin 的\n前置只部分满足——这一条同时写进了门禁消费登记，免得下次误读为「C-13 已完成」。\n\ncheck:gates 通过；gate-consumption 与 workspace-consistency 测试 46/46。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T20:13:04-07:00"}],"HeadCommit":{"Sha1":"1b940277848578801e16f969b874470ff90adce7","Message":"docs(治理): C-13 卷归属执行记录——备份 + 删卷已做，compose 半边未做\n\n决策批次 C-13 的附条件是「删卷前 dump 必须复验」，按字面执行：\n\n备份（/Users/hillao/AI产品/仓备份-2026-09-16/碰撞卷-C13/，24 MB）\n- 五个碰撞卷文件级 tar + 两个 PG 簇的 pg_dumpall，附 SHA256SUMS 与 README\n- 原卷全程 :ro，快照解到临时副本卷后在副本上起 PG 导出——在原卷上起 PG 会\n 触发 WAL 与恢复写入，而删不删当时尚未决定，不能先把它改了\n\n复验（不是「文件存在」）\n- 逻辑备份灌入全新空实例：4 库表数逐库一致，74 张表逐表 count(*) 零差异\n\n删除（核实确属历史遗留后）\n- 六个容器的工作目录全部指向迁移前旧布局，三条路径均已不存在，创建于\n 08-28—09-05；顺带清除一个漏网的同工程容器 api-nestjs-minio-init-1\n- 五卷 + 六容器清零，dev 常驻栈 13 容器未动、健康探针 200\n\n实测更正：api-nestjs_pgdata 占 47.8 MB 却是空簇（唯一库 0 张表），有数据的\n只有 enterprise-idp_pgdata（4 库 / 74 表 / ≈2088 行）。\n\n未做的半边：8 个上层应用仓的 compose 补顶层 name:，跨仓且待 N-11 裁决。\n因此 L9 仍报 1 条（它查 compose 声明，不是卷是否存在），C-14 抬内核 pin 的\n前置只部分满足——这一条同时写进了门禁消费登记，免得下次误读为「C-13 已完成」。\n\ncheck:gates 通过；gate-consumption 与 workspace-consistency 测试 46/46。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T20:13:04-07:00"},"CompareURL":"luoanwu/platform-governance/compare/ae6cf47e538de724c00c270f78ad014669585c8d...1b940277848578801e16f969b874470ff90adce7","Len":1}...
|
1789701193
|
Edit
Delete
|
|
30751
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"ae6cf47e5 {"Commits":[{"Sha1":"ae6cf47e538de724c00c270f78ad014669585c8d","Message":"docs(计划): mainline-dev-deployment 退役，四份手工记录清零（真源仓 903f019）\n\n结论由 compose 的 redpanda 启用状态、四份更新的 runtime-up 记录与\ncheck:deployed 覆盖，三者都会自动保鲜；dev 栈此后已重建多轮，09-12 的\n快照早已不是当前事实。\n\n§1 证据行补第三步；§8 增一行。未登记 16 → 13。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:24:42-07:00"}],"HeadCommit":{"Sha1":"ae6cf47e538de724c00c270f78ad014669585c8d","Message":"docs(计划): mainline-dev-deployment 退役，四份手工记录清零（真源仓 903f019）\n\n结论由 compose 的 redpanda 启用状态、四份更新的 runtime-up 记录与\ncheck:deployed 覆盖，三者都会自动保鲜；dev 栈此后已重建多轮，09-12 的\n快照早已不是当前事实。\n\n§1 证据行补第三步；§8 增一行。未登记 16 → 13。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:24:42-07:00"},"CompareURL":"luoanwu/platform-governance/compare/e81617fb827a3db190701bc33bca21cc14ce20c8...ae6cf47e538de724c00c270f78ad014669585c8d","Len":1}...
|
1789698286
|
Edit
Delete
|
|
30747
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"96bfdde50 {"Commits":[{"Sha1":"96bfdde508464b4b0e1e3fb05f77a9f488492fc0","Message":"chore(reports): 回绑 @ 903f019\n\ncheck:evidence passed：35 份已登记全部新鲜，0 过期 / 0 例外 / 0 告警；\n未登记 13 份（四份手工记录退役后的稳定态）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:23:58-07:00"},{"Sha1":"903f019dc8033d3042cc0446e054e2665782fedc","Message":"chore(governance): 退役 mainline-dev-deployment，四份手工记录清零\n\n目录负责人裁决：与今日退役的另三份同类处置。\n\n它只记 09-12 在 dev 栈启用 Redpanda 这一件事（自述 scope「Redpanda only」），\n仓内无生成脚本、不可重跑。该结论现由三处覆盖且都会自动保鲜：\n\n- stack/compose.yaml 的 redpanda 已是启用状态（非候选注释块），MS-2 的机器\n 判定直接读它\n- stack/reports/runtime-up.*.json：四份更新的 dev 运行记录（09-14 / 09-16 ×3）\n- pnpm check:deployed：活体核验运行镜像与源提交\n\ndev 栈此后已重建多轮，09-12 那份快照描述的容器与端点早已不是当前事实。\n\nevidence-scopes note 补记本次裁决（它本就未登记，无条目可撤）；\n09-12 的执行记录里那处纯文本引用补一句退役说明——L4 不报（不是 markdown\n链接），但读者会去找一个不存在的文件。\n\n至此四份无生成脚本的手工记录全部退役：identity-http、\nmainline-candidate-regression、stack-image-pins、mainline-dev-deployment。\n\ncheck:evidence passed：35 份已登记全部新鲜，未登记 14 → 13。\ngovernance 测试 277/277。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:23:31-07:00"}],"HeadCommit":{"Sha1":"96bfdde508464b4b0e1e3fb05f77a9f488492fc0","Message":"chore(reports): 回绑 @ 903f019\n\ncheck:evidence passed：35 份已登记全部新鲜，0 过期 / 0 例外 / 0 告警；\n未登记 13 份（四份手工记录退役后的稳定态）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:23:58-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/314c179366fb02a53485f4efbd0e3d3e88cd5470...96bfdde508464b4b0e1e3fb05f77a9f488492fc0","Len":2}...
|
1789698242
|
Edit
Delete
|
|
30746
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"e81617fb8 {"Commits":[{"Sha1":"e81617fb827a3db190701bc33bca21cc14ce20c8","Message":"docs(计划): MS-4 判据输入纳入新鲜度监控；Manifest 缺项 4 → 6（真源仓 314c179）\n\nrelease-manifest / sbom / image-smoke / mainline-dev-deployment 此前被归为\n「没有 provenance.gitSha」，实为写了源 SHA 但平铺在顶层，check:evidence 只认\n嵌套形状。实锤危害：Manifest 绑 ac3683d 四天，其后结论输入改了 10 个文件，\n刷新后缺项 4 → 6，多出的 images.stack.* 未锁与 SBOM 不满足 D-2 两条都是真的\n——而 MS-4 的退出条件正是读这份 Manifest。\n\n§1 Release Manifest 行按刷新后的六项缺项重采；§1 证据行改为 35 份全新鲜；\n§8 增一行，含未做项与理由（image-smoke 待下次真实运行、mainline-dev-deployment\n无生成脚本属手工记录）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:21:54-07:00"}],"HeadCommit":{"Sha1":"e81617fb827a3db190701bc33bca21cc14ce20c8","Message":"docs(计划): MS-4 判据输入纳入新鲜度监控；Manifest 缺项 4 → 6（真源仓 314c179）\n\nrelease-manifest / sbom / image-smoke / mainline-dev-deployment 此前被归为\n「没有 provenance.gitSha」，实为写了源 SHA 但平铺在顶层，check:evidence 只认\n嵌套形状。实锤危害：Manifest 绑 ac3683d 四天，其后结论输入改了 10 个文件，\n刷新后缺项 4 → 6，多出的 images.stack.* 未锁与 SBOM 不满足 D-2 两条都是真的\n——而 MS-4 的退出条件正是读这份 Manifest。\n\n§1 Release Manifest 行按刷新后的六项缺项重采；§1 证据行改为 35 份全新鲜；\n§8 增一行，含未做项与理由（image-smoke 待下次真实运行、mainline-dev-deployment\n无生成脚本属手工记录）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:21:54-07:00"},"CompareURL":"luoanwu/platform-governance/compare/a03bc2d7789a81e0da6b79e51f2abb32fff2b906...e81617fb827a3db190701bc33bca21cc14ce20c8","Len":1}...
|
1789698116
|
Edit
Delete
|
|
30745
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"314c17936 {"Commits":[{"Sha1":"314c179366fb02a53485f4efbd0e3d3e88cd5470","Message":"chore(reports): 回绑 @ 71b306f，登记面 33 → 35 全新鲜\n\nrelease-manifest 与 sbom 首次带嵌套 provenance 并受新鲜度保护。\n\nManifest 刷新后缺项 4 → 6，多出的两条都是真的、被旧快照藏了四天：\n- images.stack.*：compose 启用 runtime 服务后镜像是变量 ${PLATFORM_RUNTIME_IMAGE:?}，\n 未按 digest 锁定（PF-05 / PF-23 晋级前解决）\n- sbom：局部 lockfile 清单（268 包，仅 Node 依赖，不含基础镜像 Debian 包）\n 不满足 D-2 的完整 SBOM 要求\n\ncheck:evidence passed：35 份已登记全部新鲜，0 过期 / 0 例外 / 0 告警。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:21:00-07:00"},{"Sha1":"71b306f53390e6119590b5225b660cd4effcd3b5","Message":"chore(governance): 登记 release-manifest 与 sbom 进 evidence-scopes（35 条）\n\n承 c5d9897。三份补了嵌套 provenance 之后，其中两份现在就能登记：\n\n- release-manifest（warn）：作用域 = compose、image-digest.json、四类包的\n package.json、runtime/modules、生成器与 schema。**这是 MS-4 退出条件的判据\n 输入**，此前无人监控其新鲜度\n- sbom（warn）：作用域 = image-digest.json、生成器、runtime 锁文件\n\nimage-smoke 本轮不登记：生成器已修，但要补上形状得起真实容器重跑（脚本强制\n--env-file，需 DB/Redis），不值得为补一个形状去起一套底座——待其下一次真实\n运行后登记。note 里写明了这一条与 mainline-dev-deployment 的区别（后者没有\n生成脚本，属手工记录那一类，不是形状问题）。\n\n同时修正 note 里 ① 类的描述：原写「没有 provenance.gitSha，要先让产生脚本写\nprovenance」，对这四份是不准确的——数据早就有。\n\ngovernance 测试 277/277。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:20:23-07:00"},{"Sha1":"c5d98975f8c466d035550a4f38899db325d8b806","Message":"feat(governance): 三份报告补嵌套 provenance，让 MS-4 判据输入受新鲜度保护\n\n2026-09-17 目录负责人裁决：补嵌套 provenance，四份有 SHA 的全做。\n\n问题：release-manifest / sbom / image-smoke / mainline-dev-deployment 都写了\n源 SHA，但是**平铺**在顶层（sourceSha），而 check:evidence 只认嵌套的\nprovenance.gitSha。于是它们登记不进 evidence-scopes，游离在新鲜度之外。\n此前 evidence-scopes 的口径说明把它们归为「没有 provenance.gitSha，要先让\n产生脚本写 provenance」——对其中四份是不准确的：数据早就有，缺的是形状。\n\n实锤这不是理论风险：Manifest 绑 ac3683d（2026-09-14），其后它的结论输入改了\n10 个文件（contracts/governance 的 package.json、image-digest.json、\ncompose.yaml、client-fact 全套），四天无人报警。重跑后缺项 4 → 6，多出的两条\n都是真的——compose 启用 runtime 服务后镜像是变量未按 digest 锁定；SBOM 未随\n19150ae 镜像重生成。**一份过期的 Manifest 一直在少报，而 MS-4 的退出条件正是\n读它**（status=complete 且 runner 为 CI）。\n\n改动：\n- stack/release-manifest.schema.json 增可选 provenance 对象（不进 required，\n 向后兼容；additionalProperties:false 故必须显式加），并写明与顶层三项同源\n- release-manifest.mjs 同时写 provenance 与顶层兼容字段，实测两者一致\n- sbom.mjs / smoke-image.mjs 增 provenance: gitProvenance(root)。注意它们原有的\n sourceSha 是**镜像的**源提交，与「本报告在哪个提交产出」不是一回事，两者都留\n\nmainline-dev-deployment 做不了：全仓搜索确认它没有生成脚本，是第四份手工记录\n（与今日退役的三份同类），不是形状问题。另行处置。\n\ngovernance 测试 277/277。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:19:03-07:00"}],"HeadCommit":{"Sha1":"314c179366fb02a53485f4efbd0e3d3e88cd5470","Message":"chore(reports): 回绑 @ 71b306f，登记面 33 → 35 全新鲜\n\nrelease-manifest 与 sbom 首次带嵌套 provenance 并受新鲜度保护。\n\nManifest 刷新后缺项 4 → 6，多出的两条都是真的、被旧快照藏了四天：\n- images.stack.*：compose 启用 runtime 服务后镜像是变量 ${PLATFORM_RUNTIME_IMAGE:?}，\n 未按 digest 锁定（PF-05 / PF-23 晋级前解决）\n- sbom：局部 lockfile 清单（268 包，仅 Node 依赖，不含基础镜像 Debian 包）\n 不满足 D-2 的完整 SBOM 要求\n\ncheck:evidence passed：35 份已登记全部新鲜，0 过期 / 0 例外 / 0 告警。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:21:00-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/2dfbcb9148a7beabb926462eebb82645e548c10a...314c179366fb02a53485f4efbd0e3d3e88cd5470","Len":3}...
|
1789698065
|
Edit
Delete
|
|
30741
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"a03bc2d77 {"Commits":[{"Sha1":"a03bc2d7789a81e0da6b79e51f2abb32fff2b906","Message":"docs(计划): stack-image-pins 退役，check:evidence 首次全绿（真源仓 a9fd6e8）\n\n第三份无生成脚本的手工记录按其例外自写的「倾向 ① 退役」执行。结论已由\nrelease-manifest 的 parseComposeImages 覆盖（同为 12 条 0 未锁，未锁进\nmissing，由真脚本产出自动保鲜）。evidence-scopes exceptions 自此为空。\n\n§1 证据行改为 passed（33 新鲜 / 0 过期 / 0 例外 / 0 告警）；§8 增一行。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:01:40-07:00"}],"HeadCommit":{"Sha1":"a03bc2d7789a81e0da6b79e51f2abb32fff2b906","Message":"docs(计划): stack-image-pins 退役，check:evidence 首次全绿（真源仓 a9fd6e8）\n\n第三份无生成脚本的手工记录按其例外自写的「倾向 ① 退役」执行。结论已由\nrelease-manifest 的 parseComposeImages 覆盖（同为 12 条 0 未锁，未锁进\nmissing，由真脚本产出自动保鲜）。evidence-scopes exceptions 自此为空。\n\n§1 证据行改为 passed（33 新鲜 / 0 过期 / 0 例外 / 0 告警）；§8 增一行。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:01:40-07:00"},"CompareURL":"luoanwu/platform-governance/compare/150e625ed2a33300a3bc051fd4e60fa6e76a7947...a03bc2d7789a81e0da6b79e51f2abb32fff2b906","Len":1}...
|
1789696904
|
Edit
Delete
|
|
30740
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"2dfbcb914 {"Commits":[{"Sha1":"2dfbcb9148a7beabb926462eebb82645e548c10a","Message":"chore(reports): 回绑 @ a9fd6e8，check:evidence 全绿\n\n33 份已登记全部新鲜，0 过期 / 0 未提交输入 / 0 例外 / 0 告警。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:00:52-07:00"},{"Sha1":"a9fd6e826db77532ac3abfc149a9de822e40ea12","Message":"chore(governance): 退役 stack-image-pins，check:evidence 首次转 passed\n\n目录负责人裁决：按该报告自己的例外里写的「倾向 ① 退役」执行。\n\n它与今日退役的另两份同类：仓内无生成脚本（2026-09-14 全仓搜索确认），是\n一次性手工产物，永远无法重跑，只能靠例外掩盖。其结论「compose 12 个镜像\n是否全按 digest 锁定」已由 governance/release-manifest.mjs 的\nparseComposeImages 完整覆盖——实测同为 12 条、0 未锁，且未锁时进 missing，\n由真脚本产出、自动保鲜，严格优于被退役的手工快照。\n\n- evidence-scopes：entries 34 → 33，exceptions 1 → 0（自此为空）\n- check-evidence-freshness 的成因注释保留 2026-09-14 那条实锤（它正是本门禁\n 存在的理由），但标注该报告已退役，免得后人去找一个不存在的文件\n\ncheck:evidence：**passed**，33 新鲜 / 0 过期 / 0 输入未提交 / 0 例外。\n这是它第一次不是 partial。governance 测试 277/277。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:00:17-07:00"}],"HeadCommit":{"Sha1":"2dfbcb9148a7beabb926462eebb82645e548c10a","Message":"chore(reports): 回绑 @ a9fd6e8，check:evidence 全绿\n\n33 份已登记全部新鲜，0 过期 / 0 未提交输入 / 0 例外 / 0 告警。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T19:00:52-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/e6d286b4f61d0227fa7f42cd0f5f40e624540dbf...2dfbcb9148a7beabb926462eebb82645e548c10a","Len":2}...
|
1789696855
|
Edit
Delete
|
|
30731
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"150e625ed {"Commits":[{"Sha1":"150e625ed2a33300a3bc051fd4e60fa6e76a7947","Message":"docs(计划): 真实环境八份证据重跑通过，证据零过期（真源仓 36e0423）\n\n§1 证据新鲜度行与运行/UI 证据行按 2026-09-18 实测重采；§8 增一行。\n\n重跑：runtime-acceptance 775（地板 652 → 775）、ui-acceptance 5、\nmainline-acceptance 四套件 11/17/19/18、mainline-restore 3 库逐库 SHA-256\n一致、revocation-sla、identity-ui 2。退役（目录负责人裁决）：identity-http\n与 mainline-candidate-regression，两者仓内无生成脚本、永远无法重跑，结论\n已由自动化门禁覆盖。\n\ncheck:evidence：33 新鲜 / 0 过期 / 0 告警 / 1 例外。\n\n如实保留：invalidationPush 仍是 measured-with-harness-transport，生产传输\n未接线，不构成 MS-3 达成证据。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:58:36-07:00"}],"HeadCommit":{"Sha1":"150e625ed2a33300a3bc051fd4e60fa6e76a7947","Message":"docs(计划): 真实环境八份证据重跑通过，证据零过期（真源仓 36e0423）\n\n§1 证据新鲜度行与运行/UI 证据行按 2026-09-18 实测重采；§8 增一行。\n\n重跑：runtime-acceptance 775（地板 652 → 775）、ui-acceptance 5、\nmainline-acceptance 四套件 11/17/19/18、mainline-restore 3 库逐库 SHA-256\n一致、revocation-sla、identity-ui 2。退役（目录负责人裁决）：identity-http\n与 mainline-candidate-regression，两者仓内无生成脚本、永远无法重跑，结论\n已由自动化门禁覆盖。\n\ncheck:evidence：33 新鲜 / 0 过期 / 0 告警 / 1 例外。\n\n如实保留：invalidationPush 仍是 measured-with-harness-transport，生产传输\n未接线，不构成 MS-3 达成证据。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:58:36-07:00"},"CompareURL":"luoanwu/platform-governance/compare/689e924f3db984f90ee13c2e0dd1a0082a70c1d3...150e625ed2a33300a3bc051fd4e60fa6e76a7947","Len":1}...
|
1789696719
|
Edit
Delete
|
|
30730
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"e6d286b4f {"Commits":[{"Sha1":"e6d286b4f61d0227fa7f42cd0f5f40e624540dbf","Message":"docs: 修复退役报告导致的断链（一致性门禁 L4）\n\n09-12 的候选校验记录链向 reports/mainline-candidate-regression.latest.json，\n该报告已随 23281b8 退役，治理层一致性门禁 L4 随即报断链。改为文字说明：\n报告已退役、退役理由、结论现由 governance 测试产出（今 277/277），当时的\n154/0/1 保留为历史记录，不改写历史结论。\n\nL4 / L10 / L13 恢复 CLEAN。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:58:04-07:00"}],"HeadCommit":{"Sha1":"e6d286b4f61d0227fa7f42cd0f5f40e624540dbf","Message":"docs: 修复退役报告导致的断链（一致性门禁 L4）\n\n09-12 的候选校验记录链向 reports/mainline-candidate-regression.latest.json，\n该报告已随 23281b8 退役，治理层一致性门禁 L4 随即报断链。改为文字说明：\n报告已退役、退役理由、结论现由 governance 测试产出（今 277/277），当时的\n154/0/1 保留为历史记录，不改写历史结论。\n\nL4 / L10 / L13 恢复 CLEAN。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:58:04-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/36e0423efe0ddc105136d193bfb594efb9838d62...e6d286b4f61d0227fa7f42cd0f5f40e624540dbf","Len":1}...
|
1789696688
|
Edit
Delete
|
|
30729
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"36e0423ef {"Commits":[{"Sha1":"36e0423efe0ddc105136d193bfb594efb9838d62","Message":"chore(reports): 全链回绑 @ 1af0f22，证据零过期零告警\n\n干净树重跑根 pnpm check。check:evidence：34 份已登记，**33 份新鲜 / 0 过期 /\n0 输入未提交 / 0 告警 / 1 例外**（stack-image-pins，其例外自带出口）。\n\n这是本轮治理完善的收口：从九份过期、error 级判定，到零过期零告警。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:56:42-07:00"},{"Sha1":"1af0f2210a8ac041a26a97ff5f4965461398ebf1","Message":"feat(governance): 测试地板 652 → 775；动态区回灌今日三级验收\n\ncheck:docs-truth 抓出报告已刷新而动态区未跟上——正是今天修过的那类双真源\n漂移，这次由门禁自己拦住。\n\n- 动态区「真实 DB」：2026-09-13 / 7e26a50 / 652 → 2026-09-18 / 6108357 /\n 775，并写明隔离栈与「破坏性注入只认 ms23-ci 类专用工程」的边界\n- 动态区「浏览器」：2026-09-13 / 7e26a50 → 2026-09-18 / 003aadd，端口如实\n- 棘轮：testsPassed 652 → 775（本轮实测），ownTests 35 → 40、\n ownTestCases 185 → 212（静态 runner 自己抬的，它刻意不动其他 runner 的地板）\n\n棘轮只收紧不放宽：775 是今天真实跑出来的数，此后低于它即红。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:56:07-07:00"},{"Sha1":"23281b8f814b350af20b95ab6a4176cdbaf25b0a","Message":"chore(governance): 退役两份无生成脚本的报告，过期证据归零\n\n2026-09-17 目录负责人裁决：两份都退役（删文件 + 撤登记，不留例外）。\n\n两份都是 09-12 手工拼的一次性记录，仓内没有任何脚本能产出它们，因此永远\n无法重跑、只会一直过期；而它们的结论现在都由自动化门禁产出并正确绑定：\n\n- identity-http（8 例）：逐条比对确认与 runtime/test/e2e/identity-profile.test.ts\n 同名，今日已在 check:runtime 的 775 例里跑过并绑 6108357。其自述的\n provenanceMethod 就写着「Post-run source inspection」，本就不是门禁产物\n- mainline-candidate-regression（155 例）：作用域即 governance/{lib,test} 与\n 两个检查器，结论由 node --test governance/test/*.test.mjs 产出，今日 277/277。\n 155 这个数字自 09-12 起无人维护，保留它只会让人以为另有一道独立的候选回归门禁\n\n退役前已核：release-manifest.mjs 与 build-workbench-snapshots.mjs 都不引用\n这两份；其余引用只在 09-12 的历史记录文档与生成式快照字段里，会随重跑更新。\n\nevidence-scopes 登记 36 → 34，note 补记本次裁决与口径（同 stack-image-pins\n例外里写的「倾向 ① 退役」）。\n\ncheck:evidence：33 新鲜 / **0 过期** / 1 例外（stack-image-pins，另有出口）。\ngovernance 测试 277/277。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:54:10-07:00"},{"Sha1":"95d0ac8860038d4af6a3d15d4ca35ff0eff52749","Message":"chore(reports): 主线验收与恢复演练重跑通过 @ bc978a8\n\n在专用隔离栈 enterprise-platform-ms23-ci（PG :55485 / Redis :56395 /\nRedpanda :59105）上重跑，源库经 mainline:prepare 新建（platform_{fact,\npermission,audit,owner,consumer}_ci，全部非超级用户 / NOBYPASSRLS）。\n\n- mainline-acceptance passed，四套件 11 / 17 / 19 / 18：fact-read-http、\n fact-persistence、audit-tamper、mainline-revocation-chaos（含重复投递、\n 非连续乱序版本、毒消息 DLQ 与 broker 停机恢复）\n- mainline-restore passed，3 个库逐库 dump + 还原到新建 drill 库，表数\n 12 / 6 / …，逐库 SHA-256 一致；不删除、不重置源库与既往 drill\n- revocation-sla 随之刷新，仍 partial：invalidationPush 的 verdict 依旧是\n measured-with-harness-transport，生产传输未接线\n\n为什么换栈：恢复演练与 Redpanda 混沌用例的 allowlist 只认\nenterprise-platform-{ms23,ms23-ci,adversarial} 三个 compose 工程，\n共享的 ep-delivery 与 dev 常驻 stack 永远不匹配——这是故意的护栏，\n破坏性注入不得指向共享底座。dev 常驻栈全程未被触碰。\n\ncheck:evidence 过期 4 → 2，余下两份都没有生成脚本，另行处置。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:50:15-07:00"},{"Sha1":"bc978a8a99636b34e0bc69b018ee6ce8bd6192b9","Message":"chore(reports): identity 模块 UI 验收重跑通过 @ 828f178\n\n真实统一 Next 前端 + identity profile + 专用 PostgreSQL/Redis，2 个浏览器级\n用例全绿，由 90cda56 重绑 828f178（作用域内此前累计 21 个文件变更）。\n端口 web :7352 / api :7353。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:43:48-07:00"}],"HeadCommit":{"Sha1":"36e0423efe0ddc105136d193bfb594efb9838d62","Message":"chore(reports): 全链回绑 @ 1af0f22，证据零过期零告警\n\n干净树重跑根 pnpm check。check:evidence：34 份已登记，**33 份新鲜 / 0 过期 /\n0 输入未提交 / 0 告警 / 1 例外**（stack-image-pins，其例外自带出口）。\n\n这是本轮治理完善的收口：从九份过期、error 级判定，到零过期零告警。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:56:42-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/f9cb195319422eac06bd2111e1034e956ebcafb0...36e0423efe0ddc105136d193bfb594efb9838d62","Len":8}...
|
1789696615
|
Edit
Delete
|
|
30728
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"f9cb19531 {"Commits":[{"Sha1":"f9cb195319422eac06bd2111e1034e956ebcafb0","Message":"chore(reports): 工作台两份快照回绑 @ e94ad17\n\n经 pnpm reports:rebind 的多报告路径产出，两份同绑同一提交。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:22:33-07:00"},{"Sha1":"e94ad17db981932c5d0b08f2b99f2a786bdf36d7","Message":"feat(governance): rebind-reports 支持一个门禁多份报告，纳入工作台两份快照\n\nworkbench-catalog-snapshot 与 workbench-ops-snapshot 由同一个\nbuild-workbench-snapshots.mjs 产出，且它只 import node 内置与\ngovernance/lib/provenance.mjs——同样零依赖，本该在工具覆盖范围内。\n\n注册表 report:string → reports:string[]，三处判定与执行循环随之改为按份遍历。\n新增一条语义约束：同一门禁的多份报告要么全带回、要么一份都不带，否则两份会\n绑到不同提交，而它们描述的是同一次运行。两条负向用例分别钉住「任一份在途即\n整体跳过」与「任一份作用域有未提交输入即整体跳过」。\n\n实跑验证：两份快照同绑 aaf4836 / dirty=false。\ngovernance 测试 275 → 277。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:22:24-07:00"},{"Sha1":"aaf483645e83337d5d0a8a50b95df678adf0fb76","Message":"chore(reports): 过期证据 13 → 8，全部 error 清零 @ 4b067cd\n\n按偏差 #34 裁决落地后的第一轮清理。源码树干净（仅并行会话的 4 份 latest\n报告未提交，worktreeDirty 本就排除 *.latest.json），故可就地重跑并绑定。\n\n- port-conformance：41/41 passed，由 defd9d7 重绑 4b067cd。这是 09-17 那次\n 「跑通了却因无关文件脏着而不能采信」的收口\n- runtime 静态链 31 任务全过：naming / dual-backend-parity / fork-readiness /\n migrations 四份此前因并行会话提交 WB-1、runtime/apps 多出 workbench 而落进\n 作用域，现已重绑\n- workbench 两份快照重新生成（catalog projects=21 mapped=7/7 conflicts=0；\n ops controlled_operations=recorded）\n\ncheck:evidence 由 failed 转 partial：27 新鲜 / 8 过期 / 1 例外，0 error。\n余下 8 份全部需要真实环境才能重跑，不是本机静态能解决的：\nruntime-acceptance、ui-acceptance、revocation-sla、mainline-acceptance、\nmainline-restore、mainline-candidate-regression、identity-http、identity-ui。\n\n未触碰并行会话在途的 caddy / facts / migration-decs / module-imports 四份。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:21:25-07:00"}],"HeadCommit":{"Sha1":"f9cb195319422eac06bd2111e1034e956ebcafb0","Message":"chore(reports): 工作台两份快照回绑 @ e94ad17\n\n经 pnpm reports:rebind 的多报告路径产出，两份同绑同一提交。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:22:33-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/4b067cd4a76bfced89f76e979cf7bd36d3521a91...f9cb195319422eac06bd2111e1034e956ebcafb0","Len":3}...
|
1789694573
|
Edit
Delete
|
|
30727
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"689e924f3 {"Commits":[{"Sha1":"689e924f3db984f90ee13c2e0dd1a0082a70c1d3","Message":"docs(计划): 过期证据 9 → 8 且 error 清零；§1 证据行按 f9cb195 重采\n\n偏差 #34 裁决落地后的首轮清理：port-conformance 重绑、runtime 静态链四份\n重绑（WB-1 使 runtime/apps 变动而落进作用域）、工作台两份快照经\npnpm reports:rebind 重绑。check:evidence 由 failed 转 partial（27/8/1，0 error）。\n余 8 份全部需真实环境重跑，已逐一点名。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:22:48-07:00"}],"HeadCommit":{"Sha1":"689e924f3db984f90ee13c2e0dd1a0082a70c1d3","Message":"docs(计划): 过期证据 9 → 8 且 error 清零；§1 证据行按 f9cb195 重采\n\n偏差 #34 裁决落地后的首轮清理：port-conformance 重绑、runtime 静态链四份\n重绑（WB-1 使 runtime/apps 变动而落进作用域）、工作台两份快照经\npnpm reports:rebind 重绑。check:evidence 由 failed 转 partial（27/8/1，0 error）。\n余 8 份全部需真实环境重跑，已逐一点名。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T18:22:48-07:00"},"CompareURL":"luoanwu/platform-governance/compare/693800ceaf834b217a01466e19579af6a23925cf...689e924f3db984f90ee13c2e0dd1a0082a70c1d3","Len":1}...
|
1789694571
|
Edit
Delete
|
|
30725
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"693800cea {"Commits":[{"Sha1":"693800ceaf834b217a01466e19579af6a23925cf","Message":"chore(reports): 治理层门禁报告回绑 @ 87ec431\n\n清树重跑 run-foundation-gates：阻断性 11/11 通过，仅记录 2（工作区一致性\nL9 / L12）。三份报告 worktreeDirty=false，绑定 87ec431。\n\n（订正：本条信息初次写成「回绑 @ 9a5b3c0」，那是提交前预估的 SHA，实际\n实现提交是 87ec431。报告内容的绑定一直是对的，错的只是信息里的那串字。）\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:40:15-07:00"},{"Sha1":"87ec4315c77935405e16314c11789ddb6ae45e5a","Message":"feat(governance): G6 流外出口必须写复核期，缺失或过期即判红\n\n补上前一提交记为「仍缺一道」的缺口，不留登记而不建。\n\n根因：门禁消费登记的 G4 只断言 wireInWhen 非空，从不校验它是否已成立。\n流外出口是这份登记唯一的豁免口，而豁免口自己没有到期判据——用来防静音的\n机制在自己身上静音了。实锤就是前一提交处理的那条：check-domain-layer 的\n接入条件「D5 与 D9 清零」早已双双成立，G4 全绿，无人发现。\n\nG6 OUT_OF_FLOW_UNREVIEWED：流外条目必须写 conditionReviewBy（YYYY-MM-DD），\n缺失、格式不对或已过期即判红；当天不算过期；只对流外生效，阻断位条目每轮\n都在跑，不需要复核期。6 例负向覆盖这五种形态。\n\n守日期而不守条件是有意的：接入条件多半含人的判断（「框架同步波次扩散后」\n「再有派生仓开仓时」），机器判不了；能机器化的是「必须有人按期回来看一眼」。\n与同日落的域分层基线 B4 同形——缓期不是豁免。\n\n两条现存流外条目各自写了复核期与理由：\n- check-workspace-consistency 2026-10-31：L9 / L12 挂在活裁决 N-11 / C-14\n 上，六周内应有进展。\n- audit-platform-openings 2026-12-31：触发事件罕见，现存派生仓 0 个\n （派生仓状态.json 只剩 missing 1 + retired 18），不是六周尺度的事。\n\n台账 C-19 由「仍缺一道」改为已补，落点写明。\n\n证据：治理层 378 例 0 失败（新增 6 例负向）、阻断性门禁 11/11、\n工作区一致性仍 2 条（L9 / L12）。报告在脏树上产出，不随本提交。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:40:01-07:00"}],"HeadCommit":{"Sha1":"693800ceaf834b217a01466e19579af6a23925cf","Message":"chore(reports): 治理层门禁报告回绑 @ 87ec431\n\n清树重跑 run-foundation-gates：阻断性 11/11 通过，仅记录 2（工作区一致性\nL9 / L12）。三份报告 worktreeDirty=false，绑定 87ec431。\n\n（订正：本条信息初次写成「回绑 @ 9a5b3c0」，那是提交前预估的 SHA，实际\n实现提交是 87ec431。报告内容的绑定一直是对的，错的只是信息里的那串字。）\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:40:15-07:00"},"CompareURL":"luoanwu/platform-governance/compare/eca16cecb1b0e54e66d69268589ffa59c2fe6eff...693800ceaf834b217a01466e19579af6a23925cf","Len":2}...
|
1789688425
|
Edit
Delete
|
|
30724
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"eca16cecb {"Commits":[{"Sha1":"eca16cecb1b0e54e66d69268589ffa59c2fe6eff","Message":"chore(reports): 治理层门禁报告回绑 @ 3ca6a19\n\n清树重跑 run-foundation-gates：阻断性 11/11 通过（check-domain-layer 首次\n以阻断位计入，基线口径 0 阻断），仅记录 2（工作区一致性 L9 / L12）。\n三份报告 worktreeDirty=false，绑定 3ca6a19。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:37:02-07:00"},{"Sha1":"3ca6a19355e099d1643d6f03eeb59ae360011615","Message":"feat(governance): 域分层门禁转阻断位，按基线只拦新增\n\n查 L9 派活路径时撞见一条静音：门禁消费登记里 check-domain-layer 的接入\n条件写的是「D5（CHG-007 合并）与 D9（两份契约副本收敛）清零后转阻断位」，\n2026-09-17 实测 D5 = 0、D9 = 0——条件早已满足而无人察觉。\n\n根因：check-gate-consumption 的 G4 只断言 wireInWhen 非空，从不校验它是否\n已成立。该登记本是为治「门禁失败信号无人看」而建的，流外出口是它唯一的\n豁免口，而豁免口自己没有到期判据——用来防静音的机制在自己身上静音了。\n\n处置不能是直接翻阻断位：余下 13 条全在上层应用仓，治理层按回灌边界不代其\n提交，接入即恒红，而恒红正是该登记要治的病。原条件后半句「其余各条按\nOwner 各自的节奏消」已写明正确形态，只是从没人把它实现成基线。故：\n\n- 新增 域分层基线.json：13 条存量逐条具名 Owner 与复核期（D2 权威倒挂 3 ·\n D4 无事实出口 3 · D6 未装入基座 3 · D7 同版本分叉 1 · D10 已装内容不可\n 追溯 1 · D11 门禁零负向探针 2），来源登记指回台账 U-24 / U-32 / U-34 /\n U-36 与 §3 第 2 条。\n- 新增 scripts/lib/domain-layer-baseline.mjs：B1 新增 / B2 存量恶化 /\n B3 基线腐烂或缺 Owner / B4 过复核期，四类任一非空即阻断。B3 与 B4 守的是\n 基线自己——修好了必须删条目，过了复核期必须复核，否则基线会腐烂成豁免清单。\n- check-domain-layer 退出码改基线口径，--no-baseline 仍看未过滤的全量\n （13 条，退出 1）；registry 里该条转 run-foundation-gates 阻断位。\n\n一处要点：棘轮只认单调变坏的量。D10 的 matched（命中源仓提交的文件数）与\nD11 的 probes（负向探针条数）都是越大越好，拿来当上限会把改进判成恶化、\n反过来把修复锁死；已在 findingCount 里排除并写明理由，现只有 D2 的 count\n做棘轮。这条有单测守着。\n\n顺带更正 check-workspace-consistency 那条已过期的基线文字：L8 已归零\n（并行会话登记了具名冻结豁免），L9 的「12 仓」按当日实测改为 8 个上层\n应用仓；其接入形态改写为照此先例走基线棘轮，不再用「全清零再接入」——\n否则又是一个可能长期悬空的流外出口。\n\n仍缺一道，已记台账 C-19：G4 至今不会因「wireInWhen 已成立」而判红，\n下一个流外出口还能同样悬空。\n\n证据：治理层 377 例 0 失败（新增 12 例负向）、阻断性门禁 11/11、\n工作区一致性仍 2 条（L9 / L12）。L13 当场抓出收纳表漂移，已随本提交改数\n（根文件 62→63、scripts 50→51、tests 33→34）。报告在脏树上产出，\n不随本提交，清树重跑后另行回绑。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:36:39-07:00"}],"HeadCommit":{"Sha1":"eca16cecb1b0e54e66d69268589ffa59c2fe6eff","Message":"chore(reports): 治理层门禁报告回绑 @ 3ca6a19\n\n清树重跑 run-foundation-gates：阻断性 11/11 通过（check-domain-layer 首次\n以阻断位计入，基线口径 0 阻断），仅记录 2（工作区一致性 L9 / L12）。\n三份报告 worktreeDirty=false，绑定 3ca6a19。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:37:02-07:00"},"CompareURL":"luoanwu/platform-governance/compare/1645e7288b484ab39a4e3ebb2da44678e9b9e300...eca16cecb1b0e54e66d69268589ffa59c2fe6eff","Len":2}...
|
1789688225
|
Edit
Delete
|
|
30723
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"4b067cd4a {"Commits":[{"Sha1":"4b067cd4a76bfced89f76e979cf7bd36d3521a91","Message":"chore(reports): otel 回绑 @ e5cc39f（首次用 pnpm reports:rebind 产出）\n\n在主工作区有并行会话未提交改动的情况下，经 HEAD 干净检出生成，\nprovenance 绑 e5cc39f / worktreeDirty=false——偏差 #34-A 裁决实现的首次实用。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:32:14-07:00"},{"Sha1":"e5cc39f4dc74d56bce2b1861b15f8822ec3fdc3e","Message":"feat(governance): 偏差 #34 裁决实现——干净检出回绑工具 + 脏绑判定不再断言未核实的事\n\n2026-09-17 目录负责人裁决：A 走「干净 worktree 生成报告」（不动判据），\nB「修正断言」。\n\n#34-A｜新增 governance/rebind-reports.mjs（pnpm reports:rebind）\nprovenance.worktreeDirty 由全仓 git status 算，而 check:evidence 判过期是按\n每份报告自己的作用域算。口径不一致，导致只要仓里任何无关文件脏着就没法回绑\n任何报告——本工作区常有 2—3 个并行会话，干净窗口稀少，这是过期证据难清零的\n结构性成因。裁决不放宽 worktreeDirty，改为换个干净的地方生成：在 HEAD 开\ngit worktree 检出，在那里跑门禁，只把「绑 HEAD 且 dirty=false」的报告拷回。\n\n可行前提是 governance/ 零依赖（package.json 无 dependencies/devDependencies），\n临时检出不需要 pnpm install。实测：主工作区有 7 项未提交时，otel 报告仍绑到\n0c5d0b0 / dirty=false。\n\n三条护栏，都有负向用例：\n- 作用域内有未提交输入的门禁跳过（否则是拿 HEAD 的结论覆盖当前工作区）\n- 目标报告自己未提交的跳过（并行会话的在途产物，覆盖即销毁；--force 才 override）\n- 产出 provenance 不是干净 HEAD 绑定就不带回\n范围外的说清楚不假装：check:fixtures / check:port-conformance 与 runtime、\ncontracts 的检查需要构建产物，必须在装好依赖的检出里跑。\n\n顺带记一条坑：入口判定不能写 `import.meta.url === file://${argv[1]}`——本仓\n路径含中文，import.meta.url 是百分号编码的，比较永远不等，脚本会静默不做事。\n\n#34-B｜check-evidence-freshness 的 EVIDENCE_BOUND_DIRTY\n该分支读的是工作区里的报告文件，定级却用 currentTreeDirty，而后者走\ngitProvenance() 并显式排除 *.latest.json。于是并行会话跑门禁、只有报告文件脏\n时，当前树被算作干净，4 份未提交的在途报告被判 error 并输出「脏绑报告被提交\n了」——该断言为假（那 4 份的 HEAD 版本干净绑 8382c04）。现按报告文件自身是否\n未提交区分：在途产物降 info 并如实描述；确系已提交的脏绑仍是 error。判据未放宽，\n两条负向用例分别钉住这两侧。\n\ngovernance 测试 266 → 275。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:32:04-07:00"}],"HeadCommit":{"Sha1":"4b067cd4a76bfced89f76e979cf7bd36d3521a91","Message":"chore(reports): otel 回绑 @ e5cc39f（首次用 pnpm reports:rebind 产出）\n\n在主工作区有并行会话未提交改动的情况下，经 HEAD 干净检出生成，\nprovenance 绑 e5cc39f / worktreeDirty=false——偏差 #34-A 裁决实现的首次实用。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:32:14-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/0c5d0b09bc2d1c4baddfd59b484c5fdcbe0917ab...4b067cd4a76bfced89f76e979cf7bd36d3521a91","Len":2}...
|
1789687994
|
Edit
Delete
|
|
30722
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"1645e7288 {"Commits":[{"Sha1":"1645e7288b484ab39a4e3ebb2da44678e9b9e300","Message":"docs(计划): 偏差 #34 已裁决并实现，移入已关闭（真源仓 e5cc39f / 4b067cd）\n\n目录负责人 2026-09-17 裁决：\n- A「干净检出生成报告」：不放宽 worktreeDirty，改为 pnpm reports:rebind 在\n HEAD 开 git worktree 跑零依赖门禁、只带回干净绑定的报告\n- B「修正断言」：EVIDENCE_BOUND_DIRTY 按报告文件自身是否未提交区分，在途\n 产物降 info 并如实描述，已提交的脏绑仍 error\n\n§6 中 #34 由开放表移入已关闭一行并附实现与证据；§8 增一行记录裁决与实现细节\n（含范围外声明与三条护栏）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:33:10-07:00"}],"HeadCommit":{"Sha1":"1645e7288b484ab39a4e3ebb2da44678e9b9e300","Message":"docs(计划): 偏差 #34 已裁决并实现，移入已关闭（真源仓 e5cc39f / 4b067cd）\n\n目录负责人 2026-09-17 裁决：\n- A「干净检出生成报告」：不放宽 worktreeDirty，改为 pnpm reports:rebind 在\n HEAD 开 git worktree 跑零依赖门禁、只带回干净绑定的报告\n- B「修正断言」：EVIDENCE_BOUND_DIRTY 按报告文件自身是否未提交区分，在途\n 产物降 info 并如实描述，已提交的脏绑仍 error\n\n§6 中 #34 由开放表移入已关闭一行并附实现与证据；§8 增一行记录裁决与实现细节\n（含范围外声明与三条护栏）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:33:10-07:00"},"CompareURL":"luoanwu/platform-governance/compare/1e0ea08130bcbed848d075740bb4b5c44c1ffde4...1645e7288b484ab39a4e3ebb2da44678e9b9e300","Len":1}...
|
1789687994
|
Edit
Delete
|
|
30721
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"1e0ea0813 {"Commits":[{"Sha1":"1e0ea08130bcbed848d075740bb4b5c44c1ffde4","Message":"chore(reports): 治理层门禁报告回绑 @ 6e0d7aa\n\n清树重跑 run-foundation-gates：阻断性 10/10 通过，仅记录 3\n（工作区一致性 2 条 L9 / L12、域分层 13 条，均不参与总账）。\n三份报告 worktreeDirty=false，绑定 6e0d7aa。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:30:08-07:00"},{"Sha1":"6e0d7aa2570ce0046c8e1860044a8547f8567c23","Message":"docs(治理): L9 收口为「根因已闭、扩散无渠道」，登记 N-11 交派活\n\n核到底的结论是 L9 不该按「待修的缺陷」处理：\n\n- 模板根因早已关闭。框架 0.17.0（实现提交 313137b）给 compose 写了\n name: ${COMPOSE_PROJECT_NAME:-base-framework} 与 ${REDIS_PORT:-6379}，\n F14 在框架 pnpm check 链里守全仓每一份 compose，scaffold-derived.mjs\n 派生时重绑身份，变更请求文首已记「已受理并落地」。enterprise-platform\n 两份 compose 也已带身份。新建仓不再扩散该缺陷。\n- 但扩散没有发生，且没有渠道可走。回灌台账 §3 第 8 条指定的\n 框架新版本同步计划自述「同步波次事实上已无目标」（18 个派生仓已退役），\n 它也从不覆盖上层应用仓。L9 当日实测仍有 8 个应用仓共用 project\n api-nestjs，且这 8 个仓自己的门禁一个都不报：5 个带 F14 之前的旧版\n check-fork-readiness.mjs（文件内 F14 零命中），3 个没有该脚本——\n 它们的 pnpm check 在冲突状态下仍是绿的。\n- 台账原话「在模板修好前，任何两个仓不得同时 docker compose up」的前提\n 已不成立而风险未变，按实际判据改述为「上列 8 个仓两两之间不得同时 up」。\n\n治理层按回灌边界「不代业务应用仓提交」，不自行下手，本轮只落记录：\nC-15 与 §3 第 8 条加订正（append-only，不改写原文），新登记待裁决 N-11\n交目录负责人三选一，并写明 0.17.0 的 F14 是破坏性变更——改 project 名后\n首次 up 会建新空卷，旧卷 api-nestjs_* 的开发数据不跟着走，这正是 N-09\n当时搁置同步的那一条，选任一路线都要先答共用卷里的数据归谁。\n\nL9 保持红，不降级。\n\n证据：治理层 365 例 0 失败、阻断性门禁 10/10；工作区一致性仍 2 条\n（L9 / L12），无新增断链与措辞违规。报告在脏树上产出，不随本提交，\n清树重跑后另行回绑。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:29:41-07:00"}],"HeadCommit":{"Sha1":"1e0ea08130bcbed848d075740bb4b5c44c1ffde4","Message":"chore(reports): 治理层门禁报告回绑 @ 6e0d7aa\n\n清树重跑 run-foundation-gates：阻断性 10/10 通过，仅记录 3\n（工作区一致性 2 条 L9 / L12、域分层 13 条，均不参与总账）。\n三份报告 worktreeDirty=false，绑定 6e0d7aa。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:30:08-07:00"},"CompareURL":"luoanwu/platform-governance/compare/d5f1b641b05f0b3a351c8792b64d7215f3356d38...1e0ea08130bcbed848d075740bb4b5c44c1ffde4","Len":2}...
|
1789687811
|
Edit
Delete
|
|
30720
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"0c5d0b09b {"Commits":[{"Sha1":"0c5d0b09bc2d1c4baddfd59b484c5fdcbe0917ab","Message":"fix(runtime): 框架残留 compose 补 project 名——退出 api-nestjs 撞名（C-13 / 一致性门禁 L9）\n\nruntime/apps/api-nestjs/docker-compose.yml 未声明顶层 name，docker compose 按目录名推导为\napi-nestjs，与工作区内另外 8 个仓的同名 compose 共用容器与具名卷（api-nestjs_pgdata /\nredisdata / miniodata）；后 up 的那个会复用前一个的数据。框架仓同名文件已写成\n`name: ${COMPOSE_PROJECT_NAME:-base-framework}`，本仓这份是更早版本的残留。\n\n按同一写法补上，默认值取本仓专属 enterprise-platform-runtime-dev（不沿用框架默认，否则\n改为与框架仓撞名）。本文件是框架默认 dev compose，其端口 5432 / 6379 与本仓实际底座\n（stack/compose.yaml，project enterprise-platform-stack）无关，改名不影响任何在跑的服务。\n\n一致性门禁 L9 参与仓 9 → 8；余 8 个仓在业务应用层与企业领域应用层，按 C-13 由各仓 Owner\n与 SRE 处置。三组卷现状：api-nestjs_* 与 enterprise-idp_* 共 5 个卷，所属容器均为\ncreated / exited 非运行态；C-13 要求的 pg_dump 尚未做（仓备份-2026-09-16/ 目前只有 4 个 git bundle）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:25:43-07:00"}],"HeadCommit":{"Sha1":"0c5d0b09bc2d1c4baddfd59b484c5fdcbe0917ab","Message":"fix(runtime): 框架残留 compose 补 project 名——退出 api-nestjs 撞名（C-13 / 一致性门禁 L9）\n\nruntime/apps/api-nestjs/docker-compose.yml 未声明顶层 name，docker compose 按目录名推导为\napi-nestjs，与工作区内另外 8 个仓的同名 compose 共用容器与具名卷（api-nestjs_pgdata /\nredisdata / miniodata）；后 up 的那个会复用前一个的数据。框架仓同名文件已写成\n`name: ${COMPOSE_PROJECT_NAME:-base-framework}`，本仓这份是更早版本的残留。\n\n按同一写法补上，默认值取本仓专属 enterprise-platform-runtime-dev（不沿用框架默认，否则\n改为与框架仓撞名）。本文件是框架默认 dev compose，其端口 5432 / 6379 与本仓实际底座\n（stack/compose.yaml，project enterprise-platform-stack）无关，改名不影响任何在跑的服务。\n\n一致性门禁 L9 参与仓 9 → 8；余 8 个仓在业务应用层与企业领域应用层，按 C-13 由各仓 Owner\n与 SRE 处置。三组卷现状：api-nestjs_* 与 enterprise-idp_* 共 5 个卷，所属容器均为\ncreated / exited 非运行态；C-13 要求的 pg_dump 尚未做（仓备份-2026-09-16/ 目前只有 4 个 git bundle）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:25:43-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/ac396a0f27fd673e02fc0c2f5887ea7f06b879b4...0c5d0b09bc2d1c4baddfd59b484c5fdcbe0917ab","Len":1}...
|
1789687572
|
Edit
Delete
|
|
30719
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"ac396a0f2 {"Commits":[{"Sha1":"ac396a0f27fd673e02fc0c2f5887ea7f06b879b4","Message":"chore(governance): evidence-scopes 口径说明补第五类——有 provenance 但无生成脚本的一次性记录\n\n那段说明自 2026-09-15 起写「余 15 份未登记，都是有意不登」并逐类枚举，\n但今天实测是 16 份，且 client-identity 与 governance-truth-validation 两份\n不在任何一类里——说明本身成了过期声明，而它正是「为什么不登」的唯一依据。\n\n全仓搜索确认这两份没有生成脚本（与 caddy / workbench-snapshots 这类有\ngenerator 的登记项相反），属手工一次性记录：可以有合法 provenance，但不可\n重跑，登记后必然过期且只能靠例外掩盖。故新增第五类并写明：\n\n- 与 reports/stack-image-pins.latest.json 同因不同果——后者当年登记了，\n 于是只能挂一条带出口的例外；第五类因未登记而无需例外\n- 出口同样二选一：补生成脚本后转正式登记，或退役该报告\n- 在此之前它们只是记录，不得被引用为可重跑的机器证据\n\n计数改为带日期的复核快照。未新增或删除任何作用域登记，未改判据，未动文件\n其余格式（差异 1 行）。check:evidence 仍 partial（26 新鲜 / 9 过期 / 1 例外，\n与改前逐项相同）；governance 测试 266/266。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:23:44-07:00"}],"HeadCommit":{"Sha1":"ac396a0f27fd673e02fc0c2f5887ea7f06b879b4","Message":"chore(governance): evidence-scopes 口径说明补第五类——有 provenance 但无生成脚本的一次性记录\n\n那段说明自 2026-09-15 起写「余 15 份未登记，都是有意不登」并逐类枚举，\n但今天实测是 16 份，且 client-identity 与 governance-truth-validation 两份\n不在任何一类里——说明本身成了过期声明，而它正是「为什么不登」的唯一依据。\n\n全仓搜索确认这两份没有生成脚本（与 caddy / workbench-snapshots 这类有\ngenerator 的登记项相反），属手工一次性记录：可以有合法 provenance，但不可\n重跑，登记后必然过期且只能靠例外掩盖。故新增第五类并写明：\n\n- 与 reports/stack-image-pins.latest.json 同因不同果——后者当年登记了，\n 于是只能挂一条带出口的例外；第五类因未登记而无需例外\n- 出口同样二选一：补生成脚本后转正式登记，或退役该报告\n- 在此之前它们只是记录，不得被引用为可重跑的机器证据\n\n计数改为带日期的复核快照。未新增或删除任何作用域登记，未改判据，未动文件\n其余格式（差异 1 行）。check:evidence 仍 partial（26 新鲜 / 9 过期 / 1 例外，\n与改前逐项相同）；governance 测试 266/266。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:23:44-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/bd625a1a1535a34e265b11424a646618c543ffef...ac396a0f27fd673e02fc0c2f5887ea7f06b879b4","Len":1}...
|
1789687513
|
Edit
Delete
|
|
30718
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"d5f1b641b {"Commits":[{"Sha1":"d5f1b641b05f0b3a351c8792b64d7215f3356d38","Message":"docs(计划): 偏差 #34 补第二个缺陷；登记 evidence-scopes 口径修正（真源仓 ac396a0）\n\nEVIDENCE_BOUND_DIRTY 读工作区里的报告文件，判定严重级所用的 currentTreeDirty\n却走 gitProvenance()，而后者显式排除 *.latest.json。并行会话正在跑门禁、只有\n报告文件脏时，门禁认定「当前树干净」，把 4 份尚未提交的在途报告判成 error 并\n输出「脏绑报告被提交了」——该断言在此情形下为假：那 4 份的 HEAD 版本干净绑定\n8382c04，只有工作区版本绑 bd625a1 + dirty。对比工作区版本与 HEAD 版本即可分辨，\n门禁不该断言它没有核实的事。两点同属一条规则，建议一并处置。\n\n§8 同时登记 evidence-scopes 口径说明补第五类（真源仓 ac396a0）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:25:08-07:00"}],"HeadCommit":{"Sha1":"d5f1b641b05f0b3a351c8792b64d7215f3356d38","Message":"docs(计划): 偏差 #34 补第二个缺陷；登记 evidence-scopes 口径修正（真源仓 ac396a0）\n\nEVIDENCE_BOUND_DIRTY 读工作区里的报告文件，判定严重级所用的 currentTreeDirty\n却走 gitProvenance()，而后者显式排除 *.latest.json。并行会话正在跑门禁、只有\n报告文件脏时，门禁认定「当前树干净」，把 4 份尚未提交的在途报告判成 error 并\n输出「脏绑报告被提交了」——该断言在此情形下为假：那 4 份的 HEAD 版本干净绑定\n8382c04，只有工作区版本绑 bd625a1 + dirty。对比工作区版本与 HEAD 版本即可分辨，\n门禁不该断言它没有核实的事。两点同属一条规则，建议一并处置。\n\n§8 同时登记 evidence-scopes 口径说明补第五类（真源仓 ac396a0）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:25:08-07:00"},"CompareURL":"luoanwu/platform-governance/compare/bc850cd56895bf917bf6ec91b565f582e3dc601d...d5f1b641b05f0b3a351c8792b64d7215f3356d38","Len":1}...
|
1789687511
|
Edit
Delete
|
|
30717
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"bc850cd56 {"Commits":[{"Sha1":"bc850cd56895bf917bf6ec91b565f582e3dc601d","Message":"docs(决策): C-12 drift 升档已执行（真源仓 8382c04）\n\n决策批次 C-12 的两个前置（A-2 合并、drift 清零）均满足后执行升档：catalog:drift 的 error 与 drift 两档一律 exit 1，去掉从无调用方的 --strict 开关。逐条登记实况比原裁决所记更弱的那一点（此前连 error 都不阻断），并记 CLI 层四段负向。开发计划 §8 同步一行。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:23:08-07:00"}],"HeadCommit":{"Sha1":"bc850cd56895bf917bf6ec91b565f582e3dc601d","Message":"docs(决策): C-12 drift 升档已执行（真源仓 8382c04）\n\n决策批次 C-12 的两个前置（A-2 合并、drift 清零）均满足后执行升档：catalog:drift 的 error 与 drift 两档一律 exit 1，去掉从无调用方的 --strict 开关。逐条登记实况比原裁决所记更弱的那一点（此前连 error 都不阻断），并记 CLI 层四段负向。开发计划 §8 同步一行。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:23:08-07:00"},"CompareURL":"luoanwu/platform-governance/compare/d991c449457408aed0e9b98accb23feffdcc3d92...bc850cd56895bf917bf6ec91b565f582e3dc601d","Len":1}...
|
1789687392
|
Edit
Delete
|
|
30716
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"d991c4494 {"Commits":[{"Sha1":"d991c449457408aed0e9b98accb23feffdcc3d92","Message":"fix(governance): L8 补冻结物的显式豁免出口——审批记录条目写明，缺理由不放行\n\nL2 有 pragma 与 frozenInputs 两条出口，L8 一条也没有：冻结物改一字节即触发 L7 摘要不符，\n文件内的`一致性门禁豁免`标记对它无效，于是命中即永久噪声，与本门禁自己立的收敛原则\n（报出来的每一条都应当是能直接去改的）相反。\n\n出口改为由审批记录条目写明 consistencyExempt + consistencyExemptNote，缺 note 不放行；\n豁免以 frozenExempted 计数与 frozenExemptions 明细出现在 JSON 输出里，不静默。\n两个字段都不进 packageFingerprint（只取 path + sha256），登记前后指纹实测未变 1fdc2e1666c5。\n\n首条登记：新增基础应用MachineCatalog完整变更.patch。核验结论是误报——命中只 1 处、位于\ndiff 的上下文行（补丁并未引入），且处在一句陈述「旧断言写成了迁移双替换产物、等于没断言」的\n注释里，是对缺陷的描述而非被使用的路径。解除条件：变更包合并或撤回归档后随条目移除。\n\n验证：一致性门禁 3 → 2（余 L9 compose 撞名、L12 内核 pin 落后，均有既定归属）；\ntests/workspace-consistency.test.mjs 35/35；变更包就绪门禁与评审门禁 exit 0。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:22:29-07:00"}],"HeadCommit":{"Sha1":"d991c449457408aed0e9b98accb23feffdcc3d92","Message":"fix(governance): L8 补冻结物的显式豁免出口——审批记录条目写明，缺理由不放行\n\nL2 有 pragma 与 frozenInputs 两条出口，L8 一条也没有：冻结物改一字节即触发 L7 摘要不符，\n文件内的`一致性门禁豁免`标记对它无效，于是命中即永久噪声，与本门禁自己立的收敛原则\n（报出来的每一条都应当是能直接去改的）相反。\n\n出口改为由审批记录条目写明 consistencyExempt + consistencyExemptNote，缺 note 不放行；\n豁免以 frozenExempted 计数与 frozenExemptions 明细出现在 JSON 输出里，不静默。\n两个字段都不进 packageFingerprint（只取 path + sha256），登记前后指纹实测未变 1fdc2e1666c5。\n\n首条登记：新增基础应用MachineCatalog完整变更.patch。核验结论是误报——命中只 1 处、位于\ndiff 的上下文行（补丁并未引入），且处在一句陈述「旧断言写成了迁移双替换产物、等于没断言」的\n注释里，是对缺陷的描述而非被使用的路径。解除条件：变更包合并或撤回归档后随条目移除。\n\n验证：一致性门禁 3 → 2（余 L9 compose 撞名、L12 内核 pin 落后，均有既定归属）；\ntests/workspace-consistency.test.mjs 35/35；变更包就绪门禁与评审门禁 exit 0。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:22:29-07:00"},"CompareURL":"luoanwu/platform-governance/compare/cfaef2fc83837096a19a0d330b709017c0f322dd...d991c449457408aed0e9b98accb23feffdcc3d92","Len":1}...
|
1789687361
|
Edit
Delete
|
|
30715
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"bd625a1a1 {"Commits":[{"Sha1":"bd625a1a1535a34e265b11424a646618c543ffef","Message":"chore(reports): catalog:drift 升档后全量回绑 @ 8382c04（干净树）\n\n根 check 链 exit 0（catalog:drift 升档后 error 0 / drift 0 仍通过）、runtime 16 项静态门禁、工作台两份快照重生成。\n\n顺带清掉 7 条 EVIDENCE_BOUND_DIRTY：caddy / fixtures / facts / migration-decs / module-imports / catalog-reconciliation / otel 此前由 a7a19ec（PF-03 ⑤ 回绑）带进干净树时绑在脏树上，本次在干净树重跑后绑定 8382c04 / worktreeDirty=false。check:evidence 由 failed 回到 partial（新鲜 26 / 过期 9 / 无效 0），余 9 条过期均为需真实库或容器重跑的 warn 级。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:22:31-07:00"},{"Sha1":"8382c0432e3c0cc2f2c6f379a2374d5f6e2cc9d1","Message":"feat(governance): catalog:drift 升档为阻断（决策批次 2026-09-16 C-12）\n\nC-12 的两个前置都已满足：A-2 已应用（CHG-002 于本轮落地），drift 已清零（error 0 / drift 0 / info 5 / unknown 0）。\n\n升档前的实况比\"能发现、不能阻断\"更弱：只有 `--strict` 才对 `error` 阻断，而根 `pnpm check` 从不传该开关、全仓也没有任何调用方传过——这道门禁装上之后一次都没拦过。现改为 `error` 与 `drift` 一律 exit 1，不再需要开关；`info`（authority≠E2 的刻意保留）与 `unknown`（独立克隆缺仓外资料）不阻断，独立克隆的失败关闭由 contracts check:local 与完整 check 承担。输出行的 ✓/✗ 同步按两档判。\n\n测试补 CLI 层四段：干净 exit 0、drift exit 1（升档前此处为 0）、error exit 1、工作区不可用只记 unknown 不阻断。check-catalog-drift 6/6。\n\n另修工作台运维快照：evidence 一档原样透传 check:evidence 的结论，passed / partial / failed 都是合法取值，此前允许集合漏了 failed——证据一过期就把快照测试打红，而不是让快照如实显示过期。规则行与测试同步补上，仍不含 available / healthy 这类会被误读成\"环境可用\"的词。governance 266/266。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:21:32-07:00"}],"HeadCommit":{"Sha1":"bd625a1a1535a34e265b11424a646618c543ffef","Message":"chore(reports): catalog:drift 升档后全量回绑 @ 8382c04（干净树）\n\n根 check 链 exit 0（catalog:drift 升档后 error 0 / drift 0 仍通过）、runtime 16 项静态门禁、工作台两份快照重生成。\n\n顺带清掉 7 条 EVIDENCE_BOUND_DIRTY：caddy / fixtures / facts / migration-decs / module-imports / catalog-reconciliation / otel 此前由 a7a19ec（PF-03 ⑤ 回绑）带进干净树时绑在脏树上，本次在干净树重跑后绑定 8382c04 / worktreeDirty=false。check:evidence 由 failed 回到 partial（新鲜 26 / 过期 9 / 无效 0），余 9 条过期均为需真实库或容器重跑的 warn 级。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:22:31-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/a7a19ec54a9f4ed81c770afa85a19125130e85a5...bd625a1a1535a34e265b11424a646618c543ffef","Len":2}...
|
1789687353
|
Edit
Delete
|
|
30714
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"cfaef2fc8 {"Commits":[{"Sha1":"cfaef2fc83837096a19a0d330b709017c0f322dd","Message":"docs(计划): 登记偏差 #34——脏树绑定判据与作用域判据不一致，卡住报告回绑\n\n领取「九条过期证据里 port-conformance 是唯一纯本机项」后核出的机制缺口。\n\ncheck-evidence-freshness 对未提交输入已按作用域判（EVIDENCE_WORKTREE_CHANGED\n只在脏文件落进该报告 scope 才报），但 EVIDENCE_BOUND_DIRTY 仍取\nprovenance.worktreeDirty 这个仓级布尔（provenance.mjs 由全仓 git status 算，\n只排除 latest 报告自身）。\n\n实锤：pnpm check:port-conformance 重跑 41/41 passed，当时唯一脏文件是并行\n会话在改的 governance/check-catalog-drift.mjs 及其测试，不在该报告作用域\n（runtime/{modules,clients,packages} + 检查器 + lockfile）内，报告仍写\nworktreeDirty=true；提交后树一转干净即被判 error。因此本轮没有提交这次回绑，\n九份过期证据仍是九份。\n\n§6 增 #34 并给三个处置选项（维持现状 / 作用域化 / 折中声明式），标明待平台\n负责人裁、裁决前不得放宽也不得把脏绑报告当证据提交；§8 增一行。\n\n未改任何门禁代码。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:21:58-07:00"}],"HeadCommit":{"Sha1":"cfaef2fc83837096a19a0d330b709017c0f322dd","Message":"docs(计划): 登记偏差 #34——脏树绑定判据与作用域判据不一致，卡住报告回绑\n\n领取「九条过期证据里 port-conformance 是唯一纯本机项」后核出的机制缺口。\n\ncheck-evidence-freshness 对未提交输入已按作用域判（EVIDENCE_WORKTREE_CHANGED\n只在脏文件落进该报告 scope 才报），但 EVIDENCE_BOUND_DIRTY 仍取\nprovenance.worktreeDirty 这个仓级布尔（provenance.mjs 由全仓 git status 算，\n只排除 latest 报告自身）。\n\n实锤：pnpm check:port-conformance 重跑 41/41 passed，当时唯一脏文件是并行\n会话在改的 governance/check-catalog-drift.mjs 及其测试，不在该报告作用域\n（runtime/{modules,clients,packages} + 检查器 + lockfile）内，报告仍写\nworktreeDirty=true；提交后树一转干净即被判 error。因此本轮没有提交这次回绑，\n九份过期证据仍是九份。\n\n§6 增 #34 并给三个处置选项（维持现状 / 作用域化 / 折中声明式），标明待平台\n负责人裁、裁决前不得放宽也不得把脏绑报告当证据提交；§8 增一行。\n\n未改任何门禁代码。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:21:58-07:00"},"CompareURL":"luoanwu/platform-governance/compare/a9e489b08d6d8dd510c9f63dd663b94037d5f9f4...cfaef2fc83837096a19a0d330b709017c0f322dd","Len":1}...
|
1789687322
|
Edit
Delete
|
|
30713
|
5
|
5
|
5
|
116
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"a7a19ec54 {"Commits":[{"Sha1":"a7a19ec54a9f4ed81c770afa85a19125130e85a5","Message":"chore(reports): PF-03 ⑤ 回灌后静态门禁回绑 @ a084c5e\n\n干净树上重跑根 pnpm check（31 个 turbo 任务全过），26 份报告绑定\na084c5e / worktreeDirty=false。同时结清 41ca988 当时因并行会话在途、\n拿不到干净树而欠下的四份文档作用域报告（docs-truth、governance-docs、\nagent-adapters、dual-backend-parity）。\n\ncheck:evidence 仍为 partial：36 份已登记，26 新鲜 / 9 过期 / 1 例外。\n九条过期均为既有项，都需要真实环境重跑才能消除，不在本轮范围：\nport-conformance、runtime/ui-acceptance、revocation-sla、\nmainline-acceptance、mainline-restore、mainline-candidate-regression、\nidentity-http、identity-ui。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:18:40-07:00"},{"Sha1":"a084c5e5f61f25a4f70efca3c42a763a1e7c9b8d","Message":"docs(runtime): PF-03 ⑤ 收口——README 与 CHANGELOG 声明本目录身份与版本轴归属\n\n承 41ca988。这两份随 scaffold 带入的框架文档不含错误的验收数字，问题是\n身份：读者落到 runtime/ 会以为自己在基础框架仓。\n\nREADME：\n- 标题由「基础框架 / Base Framework」改为「enterprise-platform/runtime\n （基础框架 v0.16.1 的 scaffold 实例）」，开篇说明下文绝大部分是框架原文\n 且仍准确，并划清三条界线——版本轴属框架、状态只认本仓报告与 CLAUDE.md\n 动态区、领取顺序在开发计划不设第二份\n- 目录结构补上此前完全缺失的本仓特有部分：modules/（kit + 七模块，\n modules.json + PLATFORM_PROFILES + 一模块一库）、clients/（@juhai/client-*）、\n test/（platform-tests），apps/ 补 workbench 与 api-nestjs 的平台装配点\n\nCHANGELOG：加一段说明这份轴属于框架、随 scaffold 带入，package.json 的\n0.16.1 读作「本运行时基于框架 0.16.1」而非平台版本；平台自己的发布轴是\nRelease Manifest、@juhai/* 包版本与开发计划 §2 的里程碑，平台侧新增能力\n不在此登记。\n\n框架原文的门禁、命令与 ADR 一律未动。check:docs-truth 与\ncheck:governance-docs 均通过（README 的信封形状与 api-fastify 定位断言\n未受影响）。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:17:59-07:00"}],"HeadCommit":{"Sha1":"a7a19ec54a9f4ed81c770afa85a19125130e85a5","Message":"chore(reports): PF-03 ⑤ 回灌后静态门禁回绑 @ a084c5e\n\n干净树上重跑根 pnpm check（31 个 turbo 任务全过），26 份报告绑定\na084c5e / worktreeDirty=false。同时结清 41ca988 当时因并行会话在途、\n拿不到干净树而欠下的四份文档作用域报告（docs-truth、governance-docs、\nagent-adapters、dual-backend-parity）。\n\ncheck:evidence 仍为 partial：36 份已登记，26 新鲜 / 9 过期 / 1 例外。\n九条过期均为既有项，都需要真实环境重跑才能消除，不在本轮范围：\nport-conformance、runtime/ui-acceptance、revocation-sla、\nmainline-acceptance、mainline-restore、mainline-candidate-regression、\nidentity-http、identity-ui。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:18:40-07:00"},"CompareURL":"luoanwu/enterprise-platform/compare/8dae61f91b56f708b1c76e66f07d98efa4828a36...a7a19ec54a9f4ed81c770afa85a19125130e85a5","Len":2}...
|
1789687171
|
Edit
Delete
|
|
30712
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"a9e489b08 {"Commits":[{"Sha1":"a9e489b08d6d8dd510c9f63dd663b94037d5f9f4","Message":"docs(计划): PF-03 收口完成（真源仓 a084c5e，报告回绑 a7a19ec）\n\n- §3 PF-03 行改为「✅ 完成」，退出条件落为「三份文档均声明本目录身份与\n 版本轴归属、不含框架仓状态声明」\n- §5 ⑤ 移除已完成的「PF-03 ⑤ 文档回灌」\n- §8 增一行，含 check:evidence 仍 partial 的如实登记\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:19:27-07:00"}],"HeadCommit":{"Sha1":"a9e489b08d6d8dd510c9f63dd663b94037d5f9f4","Message":"docs(计划): PF-03 收口完成（真源仓 a084c5e，报告回绑 a7a19ec）\n\n- §3 PF-03 行改为「✅ 完成」，退出条件落为「三份文档均声明本目录身份与\n 版本轴归属、不含框架仓状态声明」\n- §5 ⑤ 移除已完成的「PF-03 ⑤ 文档回灌」\n- §8 增一行，含 check:evidence 仍 partial 的如实登记\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:19:27-07:00"},"CompareURL":"luoanwu/platform-governance/compare/15bce0f4b672fc7456786bc61d493b3ffc04e630...a9e489b08d6d8dd510c9f63dd663b94037d5f9f4","Len":1}...
|
1789687171
|
Edit
Delete
|
|
30711
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"15bce0f4b {"Commits":[{"Sha1":"15bce0f4b672fc7456786bc61d493b3ffc04e630","Message":"docs(平台治理): 收纳表随新增门禁改数——scripts 48→50、tests 32→33、reports 324→325（顶层 39→40）\n\ncheck-deferred-ops.mjs 与 lib/deferred-ops-core.mjs、deferred-ops.test.mjs、延后op.latest.json 四个新文件没同步收纳表，上一条提交（25751af）因此把 L13 三条漂移推了出去。本次补齐，L13 清零。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:18:34-07:00"}],"HeadCommit":{"Sha1":"15bce0f4b672fc7456786bc61d493b3ffc04e630","Message":"docs(平台治理): 收纳表随新增门禁改数——scripts 48→50、tests 32→33、reports 324→325（顶层 39→40）\n\ncheck-deferred-ops.mjs 与 lib/deferred-ops-core.mjs、deferred-ops.test.mjs、延后op.latest.json 四个新文件没同步收纳表，上一条提交（25751af）因此把 L13 三条漂移推了出去。本次补齐，L13 清零。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:18:34-07:00"},"CompareURL":"luoanwu/platform-governance/compare/25751af083e566216e1803734573cfdb8184e65b...15bce0f4b672fc7456786bc61d493b3ffc04e630","Len":1}...
|
1789687118
|
Edit
Delete
|
|
30710
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"25751af08 {"Commits":[{"Sha1":"25751af083e566216e1803734573cfdb8184e65b","Message":"docs(开发计划): §8 补「延后 op 门禁与两处收口」一行\n\n登记 check-deferred-ops 的建立、它当场抓出的两个真问题（CHG-003/006 同条目冲突的延后 op、CHG-003 enterprise-idp 的半应用）与随之收口的两处恒红/崩溃。一致性检查 L4 / L10 / L13 清零。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:17:51-07:00"}],"HeadCommit":{"Sha1":"25751af083e566216e1803734573cfdb8184e65b","Message":"docs(开发计划): §8 补「延后 op 门禁与两处收口」一行\n\n登记 check-deferred-ops 的建立、它当场抓出的两个真问题（CHG-003/006 同条目冲突的延后 op、CHG-003 enterprise-idp 的半应用）与随之收口的两处恒红/崩溃。一致性检查 L4 / L10 / L13 清零。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:17:51-07:00"},"CompareURL":"luoanwu/platform-governance/compare/96dbcd998453ba6720203ea63ab612caa4d34687...25751af083e566216e1803734573cfdb8184e65b","Len":1}...
|
1789687075
|
Edit
Delete
|
|
30709
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"96dbcd998 {"Commits":[{"Sha1":"96dbcd998453ba6720203ea63ab612caa4d34687","Message":"chore(reports): 延后op 回绑 @ 17b81dd（clean 树，worktreeDirty=false）\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:17:30-07:00"},{"Sha1":"17b81dde8df1c9c426b95261a8d7c1ad60054bd2","Message":"fix(governance): check-deferred-ops 改用治理层统一 provenance 口径\n\n原先自己数 `git status --porcelain` 行数判 worktreeDirty，把本轮要覆盖的 latest 报告也算成脏文件——字段随即恒为 true，一个永远为真的 provenance 字段没有信息量，也与真源仓 governance/lib/provenance.mjs、治理层 report-provenance.mjs 的既有口径（忽略本轮生成的 latest 报告）不一致。改为复用 governanceProvenance，一并得到 repository / dirtyPaths。11 例单测不变全绿。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:17:30-07:00"}],"HeadCommit":{"Sha1":"96dbcd998453ba6720203ea63ab612caa4d34687","Message":"chore(reports): 延后op 回绑 @ 17b81dd（clean 树，worktreeDirty=false）\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:17:30-07:00"},"CompareURL":"luoanwu/platform-governance/compare/22f145c7cc1d26ac35b6ce4781f2c36621c17de7...96dbcd998453ba6720203ea63ab612caa4d34687","Len":2}...
|
1789687054
|
Edit
Delete
|
|
30708
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"22f145c7c {"Commits":[{"Sha1":"22f145c7cc1d26ac35b6ce4781f2c36621c17de7","Message":"chore(reports): 门禁运行 / 门禁消费 / 统一审计 / 延后op 回绑 @ 06cdd6f（clean 树）\n\n并行会话的在途文件落地后工作树转干净（非报告类脏文件 0），重跑取 clean 绑定：阻断性门禁 10/10 通过（含新增 check-deferred-ops），仅记录 3 道（workspace-consistency 3 条既有 + domain-layer 13 条，均不参与总账）；延后op 8 条 0 违规。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:17:04-07:00"}],"HeadCommit":{"Sha1":"22f145c7cc1d26ac35b6ce4781f2c36621c17de7","Message":"chore(reports): 门禁运行 / 门禁消费 / 统一审计 / 延后op 回绑 @ 06cdd6f（clean 树）\n\n并行会话的在途文件落地后工作树转干净（非报告类脏文件 0），重跑取 clean 绑定：阻断性门禁 10/10 通过（含新增 check-deferred-ops），仅记录 3 道（workspace-consistency 3 条既有 + domain-layer 13 条，均不参与总账）；延后op 8 条 0 违规。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:17:04-07:00"},"CompareURL":"luoanwu/platform-governance/compare/06cdd6f0fdfa799a55c0c78b11bdcb0601bcb0e7...22f145c7cc1d26ac35b6ce4781f2c36621c17de7","Len":1}...
|
1789687026
|
Edit
Delete
|
|
30707
|
5
|
5
|
5
|
117
|
0
|
0
|
refs/heads/main
|
0
|
{"Commits":[{"Sha1":"06cdd6f0f {"Commits":[{"Sha1":"06cdd6f0fdfa799a55c0c78b11bdcb0601bcb0e7","Message":"chore(reports): 延后op 报告回绑 @ ce04278\n\n上一条提交里的报告绑的是其父提交；重跑后绑到真正包含 check-deferred-ops 的提交。8 条延后 op（已登记生效 2、已撤销 1）0 违规。\n\nworktreeDirty 仍为 true 且如实记录：工作树里有并行会话的在途文件（开发计划.md、workspace-consistency.test.mjs 与三份门禁运行产物），本会话无法在 clean 树上取证，故该报告只绑定本地工作区，不作为「clean HEAD 证据」引用。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:16:31-07:00"}],"HeadCommit":{"Sha1":"06cdd6f0fdfa799a55c0c78b11bdcb0601bcb0e7","Message":"chore(reports): 延后op 报告回绑 @ ce04278\n\n上一条提交里的报告绑的是其父提交；重跑后绑到真正包含 check-deferred-ops 的提交。8 条延后 op（已登记生效 2、已撤销 1）0 违规。\n\nworktreeDirty 仍为 true 且如实记录：工作树里有并行会话的在途文件（开发计划.md、workspace-consistency.test.mjs 与三份门禁运行产物），本会话无法在 clean 树上取证，故该报告只绑定本地工作区，不作为「clean HEAD 证据」引用。\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\n","AuthorEmail":"hillao@juhailaoluodeMacBook-Pro.local","AuthorName":"juhailaoluo pro","CommitterEmail":"hillao@juhailaoluodeMacBook-Pro.local","CommitterName":"juhailaoluo pro","Timestamp":"2026-09-17T16:16:31-07:00"},"CompareURL":"luoanwu/platform-governance/compare/ce042782a285d1d8a79e5c5a43636ea05f4868d3...06cdd6f0fdfa799a55c0c78b11bdcb0601bcb0e7","Len":1}...
|
1789686993
|
Edit
Delete
|