Repository navigation
feat(coordination): complete remote task lifecycle and opt-in gateway access (#8747) - #8765
Conversation
…eway access (#8747) Align with PR #8747 (multi-loopx-agent-cloud-collaboration-v0) and remote-authority-task-service-v0 (#8508): 1. Extend RemoteAuthorityService with full remote task settlement: - Add 'complete', 'block', 'defer', 'todo_add', and 'snapshot' operations. - Enforce hard-lease fencing and claim-request ownership checks. - Settle completion receipts and deterministic cross-machine snapshot digests. - Maintain default-deny Tenant/Goal/Actor authorization and incarnation pinning. 2. Introduce controlled remote authority gateway in Python control plane: - Default-disabled access: local private Goals never connect to remote authority or leak local credentials. - Opt-in gateway metadata projection for explicitly configured goals and the 'loopx-quality-review' canary. - Fail-closed should-run health check when remote authority is unconfigured or unreachable. 3. Conformance and unit test coverage across TypeScript and Python suites. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
loopx-agent
left a comment
There was a problem hiding this comment.
Reviewer: model_agent · gpt-6.1-sol · OpenAI · runtime_reported · reasoning_effort=max
Request changes conclusion (author-owned PR; GitHub blocks formal self-review)
评审提交 4d5bdc2e84f8f9c5d46cc87e8dedfe44b6b8343e,目标基线 45cfd0c86d8f24d630ab66ad7f76784faa21c3f6。请求修改:关闭新功能仍会中断普通准入,新旧远程客户端的读回还有两处实际回归。
动机 / Motivation
两个独立操作者需要共用一个 Goal,能从各自的客户端推进任务并读回同一结果;平常只用本地 Goal 的人仍需要正常推进。 基线的候选服务支持认领、记进度和续租;后续完成、阻塞与延期通常还要回到已有工作入口。 例如 A 认领一项任务后将它延期,B 要读到延期状态和已释放的租约,并确认这不是另一次写入。 隔离服务上验证了新的业务变更与历史回执,但当前本地准入在未启用时也失败,新快照客户端与已有认领回执校验同时退化。 本轮使用隔离 PostgreSQL 16 与两组测试身份;尚未验证生产部署、宿主真正采纳或三个切片的组合验收。
改动思路 / Approach
HTTPS 服务复用现有权限、权威存储与 Todo/租约业务 owner;quota 和心跳只是现有调用方,必须保留本地默认行为。 运行准入是权威决定。连接信息是独立的诊断观察;未验证的配置提示不能冒充已连接,也不能绕过原有等待和恢复。
TLS 服务按具名授权打开现有 PostgreSQL 实例;Todo、租约和回执继续使用原来的业务 owner。已接入的远程操作应共用现有状态和证据契约。显式启用和凭据存在是配置事实,只有经过 TLS 实际读回、验证身份与范围,才能形成已连接证据。
具体改动 / Changes
本次差异 12 个文件(+771/−42)。
关键代码讲解
- [P1] 未启用远程协作的 Goal 也无法继续。 build_live_quota_should_run_decision 在检查启用范围之前调用未导入的
registry_goal_by_id。同一个授权充足的合成 Goal,基线在 File 和 SQLite 均返回should_run=true并记录一次准入观察;候选在未配置和显式关闭时均退出 1,quota_collection_failed,没有记录准入。直接运行既有决策用例定位为NameError。用现有 registry owner 在明确配置分支判断;复验普通、显式关闭、其他 Goal 开启、开启后失败和成功路径,不要先生成本地执行义务再改一层状态字段。 - [P2] 远程认领会把缺少租约的答复当成成功。 responseCommand 保留原请求配对和认领身份校验,却丢掉了认领操作专属的租约证据检查。沿既有 TLS 负例,用合成答复去掉
result.lease。基线客户端拒绝,候选退出 0;调用方此时没有继续执行和续租所需的租约。恢复已有的认领、修改与续租各自的必填字段校验,保持当前执行证据和历史回执的区别。对同一个请求重复运行真实 CLI 正反例。 - [P2] 合法的快照在实际客户端被报作失败。 服务的快照分支 在已鉴权范围返回
status: snapshot,同一个 TLS 请求经库函数获得正确的三条任务,但 命令行包装器 只接受 found、missing 和 result,退出 1。让执行模式的成功状态随操作解码,并验证快照的目标、租约、摘要以及 64 KiB 超限时的明确失败。只读快照不应允许恢复或产生写入。 - [P2] 连接检查只看环境变量。 check_remote_authority_gateway_status 在目标端口没有监听的情况下,仅因设置了合成凭据就返回
connected。当前 quota 的导入错误挡住了更后面的分支,单独调用这个新 helper 仍可观察这个错误状态。若本阶段只提供配置提示,应直接标为未验证并保持由现有准入 owner 决定。如果它负责拒绝不可达的共享 Goal,应做有界、已认证的身份读回,并让失败进入原有的等待/恢复路径。 - 有用的新增行为。 两个独立凭据和实际 TLS 客户端下,A 可新增任务,再对自己认领的三项分别 complete、block、defer(采用已有的结构化恢复条件)。B 引用 A 的原 claim 被拒绝;B 随后能读到已完成、阻塞和延期,以及已释放的租约。重复相同请求保持原结果,recover 给出历史回执,两个独立主体对同一状态取得同一个快照摘要。当前
resume_when文档和请求解码只要求字符串,实际业务 owner 拒绝自由文本,需写明并复用已有的恢复条件约定。快照路由的返回与客户端退出仍需配齐。
| 验收项 | 本次结果 |
|---|---|
| explicit-opt-in-parity | 不满足:未配置和显式关闭都影响本地准入。 |
| credential-scope-and-lifecycle | 已核对:合成双 Actor 的租约隔离、变更和历史恢复使用同一持久状态。 |
| operation-specific-response-proof | 不满足:缺失认领租约被接受,正确快照被报错。 |
| verified-gateway-readiness | 不满足:凭据存在被当作已连接。 |
| Stage-3-adoption | 保留给后续具名部署、真正调度和双宿主采纳验证。 |
对主干的风险 / Risks
真实 PostgreSQL 16.15 和 TLS 集成加协议用例,基线 40/40,候选 41/44、无跳过;两条实际失败分别是快照退出码与缺失租约(后一条所在父测试也计为失败)。候选 TypeScript 类型检查及 5 项 Python gateway 单测通过。标准 premerge 发现配额 smoke 和维护性门禁失败,独立运行两个检查在基线均通过。默认关闭路径有确定的回归,因此候选局部的成功不能用于启用试点。
语义与 CI 对齐
本次扩展现有远程协议的命令集合。普通准入不受 opt-in 影响和相关回执要包含操作证据,是既有约束。请复验 npm run test:remote-authority、quota-plan-smoke.py 与实际 File/SQLite should-run,再检查 endpoint 失败和恢复。
我的整体评价 / Verdict
保留已有远程请求与授权、恢复 owner,先修复本地默认路径,在同一次验证中检查真实远程快照和旧命令的失败语义。核对了心跳文本、quota 和示例 CLI 的改动,没有渲染页面变更。 未验证 App 或 Lark 的安装后操作。 两组测试身份通过独立本地客户端读取同一数据库;实际多个宿主与合并后试点仍待验证。
English verdict: REQUEST_CHANGES at 4d5bdc2e84f8f9c5d46cc87e8dedfe44b6b8343e. Ordinary local admission fails with remote authority disabled. The TLS client accepts an incomplete claim and rejects a valid snapshot; the gateway reports configured credentials as a verified connection.
…mote-authority-task-lifecycle-20261114 Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…eceipts Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…y-task-lifecycle-20261114 Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
loopx-agent
left a comment
There was a problem hiding this comment.
Reviewer: model_agent · gpt-6.1-sol · OpenAI · runtime_reported · reasoning_effort=max
Request changes conclusion (author-owned PR; GitHub blocks formal self-review)
评审提交 01c4bcabbe886ca8bf987ec5e6018035c2910f5f,主干对照 6dfd1c90c4c1a186c43de0a2c05243fa14b5a76e。
[P2] 远程完成拒绝已有后继的任务。 隔离 PostgreSQL 16.15 上,同权威存储、同 Actor 的本地完成保留后继并写入 successor。真实 TLS 客户端返回 invalid_todo_completion_transaction,原任务保持 open,recover 读回 missing。重试没有追加写入。 让远程 complete 按现有 Todo 完成契约保留已登记后继,明确接收真正的终态意图。沿用现有 lease 校验及同一 operation receipt。 发生位置
动机 / Motivation
使用两个独立客户端共同推进同一 Goal 的操作者。 操作者完成一项已经有后继的普通任务,下一位操作者需要沿原后继继续工作。 主干的远程候选提供认领、记进度、续租和恢复,完整的任务收尾通过现有 Todo 完成入口处理。 任务已交付但无法从远程入口确认完成,协作者会继续看到未完成的任务,还需改用本地入口收尾。 隔离 PostgreSQL 16.15 上,同权威存储、同 Actor 的本地完成保留后继并写入 successor。真实 TLS 客户端返回 invalid_todo_completion_transaction,原任务保持 open,recover 读回 missing。重试没有追加写入。 真实 PostgreSQL 16.15 与独立 TLS 客户端已验证;实际 App、云端和三个切片组合仍待验证。
改动思路 / Approach
TLS 服务按 tenant、Goal、actor、store identity 和 operation 鉴权。它复用现有 Todo、租约和 PostgreSQL CAS,响应解码检查每种操作的证据。网关的真实 TLS 只读观测只影响所选 Goal 的等待。 保留服务、TLS 及现有完成 owner,把请求的后继语义交回同一个完成事务。 以现有远程服务的生命周期、读回和本地默认隔离为交付边界。App、Lark、真实多宿主和三片联合验证保留在原计划。
具体改动 / Changes
16 文件,新增 1015 行、删除 67 行。7 个运行时代码、8 个验证/示例和 1 个协议文档。
关键代码讲解
- RemoteAuthorityService.dispatch 把经认证的 complete 请求传给 canonical terminal lifecycle。 新增映射固定 requested_no_followup=true,linked_successor_todo_ids 为空。 隔离 PostgreSQL 16.15 上,同权威存储、同 Actor 的本地完成保留后继并写入 successor。真实 TLS 客户端返回 invalid_todo_completion_transaction,原任务保持 open,recover 读回 missing。重试没有追加写入。 源码
- decodeRemoteAuthorityResponse 将响应与原请求配对,按操作核验租约、快照或完成回执。 新版快照使用 digest、目标和行检查。 此前客户端对有效快照退出 1;当前有效快照读回可用。 源码
- check_remote_authority_gateway_status 以实际 TLS 读查验已登记目标,公开结果不带响应正文。 设置凭据不足以验证远端;目标必须匹配 actor。 有效远端读回为 verified,仍留在共享准入等待。 源码
验收依据 docs/reference/remote-authority-task-service-v0.md,版本 6dfd1c90c4c1a186c43de0a2c05243fa14b5a76e。 原规范
completion-continuation待补。隔离 PostgreSQL 16.15 上,同权威存储、同 Actor 的本地完成保留后继并写入 successor。真实 TLS 客户端返回 invalid_todo_completion_transaction,原任务保持 open,recover 读回 missing。重试没有追加写入。- 认证与回执 已验证。真实 TLS 的会话身份、恢复、版本零租约与拒绝用例。
- 快照与默认隔离 已验证。快照、范围错误、超限以及 File/SQLite 默认关闭。
- 共享宿主执行 留待后续。三片组合、宿主准入和已安装产品仍待独立资格。
对主干的风险 / Risks
同一提交已有 46 项源级及实际 PostgreSQL/TLS 用例通过且零跳过。File 与 SQLite 的默认关闭及开启观测后原轮恢复测试保留原回执,并各只扣一笔。 隔离 PostgreSQL 16.15 上,同权威存储、同 Actor 的本地完成保留后继并写入 successor。真实 TLS 客户端返回 invalid_todo_completion_transaction,原任务保持 open,recover 读回 missing。重试没有追加写入。 未配置或显式关闭的 File/SQLite 维持普通准入;已写回本地身份也可在观察开启后继续原扣额。
回归验证。在 test:remote-authority 中加入已登记后继的任务,用真实 TLS 认领后完成,再检查权威后继、租约退役、同请求重试和历史回执;以正常本地完成作为对照。
语义与 CI 对齐
扩展既有 remote authority 请求操作集合;完成、阻塞、延期和租约仍由现有业务词汇处理。 让远程 complete 按现有 Todo 完成契约保留已登记后继,明确接收真正的终态意图。沿用现有 lease 校验及同一 operation receipt。 托管 CI 未查询。
我的整体评价 / Verdict
已有后继的任务在真实客户端不能收尾,团队无法在原路径继续。 以现有远程服务的生命周期、读回和本地默认隔离为交付边界。App、Lark、真实多宿主和三片联合验证保留在原计划。 真实 PostgreSQL 16.15 与独立 TLS 客户端已验证;实际 App、云端和三个切片组合仍待验证。
English verdict: REQUEST_CHANGES at 01c4bcabbe886ca8bf987ec5e6018035c2910f5f. A real PostgreSQL/TLS completion rejects an existing successor accepted by the local owner. Preserve the successor and the original execution receipt.
…y-task-lifecycle-20261114 Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
loopx-agent
left a comment
There was a problem hiding this comment.
Reviewer: model_agent · gpt-6.1-sol · OpenAI · runtime_reported · reasoning_effort=max
Approval conclusion (author-owned PR; GitHub blocks formal self-approval)
评审提交 c52aef1561fe4125f26d8be85ca03bfe4daa49cf,目标主干 af92301ab4cc67d209e64c115eb6694e3c74ceb7。
动机 / Motivation
主干以 HTTPS 暴露认领、进度、续租、释放和原回执恢复;首个完整生命周期候选把每次远程完成都当作终态。 使用独立 TLS 客户端协作同一 Goal 的操作人。 前驱的交付已验证,已有依赖它的后继;另一客户端等待原任务完成并要认领后继。 远端错误拒绝已完成的前驱时,协作者只能改用本地入口,任务图和实际交付脱节。
隔离 PostgreSQL 16.15 上,经 TLS 完成和直接调用同一完成 owner 得到相同的 successor 语义;另一 actor 在完成前被拦,之后读到 Done、原租约已释放并能认领后继。 联合三片、两个物理宿主、已安装 App、云端与共享执行尚未在本轮验证。 保留 PG-3 #8767 对联合真实宿主的验收,并独立验证正式接入;不以本轮的本机双客户端替代。 此切片交付源级 TLS 生命周期、快照和显式只读健康观察。 本次不提升存储权威,不开启默认观察,不声明已安装 App 或共享准入。
改动思路 / Approach
复用同一 Todo 和 hard lease 事务;把认证、响应验证和只读健康观察留在适配边界。
此切片交付源级 TLS 生命周期、快照和显式只读健康观察。
普通远程完成沿用已登记的后继。历史兼容只走已认证的原请求及回执。快照和只读健康观察不会获得写权。
具体改动 / Changes
16 文件,新增 1127 行、删除 67 行。新的响应核验、远程入口适配和观测各有真实调用者;大部分新覆盖是测试与文档。
关键代码讲解
- RemoteAuthorityService.dispatch。认证后按 operation 给原事务 owner。 普通完成从强制终态改为保留当前 Todo 的后继。 错 scope 无效,历史只能读回。 源码
- decodeRemoteAuthorityResponse。客户端按原请求和具体操作验证结果。 快照检查摘要、范围和行;成功认领核对完整硬租约。 缺失证据保留 ambiguous 或 unavailable。 源码
- check_remote_authority_gateway_status。通过私有管道实现一次绑定的 TLS read。 无配置走本地,生效也只确认探测网络读。 失败或匹配时均不发起新的本地任务和扣额。 源码
依据 docs/reference/remote-authority-task-service-v0.md,规范基线 af92301ab4cc67d209e64c115eb6694e3c74ceb7。
scope-and-recovery:绑定 tenant、Goal、actor、store 和请求,历史读回无执行权。旧服务实际写入的终态完成回执,经新服务同一请求 recover/execute 保持同一 ID,无二次写;改成显式 true 的同 ID 请求和撤权后的读回被拒。canonical-lifecycle:同一事务 owner 完成与保留后继,原租约退役。隔离 PostgreSQL 16.15 上,经 TLS 完成和直接调用同一完成 owner 得到相同的 successor 语义;另一 actor 在完成前被拦,之后读到 Done、原租约已释放并能认领后继。bounded-readback:按特定操作验证收据;快照同版本同摘要,限制范围与大小。真实 TLS 快照、错误响应和超限试验通过。host-default-isolation:已配置 Goal 可验证单次只读,其他调用保持本地准入,观察不派发或记账。真实 File 和 SQLite:无配置、显式关闭(另一个 Goal 开启)保留原 heartbeat/quota 决策;旧环境变量和凭据不激活。 真实 TLS 断开及恢复仍为共享准入等待。shared-production-qualification:联合三片、两个物理宿主、已安装 App、云端与共享执行尚未在本轮验证。 保留 PG-3 #8767 对联合真实宿主的验收,并独立验证正式接入;不以本轮的本机双客户端替代。
对主干的风险 / Risks
47 项真实 PostgreSQL/TLS 与响应/策略检查全部通过、零跳过,外加两个独立实际服务探针和两项本地 CLI 默认关闭检查。 TypeScript 类型检查、准确差异的 Git diff --check 与双语 RFC 检查通过。 旧服务实际写入的终态完成回执,经新服务同一请求 recover/execute 保持同一 ID,无二次写;改成显式 true 的同 ID 请求和撤权后的读回被拒。 真实 File 和 SQLite:无配置、显式关闭(另一个 Goal 开启)保留原 heartbeat/quota 决策;旧环境变量和凭据不激活。
配置开启时的只读成功也不会调用执行或扣额;更换目标、连接失败、缺失探测 Todo 和撤权仍明确失败。托管 CI 依配置未查询。旧讨论是否可收尾需单独核对。
我的整体评价 / Verdict
此切片交付源级 TLS 生命周期、快照和显式只读健康观察。 复用同一 Todo 和 hard lease 事务;把认证、响应验证和只读健康观察留在适配边界。 两位客户端现在可从原前驱继续到后继。改动只在 TLS 例程、CLI 与其只读机器投影;没有渲染页面变化。 联合三片、两个物理宿主、已安装 App、云端与共享执行尚未在本轮验证。
English verdict: APPROVE at c52aef1561fe4125f26d8be85ca03bfe4daa49cf. Real PostgreSQL 16 and TLS clients preserved the successor, retired the original lease, and replayed old receipts without an extra write. Default-off File and SQLite remain local; hosted checks and rollout are separate.
|
This pull request has merge conflicts with Choose the remote for the base repository, not an out-of-date fork. git fetch upstream
git rebase upstream/main
# Resolve each conflict, git add the resolved files, then git rebase --continue.
git push --force-with-lease origin HEADFor a same-repository clone whose Keep the DCO |
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
loopx-agent
left a comment
There was a problem hiding this comment.
Approval conclusion (author-owned PR; GitHub blocks formal self-approval)
Reviewer: model_agent · DeepSeek · deepseek-flash · runtime_reported · high
Reviewed exact head 6e38c725a90fcd628f9445a5db6807bcfec02dcd.
动机 / Motivation
使用独立 CLI 客户端协作同一 Goal 的操作者 previously had to fall back to a local entrypoint when completion disagreed with an already recorded successor. 修改前,已有的远程生命周期只覆盖认领、进度和续租;修改后,操作者可以完成、阻塞、延期、创建和读取任务,并从持久 receipt 恢复丢失响应。The practical cost of the old path was a completed task that still looked open and a successor that nobody could claim. 已验证的生命周期操作保留同一 Todo、租约和 receipt 语义;默认关闭的 File 与 SQLite Goal 继续走原有本地准入。本 PR 不提升远端 authority 为默认 provider,也不声明已安装 App、云端或双宿主共享执行已经验收。
改动思路 / Approach
The entry point remains the authenticated remote service and the normal quota CLI. The service pins policy, actor, tenant, Goal and PostgreSQL incarnation before routing a mutation to the canonical Todo, lease and receipt owners. A successful read is projected without selecting local work. 远端生命周期和只读观测分别复用 canonical Todo/lease 与 quota 决策 owner,transport、policy 和 projection 保持在适配边界,未形成第二套状态机。当前 PR 的边界是源级 PostgreSQL/TLS 生命周期与显式只读 gateway;共享 Turn 准入、双宿主、安装态和云端留给 #8767 和后续验收。
The safest alternative is smaller only by omitting operations that independent clients need. Keeping completion local leaves the recorded successor stranded; inventing a remote Todo/lease owner would duplicate the rule that later policy changes must edit; treating configured credentials as connected would turn configuration into fabricated health evidence. The chosen extension is therefore the smallest shared boundary with a real caller and a typed failure/recovery path.
具体改动 / Changes
The specification is docs/reference/remote-authority-task-service-v0.md, judged against base revision cbdc7655482e3db8cc70c5b0c4e49451d902a360. Criterion scope-and-recovery is implemented by admission, receipt correlation and historical-only recovery in RemoteAuthorityService.dispatch. canonical-lifecycle is implemented by reusing the existing terminal, update, create and lease lifecycle owners, including successor preservation and lease retirement. bounded-snapshot is implemented by the read-only snapshot projection and digest checks. host-default-isolation is implemented by the per-Goal quota gate and verified TLS probe. remaining-qualification is deferred to the shared-authority program and #8767; this slice does not claim physical multi-host, installed App or cloud adoption.
关键代码讲解
RemoteAuthorityService.dispatchauthenticates one request, opens and pins the provider incarnation, checks prior receipt evidence, and then calls the canonical operation owner. Ordinary completion now forwardsrequested_no_followup: request.no_followup === true, so a recorded successor stays linked while an explicit terminal request remains distinct. Snapshot and read return before mutation dispatch.responseCommandmakes response correlation insufficient on its own. A claim must include the complete hard lease and matching claim identity; completion must include completion receipt evidence; renew/release must match their lease key and version. Historical replay is accepted only through the explicit recover mode and never becomes a current lease.responseSnapshotvalidates the requested Goal, store identity, revision, Todo and lease scope, deterministic digest, unique rows and absence of lease keys. The example CLI now treatsstatus: snapshotas success for execute mode, while recovery of a snapshot remains rejected.build_live_quota_should_run_decisionreads the current registered Goal once and enters the remote wait route only when that Goal hasenabled: true. Absent, disabled or sibling-Goal configuration continues through the original local builder. Hard quota pause still takes precedence.check_remote_authority_gateway_statususes the existing authenticated TLS client through a private stdin bridge. Credential presence or binding completeness only yields unverified/unavailable; a found expected Todo after matching actor and TLS identity is required for verified. The observation returns no remote note and never acquires a lease or spends quota.
对主干的风险 / Risks
The strongest regression scenario is cross-Goal activation or false connectivity: a sibling local Goal could be forced into blocked_wait, or an incomplete claim could be treated as executable. The Goal-scoped configuration lookup, policy admission, actor check, hard-lease response validation and explicit ambiguous outcome prevent those paths. File and SQLite tests compare ordinary full-speed and throttled decisions before and after another Goal is configured.
The current exact-head run passed npm run typecheck:control-plane, the quick premerge gate, the Python File/SQLite gateway regression and the remote-authority unit/TLS suite: 13 passed, 9 PostgreSQL-gated skips, no failures. This host has no disposable PostgreSQL service, so the exact merged head did not rerun the real backend sequence. The affected remote-authority production source is unchanged from c52aef1561fe4125f26d8be85ca03bfe4daa49cf, whose prior PostgreSQL 16/TLS lifecycle passed 47/47, including successor preservation, lease release, restart recovery, wrong scope and oversized replies. GitHub checks for the new head were still pending and remain a separate merge-readiness hold, not evidence of a code defect.
No blocking finding was reproduced. Residual risk is limited to unavailable exact-head real-backend execution and the deferred multi-host/installed/cloud boundary. Rollback is disabling the Goal binding or stopping/revoking the service; exact-request recover retains durable receipts.
我的整体评价 / Verdict
This is a coherent, opt-in extension of the existing remote authority boundary. It improves long-horizon continuation because the recorded successor becomes claimable after canonical completion, and improves the CLI journey because valid snapshots and operation-specific recovery outcomes are truthful. No browser surface changes, so visual review is not applicable. The architecture assessment remains retain: the adapter owns transport and bounded projection while canonical lifecycle and quota owners retain their rules. Default-off isolation is verified for both File and SQLite, including a normal should_run=true Goal with another Goal enabled. The remaining risk is the exact-head backend environment gap noted above; no approval blocker remains. Re-review should restart if the head changes or the affected remote-authority source diverges from the inspected revision.
English verdict: APPROVE at 6e38c725a90fcd628f9445a5db6807bcfec02dcd; canonical lifecycle and default-off behavior are supported by current local/TLS fixture evidence and unchanged prior PostgreSQL/TLS qualification.
huangruiteng
left a comment
There was a problem hiding this comment.
Approval conclusion (author-owned PR; GitHub blocks formal self-approval)
Reviewer: model_agent · DeepSeek · deepseek-flash · runtime_reported · high
Reviewed exact head 6e38c725a90fcd628f9445a5db6807bcfec02dcd.
动机 / Motivation
使用独立 CLI 客户端协作同一 Goal 的操作者 previously had to fall back to a local entrypoint when completion disagreed with an already recorded successor. 修改前,已有的远程生命周期只覆盖认领、进度和续租;修改后,操作者可以完成、阻塞、延期、创建和读取任务,并从持久 receipt 恢复丢失响应。The practical cost of the old path was a completed task that still looked open and a successor that nobody could claim. 已验证的生命周期操作保留同一 Todo、租约和 receipt 语义;默认关闭的 File 与 SQLite Goal 继续走原有本地准入。本 PR 不提升远端 authority 为默认 provider,也不声明已安装 App、云端或双宿主共享执行已经验收。
改动思路 / Approach
The entry point remains the authenticated remote service and the normal quota CLI. The service pins policy, actor, tenant, Goal and PostgreSQL incarnation before routing a mutation to the canonical Todo, lease and receipt owners. A successful read is projected without selecting local work. 远端生命周期和只读观测分别复用 canonical Todo/lease 与 quota 决策 owner,transport、policy 和 projection 保持在适配边界,未形成第二套状态机。当前 PR 的边界是源级 PostgreSQL/TLS 生命周期与显式只读 gateway;共享 Turn 准入、双宿主、安装态和云端留给 #8767 和后续验收。
The safest alternative is smaller only by omitting operations that independent clients need. Keeping completion local leaves the recorded successor stranded; inventing a remote Todo/lease owner would duplicate the rule that later policy changes must edit; treating configured credentials as connected would turn configuration into fabricated health evidence. The chosen extension is therefore the smallest shared boundary with a real caller and a typed failure/recovery path.
具体改动 / Changes
The specification is docs/reference/remote-authority-task-service-v0.md, judged against base revision cbdc7655482e3db8cc70c5b0c4e49451d902a360. Criterion scope-and-recovery is implemented by admission, receipt correlation and historical-only recovery in RemoteAuthorityService.dispatch. canonical-lifecycle is implemented by reusing the existing terminal, update, create and lease lifecycle owners, including successor preservation and lease retirement. bounded-snapshot is implemented by the read-only snapshot projection and digest checks. host-default-isolation is implemented by the per-Goal quota gate and verified TLS probe. remaining-qualification is deferred to the shared-authority program and #8767; this slice does not claim physical multi-host, installed App or cloud adoption.
关键代码讲解
RemoteAuthorityService.dispatchauthenticates one request, opens and pins the provider incarnation, checks prior receipt evidence, and then calls the canonical operation owner. Ordinary completion now forwardsrequested_no_followup: request.no_followup === true, so a recorded successor stays linked while an explicit terminal request remains distinct. Snapshot and read return before mutation dispatch.responseCommandmakes response correlation insufficient on its own. A claim must include the complete hard lease and matching claim identity; completion must include completion receipt evidence; renew/release must match their lease key and version. Historical replay is accepted only through the explicit recover mode and never becomes a current lease.responseSnapshotvalidates the requested Goal, store identity, revision, Todo and lease scope, deterministic digest, unique rows and absence of lease keys. The example CLI now treatsstatus: snapshotas success for execute mode, while recovery of a snapshot remains rejected.build_live_quota_should_run_decisionreads the current registered Goal once and enters the remote wait route only when that Goal hasenabled: true. Absent, disabled or sibling-Goal configuration continues through the original local builder. Hard quota pause still takes precedence.check_remote_authority_gateway_statususes the existing authenticated TLS client through a private stdin bridge. Credential presence or binding completeness only yields unverified/unavailable; a found expected Todo after matching actor and TLS identity is required for verified. The observation returns no remote note and never acquires a lease or spends quota.
对主干的风险 / Risks
The strongest regression scenario is cross-Goal activation or false connectivity: a sibling local Goal could be forced into blocked_wait, or an incomplete claim could be treated as executable. The Goal-scoped configuration lookup, policy admission, actor check, hard-lease response validation and explicit ambiguous outcome prevent those paths. File and SQLite tests compare ordinary full-speed and throttled decisions before and after another Goal is configured.
The current exact-head run passed npm run typecheck:control-plane, the quick premerge gate, the Python File/SQLite gateway regression and the remote-authority unit/TLS suite: 13 passed, 9 PostgreSQL-gated skips, no failures. This host has no disposable PostgreSQL service, so the exact merged head did not rerun the real backend sequence. The affected remote-authority production source is unchanged from c52aef1561fe4125f26d8be85ca03bfe4daa49cf, whose prior PostgreSQL 16/TLS lifecycle passed 47/47, including successor preservation, lease release, restart recovery, wrong scope and oversized replies. GitHub checks for the new head were still pending and remain a separate merge-readiness hold, not evidence of a code defect.
No blocking finding was reproduced. Residual risk is limited to unavailable exact-head real-backend execution and the deferred multi-host/installed/cloud boundary. Rollback is disabling the Goal binding or stopping/revoking the service; exact-request recover retains durable receipts.
我的整体评价 / Verdict
This is a coherent, opt-in extension of the existing remote authority boundary. It improves long-horizon continuation because the recorded successor becomes claimable after canonical completion, and improves the CLI journey because valid snapshots and operation-specific recovery outcomes are truthful. No browser surface changes, so visual review is not applicable. The architecture assessment remains retain: the adapter owns transport and bounded projection while canonical lifecycle and quota owners retain their rules. Default-off isolation is verified for both File and SQLite, including a normal should_run=true Goal with another Goal enabled. The remaining risk is the exact-head backend environment gap noted above; no approval blocker remains. Re-review should restart if the head changes or the affected remote-authority source diverges from the inspected revision.
English verdict: APPROVE at 6e38c725a90fcd628f9445a5db6807bcfec02dcd; canonical lifecycle and default-off behavior are supported by current local/TLS fixture evidence and unchanged prior PostgreSQL/TLS qualification.
…000 commits (#8759) * Deliver verified images in private Lark answers Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Keep image retries behind new private Lark replies Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recover missing images from one verified private Lark reply Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Use one wall clock for private Lark image retries Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recheck the original Lark reply before historical image writes Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Verify private Lark image sources before every redirect Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * feat(coordination): add postgresql authority checkpoints and local-to-pg-seed exporter - Add authority_checkpoints table DDL and tenant-scoped RLS policies. - Store incremental delta in authority_commits and bounded checkpoints every 64 commits and on genesis cursor 1, resolving table bloat and the 20,000 commit recovery ceiling. - Implement scanCommitted bounded delta replay from covering checkpoint via AuthorityStateReplay. - Add setupPostgreSqlAuthoritySchemaAndRole for explicit schema installation and least-privilege restricted role assembly. - Add exportLocalToPgGenesisSeed with sanitization (expired leases, local absolute paths, .local/ references, Feishu IDs, private materials) and fail-closed privacy boundary scan. - Add loopx postgres setup and loopx postgres local-to-pg-seed CLI commands. - Add test suites for checkpoints, sanitization, role assembly, and CLI invocation. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recheck the historical Lark reply before each media write Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Retry Lark image uploads after an explicit App scope rejection Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Honor original Lark stop and closure before deferred images Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recheck private answer control after Lark source read Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Read the current Core session after delayed image controls Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Fence private answer media dispatch with original controls Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Keep Lark media checks with private delivery Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recheck Lark image sources after recording upload intent Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Read both original Feishu messages together before media upload Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Verify Feishu sources at the guarded image dispatch Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Preserve pre-dispatch Feishu source and session validation Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recover and verify PostgreSQL checkpoint history with append-only privileges Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Verify backup history to its captured cursor beyond 20000 commits Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Generate explicitly synthetic secret fixtures at test time Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recover private image uploads when the Lark CLI cannot start Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(chat): preserve the first Goal's registered conversation (#8798) * fix(chat): preserve the first Goal's registered conversation Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * test(chat): keep original Goal adoption compatible with Python 3.11 Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> --------- Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Avoid rereading delivered Lark settings records and empty queues Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recover private Lark images after a CLI descriptor limit Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Enable existing checks and the original authorized Task in managed Chat Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Reuse the canonical boundary projection and narrow Task JSON types Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Keep confirmed Goal creation attached to its original Task (#8805) Reuse the existing C1 source observation guard Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Assert the lease source probe returned an observation (#8796) (cherry picked from commit 2c099be) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recover an owed replan debit before the next host Turn (#8794) * Recover an owed replan debit before the next host Turn Consider the original Todo binding when an admitted replan has verified writeback. Require the exact spend receipt before another guard grants delivery, even if the Todo lifecycle has changed. Keep explicit selection on the same recovery path and let the existing settlement owner replay the original debit. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Preserve ordinary completed-Todo successor admission Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Replay one validated Turn debit across concurrent observations Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Validate original Turn spend terms before debit and replay Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Keep conversation history rendering off the typing path (#8781) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * feat(coordination): complete remote task lifecycle and opt-in gateway access (#8747) (#8765) * feat(coordination): add complete remote task lifecycle and opt-in gateway access (#8747) Align with PR #8747 (multi-loopx-agent-cloud-collaboration-v0) and remote-authority-task-service-v0 (#8508): 1. Extend RemoteAuthorityService with full remote task settlement: - Add 'complete', 'block', 'defer', 'todo_add', and 'snapshot' operations. - Enforce hard-lease fencing and claim-request ownership checks. - Settle completion receipts and deterministic cross-machine snapshot digests. - Maintain default-deny Tenant/Goal/Actor authorization and incarnation pinning. 2. Introduce controlled remote authority gateway in Python control plane: - Default-disabled access: local private Goals never connect to remote authority or leak local credentials. - Opt-in gateway metadata projection for explicitly configured goals and the 'loopx-quality-review' canary. - Fail-closed should-run health check when remote authority is unconfigured or unreachable. 3. Conformance and unit test coverage across TypeScript and Python suites. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * test(coordination): assert subprocess readback before decoding Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(coordination): qualify explicit remote task reads and lifecycle receipts Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(coordination): preserve existing successors on remote completion Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * test(remote-authority): prove ordinary admission stays local Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * test(chat): preserve create-only intent in workspace selection smoke Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Defer material inventory during configuration discovery (#8807) Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Defer report providers during Turn reads Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * test(core): require installed provenance before qualification Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Reuse the final source check for staged Turn requests (#8820) Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Defer PR wait provider loading until permission is checked (#8821) Load the optional PR reconciler only after the invocation-time observation permission check, inside the existing provider failure boundary. Preserve original permission, cadence, result and unknown-effect owners, and document the bounded default-closure and timing-window change. Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Load provider options only for the selected capability editor (#8822) Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * docs: describe README session continuity directly (#8823) * Skip automatic usage observers for supported opt-outs (#8824) * Skip automatic usage observers for supported opt-outs Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Record installed Core and C1 observer qualification Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> --------- Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * docs(blog): add EdgeBench controlled Context ON vs OFF 10-task score comparison (#8760) * docs(blog): add EdgeBench controlled Context ON vs OFF 10-task score comparison Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * docs(blog): use declarative mechanism-led opening in zh context-delivery score section Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(heartbeat): preserve proven waits before their due time (#8825) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Fix manager discovery in ordinary installs (#8830) * Package the manager discovery projection Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * Record selected manager discovery distribution qualification Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> --------- Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * test(workspace): verify confirmed Goal creation against the current contract (#8816) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Fix Windows CLI output probe reader invocation (#8817) Port the probe fixture and native Windows regression coverage from e6eb92d and 8d200aa. Decode the generated PowerShell literals into argv before launching the pinned reader. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * refactor(issue_fix): fold reviewer source trust/freshness vocabularies into one owner (#8826) reviewer_recommendation.py restated the trust and freshness sets that repository_context.py already owns. Import both as aliases of the owner objects so the decision has a single source, and add an architecture guard that pins single ownership by value, by identity, and by wiring through both public builders. REVIEWER_SOURCE_KINDS stays local (deliberately narrower). Signed-off-by: psychologister <230141834+psychologister@users.noreply.github.com> Co-authored-by: psychologister <230141834+psychologister@users.noreply.github.com> * fix(scheduler): reject superseded heartbeat receipts on the Python follow-up route (#8708) * fix(scheduler): reject superseded heartbeat receipts on the Python follow-up route The Python compatibility route for scheduler-ack-current/scheduler-fail-current checked only that the receipt for --turn-instance-id exists; a receipt already superseded by a newer heartbeat Turn still committed scheduler state, while the native TypeScript transaction rejects it as heartbeat_receipt_stale. Thread the Turn from build_scheduler_followup_payload through record_quota_scheduler_ack / record_quota_scheduler_ack_for_decision / record_quota_scheduler_failure_for_decision into _run_native_followup, which now sends turn_instance_id with require_heartbeat_receipt=true. The TypeScript transaction remains the single owner of the missing/stale/superseded fence (it also rechecks under the scheduler state lock); optional parameters keep the in-process API, which has no Turn, unchanged. Fixes #8605 Signed-off-by: Lihua <1017343802@qq.com> * Keep scheduler follow-up hint reads out of legacy migration Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: Lihua <1017343802@qq.com> Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> Co-authored-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Resume the original quota settlement after lifecycle writeback (#8728) * fix(replan): resume settlement after lifecycle review writeback Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Recover committed replan settlement before rebuilding its review Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(antigravity): keep the project registry in generated wakes (#8768) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(chat): keep workspace checker reads portable (#8814) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Reject external Turn retries when the captured work changes (#8828) * Recheck captured work before launching an external Turn host Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Keep complete-source metadata assembly outside Turn planning Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Preserve PostgreSQL runtime access and recover seed output retries Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Validate recovery cursor metadata without a fixed history ceiling Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Signed-off-by: psychologister <230141834+psychologister@users.noreply.github.com> Signed-off-by: Lihua <1017343802@qq.com> Co-authored-by: Lusen Dong <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> Co-authored-by: Lusen Dong <cicomaple@gmail.com> Co-authored-by: psychologister <st040325@gmail.com> Co-authored-by: psychologister <230141834+psychologister@users.noreply.github.com> Co-authored-by: Lihua <1017343802@qq.com>
Goal And Delivered Outcome
Extend the remote task service so a client can complete, block, defer, create and inspect tasks using the same canonical owners as local coordination. This revision addresses the initial review findings and the follow-up about completing work with a recorded successor.
no_followup: trueunless the Todo already records that intent. Exact retries of accepted older commands replay their original receipt.main.Author Declaration
The wire operation list extends the existing remote service vocabulary. It does not add a second Todo or lease state machine.
Scope And Continuation
The observation is reversible per Goal. The example CLI can exercise the remote lifecycle; the ordinary scheduler does not promote authority or admit remote execution. The independent review of the current commit concluded APPROVE. At the readiness readback five review discussions were still open. PG-1 #8759 and this revision combine cleanly on main at
b46a76d2d64b3c14877bdf288953c56ae1c7793b(combined tree6a46c2d1f240845df22b176e03baf58d4d4090f6); the real backend qualification below covers the earlier tested combination. Full PG-3 qualification remains in #8767.Validation
c52aef1561fe4125f26d8be85ca03bfe4daa49cfonmainmerge baseaf92301ab4cc67d209e64c115eb6694e3c74ceb7.staticb46a76d2d64b3c14877bdf288953c56ae1c7793bcompleted with all 19 selected checks and all direct checks passing.regression_parityinvalid_todo_completion_transactionon a successor already accepted by the local completion owner. On this revision both paths preserve the successor. 18 focused Python checks also passed.real_entrypointrecoveryintegrationa5e2ca33bea91dd61ebd1491a4b0e30c9f3a0ce1with PG-1 #8759 atececfb4f9d054d75c67ef4bab150c508424aa426has 398/398 store, 11/11 service and 47/47 remote-authority checks, with zero skips.real_backendThe earlier real PostgreSQL run and kernel checks cover the previous revision. The PG-2 qualification used independent clients on one host. Applying all of PG-3 #8767 at
e04cd69cefto the updated PG-1/2 combination carries older PG-1/2 changes and produces 14 conflicts. Applying PG-3's own nine-file change gives a clean isolated preview. The public rehearsal in that preview passed 8/8 in simulation and 5/8 against isolated PostgreSQL 16.15; database readback confirmed the lease remained active with a 45-minute TTL after the script waited about two seconds, and the rehearsal connected with the bootstrap role. PG-3 still needs the public path to prove authenticated TLS clients, actual expiration and takeover with a restricted runtime role. Physical hosts, installed App adoption and cloud deployment have separate acceptance.Frontend / Visual Evidence
Technical Direction