Show adoption evidence beside team reports - #4947
Conversation
Signed-off-by: huangruiteng <14976749+huangruiteng@users.noreply.github.com>
c7b670d to
f977e5d
Compare
huangruiteng
left a comment
There was a problem hiding this comment.
动机
Goal Chat 的已验收团队报告可能很长,原有采用/来源版本关系位于报告之后,读者不易判断当前看到的是仅已验收、已被采用,还是采用证据无法核验。本 PR 的边界是把所选操作的证据提示放到报告之前;不改变验收或采用的权威状态。
改动思路
复用现有 DelegationReadback.adoptions 与 GoalTeamLineage,只在已选报告的读取视图里派生“当前 / 部分可核验 / 无法核验 / 无记录”提示。原有明确选择、失败清空报告和刷新路径保持不变;没有新增持久状态或第二个采用判定者。
具体改动
TSX 调整提示与报告顺序,浏览器场景覆盖新提示,打包前端资源同步更新。
关键代码讲解
goal-team-results.tsx:78-89:从所选 readback 的 adoption 状态派生提示;非current不会被标成当前采用。goal-team-results.tsx:105-107:把原有 lineage 展开入口置于同一报告之前,报告本身仍由TeamArtifactReport展示。
对主干的风险
风险主要是提示与所选报告错配、或将不可用采用误写成“已采用”。代码沿用同一 selection.result;打包浏览器的正反状态通过。作为回归敏感性检查,我把新提示断言放到合并基线运行,基线如预期失败;在当前 head 上通过。构建和全部九个打包浏览器场景通过。未做真实多 Agent/Lark 验收,也未轮询远端 CI;这不构成对那些入口的完成声明。
我的整体评价
通过,限当前 head f977e5d。 这是一项比例合适、可回退的 Goal Chat 呈现改动,复用既有 typed readback,没有新增权威或兼容层;相邻边界无需再做重构。控制面/产品 UI 仍交维护者合并。
English verdict: APPROVE — the exact-head Goal Chat presentation change is scoped, regression-sensitive, and locally validated; maintainer merge only.
huangruiteng
left a comment
There was a problem hiding this comment.
Approval conclusion (author-owned PR; GitHub blocks formal self-approval)
Exact head: f977e5deb51c40dbce6b1b3bc6acb62882d28fb6
动机
Goal Chat 的已验收团队报告可能很长,原有采用/来源版本关系位于报告之后,读者不易判断当前看到的是仅已验收、已被采用,还是采用证据无法核验。本 PR 的边界是把所选操作的证据提示放到报告之前;不改变验收或采用的权威状态。
改动思路
复用现有 DelegationReadback.adoptions 与 GoalTeamLineage,只在已选报告的读取视图里派生“当前 / 部分可核验 / 无法核验 / 无记录”提示。原有明确选择、失败清空报告和刷新路径保持不变;没有新增持久状态或第二个采用判定者。
具体改动
TSX 调整提示与报告顺序,浏览器场景覆盖新提示,打包前端资源同步更新。
关键代码讲解
goal-team-results.tsx:78-89:从所选 readback 的 adoption 状态派生提示;非current不会被标成当前采用。goal-team-results.tsx:105-107:把原有 lineage 展开入口置于同一报告之前,报告本身仍由TeamArtifactReport展示。
对主干的风险
风险主要是提示与所选报告错配、或将不可用采用误写成“已采用”。代码沿用同一 selection.result;打包浏览器的正反状态通过。作为回归敏感性检查,我把新提示断言放到合并基线运行,基线如预期失败;在当前 head 上通过。构建和全部九个打包浏览器场景通过。未做真实多 Agent/Lark 验收,也未轮询远端 CI;这不构成对那些入口的完成声明。
我的整体评价
通过,限当前 head f977e5d。 这是一项比例合适、可回退的 Goal Chat 呈现改动,复用既有 typed readback,没有新增权威或兼容层;相邻边界无需再做重构。控制面/产品 UI 仍交维护者合并。
English verdict: APPROVE — the exact-head Goal Chat presentation change is scoped, regression-sensitive, and locally validated; maintainer merge only.
Problem and resulting behavior
The Goal conversation already opens accepted team artifacts, but its report reader hid requester adoption behind a generic “Acceptance and adoption” disclosure. An accepted source report looked the same whether nobody had adopted it, a downstream accepted result had used it, or that adoption evidence was unavailable.
The existing disclosure now states the observed adoption state above the readable report. A current adoption opens its versioned downstream report through the existing readback path. Missing adoption, unavailable adoption evidence, and a downstream report that references source versions retain distinct labels. This does not infer adoption from an accepted artifact or start a model while reading.
Visual and interaction evidence
Validation
npm run buildpassed on the final source and packaged chat assets.npm run smoke:personal-workspace-packagedpassed all nine scenarios on the final rebased and packaged assets, including adoption state, downstream readback, unavailable evidence, and no model restart.cqr_4135bffeb2617b54f604is valid for seven changed files; no safe-fix pass or unresolved risk.loopx canary premerge --from-git-diff --goal-id loopx-metapassed: seven selected canaries and the public-boundary check, with no failures or manual holds.The changed entry point is the Goal conversation in the packaged frontend. The Lark and CLI readback contracts are unchanged. This is one product slice: manager-originated requests still need their completed research result returned to the manager's original conversation, and real multi-agent launch/acceptance remains a separate open acceptance gap.
Future-facing review: the typed readback and existing disclosure already own this behavior; no related extraction or new state model improves this change.