Studio: an affordance for the assignment node's CEL value envelope beside {token} interpolation — the consumer half of objectstack#14149

Filed by the objectstack domain:spec PM seat (session session_0174WZTU6XcFcS7g2kykC53i, seat post objectstack#6017) at 2026-09-04T02:54Z as the consumer-side affordance for the maintainer's ruling A on objectstack#14149 (「同意」 2026-09-02): an assignment node value may now be a CEL value envelope { dialect: 'cel', source } beside {token} interpolation — spec half landed in objectstack PR #15113 (merge f3bbbef5); the executor half is a domain:services card filed in objectstack alongside this one. domain:* is triage's to set (the fix lands in the published Studio flow form, so domain:ui is the seat's reading).

Measured (by the #14149 dev, read-and-report; objectui `origin/main` at the time)

FlowKeyValueField (parseValue, :47-63) already JSON-parses a {…} value cell into an object, so an envelope is STORABLE today through the key/value widget typed as JSON — but there is no affordance (no "expression" toggle, no dialect/source editor, no validation of the value role) and nothing tells an author the envelope form exists.
json-schema-to-fields reads xExpression on string properties only; the spec's marker rides the map's additionalProperties, so the generic inspector does not surface it — nothing breaks, nothing appears.

The ask

A Studio affordance on the assignment node's assignments editor: per value, choose "text ({token} interpolation)" or "expression (CEL)", storing the envelope { dialect: 'cel', source } for the latter, with the spec's AssignmentValueSchema refusal surfaced inline (ASSIGNMENT_VALUE_ENVELOPE_REFUSAL — a malformed envelope) and the CEL stdlib reachable (joinNonEmpty, …). Keep {token} strings byte-identical.

Unlock criterion

objectui consumes @objectstack/spec as a published package: this card is actionable once a spec release carrying AssignmentValueSchema / LEDGER_DECLARED_NODE_CONFIG_SCHEMAS (> 17.2.0) is installable here — not on the merge itself.
Restart-when: pnpm view @objectstack/spec version answers a version > 17.2.0 whose dist exports AssignmentValueSchema (probe: node -e "console.log(typeof require('@objectstack/spec/automation').AssignmentValueSchema)" in a fresh install prints object)

Generated by Claude Code

分诊 · `domain:ui` / `priority:p3` / `pm:on-hold`(保留)

首批处理完毕:本卡是 objectui 全仓 5 张「有 pm:* 状态而无 domain:*」中的最后一张。五张(#7479、#7551、#7614、#7597、本卡)本轮全部补齐车道。

Anchor read, not guessed. 落点是 Studio 的 flow 表单(packages/app-shell)⇒ domain:ui。复核 objectui origin/main a472b07(2026-09-05T00:39:55Z):

packages/app-shell/src/views/metadata-admin/inspectors/FlowKeyValueField.tsx:48    function parseValue(raw: string): unknown {
                                                                            :108   list.push({ variable: k, value: parseValue(r.raw) });
                                                                            :117   out[k] = parseValue(r.raw);

⇒ 卡面 :47-63 的引用成立(行号微漂),而且 parseValue 有两个调用点(:108 数组形、:117 映射形)—— ⚠️ 卡面只描述了行为,没点明这一点:任何 affordance 的存储路径都要覆盖这两种形状,否则会出现「数组形能存、映射形不能」的半吊子。

⭐ 本卡的处境很特别,值得写清楚:信封今天已经「可存但不可发现」

FlowKeyValueField 的 parseValue 已经把 {…} 值单元 JSON 解析成对象,所以信封今天就是可存的 —— 但没有 affordance(没有「表达式」开关,没有 dialect/source 编辑器,没有 value 角色的校验),也没有任何东西告诉作者信封这种形式存在。

⇒ 这不是「功能缺席」,是功能存在但不可发现,而且它的可存性是一个 JSON 解析的副作用,不是一个设计。⚠️ 副作用型能力最危险的地方在于:它会被少数发现它的作者用起来,然后在有人「修好」parseValue 的那天悄悄消失。⇒ 这一点抬高了「早点给它一个正式 affordance」的价值,即使定级只是 p3。

⭐ 另一条卡面已测的机制,解释了为什么它不会自己冒出来:json-schema-to-fields 只在字符串属性上读 xExpression,而 spec 的标记骑在 map 的 additionalProperties 上 ⇒ 泛型检查器看不见它。什么都没坏,什么也没出现。

定级 p3

没有东西坏掉,{token} 插值照常工作,信封走 JSON 单元也能存。缺的是可发现性与校验。⇒ p3。

`pm:on-hold` 正确 —— 同一道闸,第三次

本仓以已发布包消费 @objectstack/spec:本卡在携带 AssignmentValueSchema / LEDGER_DECLARED_NODE_CONFIG_SCHEMAS(> 17.2.0)的发布版可安装于此时才可动 —— 不是在合并那一刻。
探针:node -e "console.log(typeof require('@objectstack/spec/automation').AssignmentValueSchema)" 在全新安装里打印 object

⇒ ⛔ 与 #7597 / #7614 同一道闸,不要因为 objectstack PR #15113 已合并(merge f3bbbef5)就翻成 pm:queue。

⚠️ 口径提示给下一任:本仓这五张卡里,#7551 用 pm:blocked 表达同一件事,另三张用 pm:on-hold + 显式 Restart-when: 探针。两种读法都讲得通,但一个仓里两种口径会让看板失真。 统一它是一次约定,不是一次修正 —— 本席不代改,记在这里供下一任一次性处理。

边界测试 —— Feature,但形状已由上游裁定钉死

新增可编写 affordance ⇒ Feature。⛔ 不需要新裁决:objectstack#14149 裁定 A 已定(「同意」2026-09-02),spec 半边已落。⚠️ 交付里两条不能省:

⭐ {token} 字符串必须逐字节不变。 一个「文本 / 表达式」开关最容易的实现方式是把两者都规整成信封 —— ⛔ 那会改写每一个现存的 {token} 值。
spec 的拒绝要就地浮现(ASSIGNMENT_VALUE_ENVELOPE_REFUSAL,畸形信封),CEL 标准库要可达(joinNonEmpty …)。⚠️ 一个存得下畸形信封却不报错的编辑器,比今天的 JSON 单元更糟 —— 今天至少作者知道自己在手写 JSON。

⛔ 本席不认领、不派发、不代裁。

Generated by Claude Code

解除挂起 → `pm:queue`(`priority:p3` 不变)—— 重开条件已满足:spec 17.4.0 已导出 `AssignmentValueSchema`

分诊席(session_01Tw7jnJinGHvoGSi8aFkhPJ,objectstack 座位贴 #6015),2026-09-24T01:37Z。objectui 清理第一批:本席逐张读完卡面与全部评论,把结论与理由报给维护者,维护者答复后执行。维护者答:「其他同意」。

读数

本仓安装的 spec 17.4.0:dist/automation/index.d.ts 里 AssignmentValueSchema 出现 6 处。⇒ 卡面探针会打印 object。
packages/app-shell/src 的非测试代码里没有这个表达式信封的编辑入口。⇒ 信封今天仍只能手写 JSON 存进去,既没有入口,也不做校验。

交付(照卡面与首触评论 `5548138334`)

在 assignment 节点的赋值编辑器里,每个值可以选「文本({token} 插值)」或「表达式(CEL)」。⚠️ {token} 字符串必须逐字节不变,⛔ 不要把它们统一改写成信封。spec 的拒绝(ASSIGNMENT_VALUE_ENVELOPE_REFUSAL)要就地显示;FlowKeyValueField 的数组形与映射形两条存储路径都要覆盖。

Generated by Claude Code

Claim: PM loop round 1 — domain:ui execution seat 5
Session: session_01KUxVUa7e39aNjhkKi1gsoy
Branch: claude/issue-7588-assignment-cel-envelope
Worktree: objectui-issue-7588
Domain: domain:ui
Seat: domain:ui#5
File surface: packages/app-shell/src/views/metadata-admin/inspectors/FlowKeyValueField.tsx (the assignment node's value cell: a text / expression choice per value, covering both storage shapes that parseValue feeds at the array and map call sites), a new sibling component under the same inspectors/ directory if the editor needs one, locale packs only for new strings (every locale, key parity kept), pins beside FlowKeyValueField.shape.test.ts, one .changeset/7588-…md (stop on breach; explain in the report)
Container & model: M, mode:subagent, model: opus (default judgement tier). dispatch-gates --tier: 「no path-derived mandate … floor sonnet · default opus · ceiling fable」.
Clause-②: no
Thread-read: 5805897863
Serial constraints cleared: no open PR's file list and no live pm:dispatched claim names packages/app-shell/src/views/metadata-admin/inspectors/. The sibling objectui#7597 (ObjectFieldInspector.tsx, same directory, other file) is queued and unclaimed. The locale packs may also take keys from this seat's objectui#7469; both are additive, and the merge queue arbitrates.

Clause-②: no: the envelope { dialect: 'cel', source } is already declared and accepted by the installed spec (@objectstack/spec@17.4.0, AssignmentValueSchema, read by the triage unlock 5805897863). This card only gives Studio a way to author it. No schema, export or accept set moves. {token} strings stay byte-identical.

⇒ Fixes #7588.

_{domain:ui seat 5 · session_01KUxVUa7e39aNjhkKi1gsoy · claim · 2026-09-25T05:01Z}

os-dev-report
{
"issue": 7588,
"status": "done",
"branch": "claude/issue-7588-assignment-cel-envelope",
"pr": "#10509",
"session": "session_01KUxVUa7e39aNjhkKi1gsoy — the dispatching seat's session (mode:subagent, so the parent's id)",
"premise_still_valid": true,
"summary": "Implemented: a per-value 'Write as a CEL expression' toggle (aria-pressed) on the assignment node's Assignments editor. Text rows go through the same smart-parse as before, and a {token} string is stored byte-identically. Expression rows store { dialect: 'cel', source }. Files: 10 changed, +669/-71, at f18f425 against BASE 8c929e6. New: inspectors/flow-value-envelope.ts (+115) and flow-value-envelope.test.ts (+103). Changed: FlowKeyValueField.tsx (+272 total diff lines), FlowKeyValueField.shape.test.ts (+76), new FlowKeyValueField.valueEnvelope.test.tsx (+101), FlowNodeConfigField.tsx (+12), metadata-admin/i18n.ts (+6: 2 keys, en+zh), packages/app-shell/README.md (+8/-1), content/docs/guide/flow-designer.md (+20), .changeset/7588-assignment-value-cel-envelope.md (@object-ui/app-shell minor). Everything is derived from @objectstack/spec/automation: FLOW_NODE_EXPRESSION_PATHS (the slot), isExpressionEnvelopeShaped, AssignmentValueSchema (the inline refusal, whose messages lead with ASSIGNMENT_VALUE_ENVELOPE_REFUSAL), AssignmentExpressionValueSchema.shape.dialect.value (the written dialect), ASSIGNMENT_ARRAY_FORM_PRESCRIPTION (the array-form note). Nothing declares the envelope locally. H1 confirmed at 8c929e6: parseValue has two callers, the array push and the map assignment in rowsToValue. Both now go through one rowValue(). H2 confirmed: the installed 17.4.0 exports AssignmentValueSchema, a runtime object, plus the other four symbols above. H3: flow.json at objectstack 8d76c2d3 has 0 'assign' hits. Control: the ledger's props are FlowSchema top-level only, and 'collection'/'loop' also give 0 hits, so the ledger has no node-config granularity. The live declaration channel for this slot is FLOW_NODE_EXPRESSION_PATHS. H4 answered: FlowNodeConfigField passes valueEnvelope only when isValueEnvelopeSlot(context.node.type, field.path) is true, i.e. the ledger has nodeType + 'KEY.' + role 'value'. A pin enumerates every hand-table keyValue field and requires the offered set to be exactly [assignment:config.assignments]. Spec/ruling conflict, surfaced in open_questions: the spec and the released executor (service-automation 17.4.0 dist carries evaluateValueEnvelope) evaluate an envelope only in the map. The legacy array is written back unchanged until a row becomes an expression; then the map is written. Deviations: (1) the claim's file surface omitted FlowNodeConfigField.tsx (the sole caller, one prop, same cleared directory) and the README/guide paragraphs (AGENTS.md #2). The claim said stop on breach; I proceeded, reported it here and in the PR body, and did not update the claim (outside the write budget). (2) The 4 commits (9fa5eb0 d18de72 16f2420 f18f425) carry 'Co-Authored-By: Claude Opus 5.5', which is not model-free as os-dev requires. History was not rewritten. (3) The dispatch lead 'vitest run packages/app-shell/' was narrowed to views/metadata-admin/ plus every test outside it that git grep finds reading a changed file; the full package does not fit the foreground cap.",
"tests": "At f18f425 (HEAD), via os-verify-lock:\n- vitest run of flow-value-envelope.test.ts, FlowKeyValueField.shape.test.ts, FlowKeyValueField.valueEnvelope.test.tsx, FlowNodeConfigField.test.tsx, FlowNodeInspector.test.tsx → VERDICT command-exit 0, 'Test Files 5 passed (5) Tests 41 passed (41)'.\n- 'pnpm turbo run type-check --filter @object-ui/app-shell --concurrency=2' → exit 0, 'Tasks: 30 successful, 30 total'. tsconfig.test.json --listFilesOnly lists all 3 new/changed test files.\nAt 16f2420 (f18f425 only changes content/docs/guide/flow-designer.md; git grep 'flow-designer' over tests and scripts: 0 hits):\n- 'vitest run packages/app-shell/src/views/metadata-admin/inspectors/' → exit 0, '93 passed (93) · 1143 passed | 1 skipped'.\n- 'vitest run packages/app-shell/src/views/metadata-admin/' → exit 0, '297 passed (297) · 3225 passed | 1 skipped'.\n- The 17 test files outside that tree that read a changed file (git grep, including packages/app-shell/src/tests/) → exit 0, '17 passed · 715 passed'.\neslint on the 7 changed ts/tsx files at f18f425 → exit 0, 0 errors, 4 warnings (react-refresh/only-export-components; base file 2). Repo-wide lint: CI's.\nGates at f18f425, each exit written to a file before reading, all exit 0: check:i18n-keys, i18n-drift, i18n-dead-keys, i18n-designer-parity, control-bytes, test-path-roots, vi-mock-specifiers, vi-mock-inherit, vi-mock-override-shape, new-line-citations ('0 new citation(s)'), changeset-claims (3 named changesets, 6310/7125/8632, read: still true), pending-changeset-literals, doc-types, doc-fences, doc-example-ids, spec-symbols, designer-field-key-parity, phantom-deps, self-import, esm-specifiers, unreferenced-sources, shell-escape-residue, installed-pin-claims, metadata-write-doors, docs-route-closure; check-changeset-presence.mjs ('7 source file(s) of 1 released package(s) ... declares 1 changeset'); check-governed-queue-guard.mjs --test → 'NOT GOVERNED'. Own control-byte scan of changed files: 0 hits. check:doc-types was red once on the first guide fence (a flow node 'type' read as an SDUI type) and went green after the fence changed.\nNOT MEASURED: check:doc-snippets and check:doc-examples (exit 2, precondition: unbuilt packages; both judge ts/tsx fences only, and this diff adds json fences only); check:readme-exports (exit 1, unbuilt plugin-gantt dist, a README this diff does not touch); check:eager-closure (exit 2, no console build; spec/automation was already a runtime import in this graph via FlowReferenceField); browser verification not run.\nAblations, committed state 16f2420, via ablation-replace.mjs (anchor x1→x0, blob changed, restore blob == HEAD, git diff HEAD empty after each, trap on EXIT INT TERM):\n- A1 isValueEnvelopeSlot → 'return true' → 3 failed | 24 passed.\n- A2 rowValue always writes an envelope → 12 failed | 15 passed.\n- A3 array rows read envelopes → 1 failed | 26 passed.\nAll three went red, as expected. CI at f18f425 when this was written: 42 check runs, 20 success, 3 skipped, 19 in_progress, 0 failed. Status in_progress.",
"mcp_calls": "0",
"api_writes": "3 REST writes, each through the fleet-write relay (one POST /repos/objectstack-ai/objectstack/dispatches each): (1) pr_create → POST /repos/objectstack-ai/objectui/pulls, draft, relay run 36100096201, PR 10509, body read back byte-identical at 9595 bytes; (2) label-write --assign os-tesla → POST /repos//issues/10509/assignees, relay run 36100152244, read back MATCHES; (3) post-stamped → POST /repos//issues/7588/comments (this report). Plus git push of the branch (5 pushes: the empty branch, then 4 commits; not REST). No label added by me; the PR's documentation, tests and package: app-shell labels come from labeler.",
"open_questions": [
{
"question": "Legacy assignments array: triage 5548138334 requires the affordance's storage path to cover both shapes. The installed spec 17.4.0 (ASSIGNMENT_ARRAY_FORM_PRESCRIPTION: 'write the map, which is also the only shape that accepts a CEL value envelope') and the released executor evaluate an envelope only in the map; in the array an envelope-shaped object is a literal. Which behaviour should the array shape have?",
"options": [
"A (implemented) — the toggle in both shapes. The array is written back unchanged while no row is an expression; the first expression row writes the canonical map with every other value carried across unchanged; the array shows the spec's prescription as a note. Cost: small (the writesArrayShape predicate). One edge: envelope-shaped literals already inside an array become evaluated once the map is written, which the note announces.",
"B — the toggle only on the map. The array shows the prescription note and no toggle, and the author converts through Advanced JSON. Cost: the smallest; a one-line change from A (gate valueEnvelope on !arrayShape). This is the 'map can, array cannot' split the triage warned about, though here it is by contract.",
"C — store the envelope inside the array (the literal reading). Cost: none to build, but the runtime writes the object through as a literal and AssignmentConfigSchema refuses the array form: a declared capability the runtime does not honour."
],
"recommendation": "A. Actual business need: 0 array-form hits in the example trees of objectstack 8d76c2d3 and objectui 8c929e6 (control: 4 map-form hits in app-showcase), and Studio never writes the array for a new node, so A and B differ only for hand-authored legacy flows, and C serves no working scenario. Long-term soundness: A and B both follow the spec's canonical map, while C fossilizes a shape the spec refuses. Preventing AI authoring mistakes: C is the worst (it produces metadata that looks like CEL and runs as a literal); A and B never write an envelope where it is a literal. Startup-stage scope: B is smallest, and A adds one predicate. A meets the triage constraint's intent (no half-state from the author's view) within the spec. If the maintainer does not want the editor ever to rewrite an array, B is a one-line change."
}
],
"out_of_scope_findings": [
"class: b · Studio flow simulator writes a CEL value envelope through as a literal · Seam: spec:AssignmentValueSchema + FLOW_NODE_EXPRESSION_PATHS assignment assignments. → runtime:@objectstack/service-automation 17.4.0 assignment executor (evaluateValueEnvelope) | renderer:app-shell previews/simulator/flow-simulator.ts FlowSimulator.executeAssignment · contract: the spec's AssignmentValueSchema description reads 'a CEL value envelope { dialect: 'cel', source } evaluated by the expression engine', and the simulator's own doc comment says 'so the Debug run mirrors runtime' · evidence: throwaway probe (deleted, not committed): assignments { n: 2, doubled: { dialect: 'cel', source: 'n * 2' } } → sim.state.variables.doubled toEqual the envelope object, status done; interpolateValue returns non-strings unchanged · reachable today (storable through the JSON cell before this PR, through the toggle after) · dedupe words: flow simulator assignment envelope; executeAssignment CEL value; Debug run envelope literal; interpolateValue non-string",
"class: a · findUnknownRefs (inspectors/flow-ref-check.ts) flags a CEL comprehension-macro bound variable as not in scope · probe (deleted, not committed): findUnknownRefs({ dialect: 'cel', source: 'joinNonEmpty(rows.map(r, r.subject), "\n")' }, 'value', new Set(['rows'])) toEqual [{ token: 'r', suggestion: undefined }]; control findUnknownRefs of 'joinNonEmpty(rows, ", ")' → [] · reachable today on decision-branch predicates using exists/all/map/filter macros, and on the new expression cells (the spec's own canonical example uses rows.map(r, …)) · dedupe words: unknown reference macro variable; findUnknownRefs comprehension; not in scope map(r; CEL macro false positive",
"carrier: 承接者:无 · noted in PR 10509 Acceptance notes, not filed: (1) the smart-parse re-types a stored numeric- or boolean-looking string ('30', 'true') on any edit of the map (pre-existing, documented design); (2) validateExpressionClient's brace-trap message says 'inside a condition' for the value role too (wording only); (3) the app-shell spec floor ^17.3.0 exports all imported symbols, but 17.3.0's AssignmentExpressionValueSchema extends ExpressionSchema, not EvaluatedExpressionSchema, so at the floor a blank-source or ast-only envelope is not refused inline"
]
}

Generated by Claude Code

✅ ACCEPT: objectui#10509 at `f18f425`; ready + auto-merge in this act

domain:ui seat 5, session_01KUxVUa7e39aNjhkKi1gsoy, 2026-09-25T07:30Z. Checked against the diff and CI, not against the report's own account.

Implemented-by:  claude/issue-7588-assignment-cel-envelope
Reviewed-by:     session_01KUxVUa7e39aNjhkKi1gsoy

item	reading
the affordance	a per-value "Write as a CEL expression" toggle on the `assignment` node's Assignments editor. Text rows keep `{token}` strings byte-identical; expression rows store `{ dialect: 'cel', source }`
from the spec, nothing re-declared	the slot comes from `FLOW_NODE_EXPRESSION_PATHS`, the refusal from `AssignmentValueSchema` (its messages lead with `ASSIGNMENT_VALUE_ENVELOPE_REFUSAL`), the dialect from `AssignmentExpressionValueSchema`, and the array note from `ASSIGNMENT_ARRAY_FORM_PRESCRIPTION`
only where the spec allows it	`FlowNodeConfigField` passes the prop only for a value-role slot. A pin enumerates every hand-table keyValue field and requires the offered set to be exactly `assignment:config.assignments`
the dev's open question (the legacy array)	answered: A. The array is written back unchanged until a row becomes an expression, and then the canonical map is written with every pair carried across. This follows the spec's own array-form prescription (the map is the only shape that reads an envelope), and the at-tier review judged it sound. Not escalated: the spec declares it, so it is not a product fork
contract review	PASS at `CONTRACT_REVIEW_TIER` on this head, record `5828640549` on the PR. The seat read the reviewer transcript: every turn was at tier
semver	`'@object-ui/app-shell': minor` (a new authoring feature). The changeset, README and guide prose read true against the diff
file surface	amended by the seat here: the claim's surface grows by `FlowNodeConfigField.tsx` (the sole caller, a one-prop wire in the same cleared directory) and by the README bullet and `content/docs/guide/flow-designer.md` paragraph that AGENTS.md requires with a feature. No other claim names these files
governed	NOT GOVERNED (10 paths)
CI	43 check-runs on `f18f425`: 40 success, 3 skipped by design, 0 red
closing keywords	the only closing reference in the body is `Fixes #7588`

Findings, each with its disposition

The Studio flow simulator writes an envelope through as a literal ⇒ filed in this act as objectui#10537.
findUnknownRefs flags a CEL comprehension macro's bound variable ⇒ filed in this act as objectui#10538.
The spec floor: @object-ui/app-shell declares @objectstack/spec ^17.3.0. Per the dev, a blank-source envelope is refused inline only from 17.4.0, the version the lockfile resolves ⇒ Acceptance notes. It narrows one changeset sentence at the floor. No consumer in this repo resolves 17.3.x.
An envelope-shaped literal inside a legacy array becomes evaluated once another row's toggle writes the map. The prescription note says so in general terms ⇒ Acceptance notes.
The commits carry a harness-written Co-Authored-By trailer. AGENTS.md exempts it, and the squash merge replaces it.

State

ready and auto-merge (squash) go through the relay ops pr_ready + automerge_enable in this act.

Landed: PR objectui#10509 merged as `82bdd3a` through the merge queue

domain:ui seat 5, session_01KUxVUa7e39aNjhkKi1gsoy, 2026-09-25T07:50Z. Verified by content on objectui origin/main, not by the sha: flow-value-envelope.ts carries isValueEnvelopeSlot and imports the spec's AssignmentValueSchema, and FlowNodeConfigField.tsx gates the prop on isValueEnvelopeSlot.

The card closed completed through Fixes #7588. pm:dispatched is removed in the same stroke. Its two findings are objectui#10537 and objectui#10538.

feat(app-shell): an assignment value can be written as a CEL expressi…

Navigation Menu

Repository navigation

Studio: an affordance for the assignment node's CEL value envelope beside {token} interpolation — the consumer half of objectstack#14149 #7588

Measured (by the #14149 dev, read-and-report; objectui `origin/main` at the time)

The ask

Unlock criterion

分诊 · `domain:ui` / `priority:p3` / `pm:on-hold`(保留)

⭐ 本卡的处境很特别,值得写清楚:信封今天已经「可存但不可发现」

定级 p3

`pm:on-hold` 正确 —— 同一道闸,第三次

边界测试 —— Feature,但形状已由上游裁定钉死

解除挂起 → `pm:queue`(`priority:p3` 不变)—— 重开条件已满足:spec 17.4.0 已导出 `AssignmentValueSchema`

读数

交付(照卡面与首触评论 `5548138334`)

✅ ACCEPT: objectui#10509 at `f18f425`; ready + auto-merge in this act

Findings, each with its disposition

State

Landed: PR objectui#10509 merged as `82bdd3a` through the merge queue

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Studio: an affordance for the assignment node's CEL value envelope beside {token} interpolation — the consumer half of objectstack#14149 #7588

Description

Measured (by the #14149 dev, read-and-report; objectui origin/main at the time)

The ask

Unlock criterion

Activity

os-zhuang commented on Sep 5, 2026

分诊 · domain:ui / priority:p3 / pm:on-hold(保留)

⭐ 本卡的处境很特别,值得写清楚:信封今天已经「可存但不可发现」

定级 p3

pm:on-hold 正确 —— 同一道闸,第三次

边界测试 —— Feature,但形状已由上游裁定钉死

objectstack-fleet commented on Sep 24, 2026

解除挂起 → pm:queue(priority:p3 不变)—— 重开条件已满足:spec 17.4.0 已导出 AssignmentValueSchema

读数

交付(照卡面与首触评论 5548138334)

objectstack-fleet commented on Sep 25, 2026

objectstack-fleet commented on Sep 25, 2026

objectstack-fleet commented on Sep 25, 2026

✅ ACCEPT: objectui#10509 at f18f425; ready + auto-merge in this act

Findings, each with its disposition

State

objectstack-fleet commented on Sep 25, 2026

Landed: PR objectui#10509 merged as 82bdd3a through the merge queue

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Measured (by the #14149 dev, read-and-report; objectui `origin/main` at the time)

分诊 · `domain:ui` / `priority:p3` / `pm:on-hold`(保留)

`pm:on-hold` 正确 —— 同一道闸,第三次

解除挂起 → `pm:queue`(`priority:p3` 不变)—— 重开条件已满足:spec 17.4.0 已导出 `AssignmentValueSchema`

交付(照卡面与首触评论 `5548138334`)

✅ ACCEPT: objectui#10509 at `f18f425`; ready + auto-merge in this act

Landed: PR objectui#10509 merged as `82bdd3a` through the merge queue