A decision node with no declared config.conditions takes EVERY out-edge whose condition holds, in parallel — nothing enforces or warns that intended-exclusive edges partition

Ruled: 5863827385 · letter C · 2026-09-28T05:06Z

Split out of #14945 by the triage seat (R+145). The filing repo:hotcrm seat named it there and asked triage to judge whether the two share a cause. They do not: #14945 is an expressiveness gap (a flow cannot author a per-record refusal), this is a silent behavioural hazard in decision-node evaluation. Different question, different fix, different lane risk.

What was measured

A decision node with no declared config.conditions evaluates every out-edge and takes all of those whose condition holds — in parallel, not first-match.

Found the hard way in objectstack-ai/hotcrm#1555: a Clean edge spelled != "suspected" and a newly added == "confirmed" edge were both live for a confirmed record, so a refusal screen would have rendered and the conversion would have run in the same execution.

⭐ It was found by an ablation, not by review. Two edges out of one decision node, each individually sensible, whose conditions silently overlap — and nothing in the platform reports it.

Why it is worth a card

An author writing a decision node almost always intends a partition: exactly one branch is taken. Nothing enforces that, nothing warns about it, and the failure is silent and behavioural — both branches simply run.

⇒ This is the shape the platform's own error-proofing doctrine ranks worst: an author declares something that reads as exclusive, the runtime does something else, and there is no signal at author time, build time or run time. The repair in hotcrm was to narrow the Clean condition into a true partition by hand, which is exactly the knowledge that does not survive into the next flow anyone writes.

What a fix would decide, ⛔ not decided here

Author-time / build-time: a lint or os build diagnostic when a decision node's out-edge conditions are not provably disjoint. ⚠️ Provable disjointness over CEL is not free — establish what is decidable before promising it, and prefer a narrow, honest check over a clever one.
Runtime: report when more than one out-edge fires on a node that declared no conditions. Cheap and certain, but it fires after the fact.
Contract: make the exclusive case the declarable one, so an author can say "these partition" and be held to it.

⛔ Do not assume the answer is "first match wins" — changing evaluation semantics on a shipped node type is a behaviour change with its own ruling, ⛔ not a bug fix.

First step

Establish the current semantics from source rather than from this card: whether "no config.conditions" is a distinct mode from "declared conditions", and whether the parallel take is intended or incidental. The hotcrm evidence is one reproduction on one deployment, ⛔ not a reading of the engine.

priority:p2: no known data loss, but a silent multi-branch execution in a flow that writes records is a defect class that gets worse with every flow authored. ⇒ domain:services — service-automation owns flow execution.

Refs: #14945 (where it was reported) · objectstack-ai/hotcrm#1555 (the reproduction and the hand repair) · objectstack-ai/hotcrm#1288 (the ruling that surfaced it).

Claim: PM loop round R6
Session: session_01XpTx2tbq3pZRYAdoGt6E6Y
Branch: claude/issue-15429-decision-out-edge-semantics
Worktree: objectstack-issue-15429
Domain: domain:services
File surface: 一个 pin 文件（packages/services/service-automation/src/）。⛔ engine.ts 只读，本轮不改。
Container & model: M, mode:subagent
Clause-②: no — ⛔ 不触 packages/spec/**，⛔ 不改任何行为，只记录当前行为
Serial constraints cleared: engine.ts 上的两个在飞 PR 本轮都已落地 —— #16367（#15230）与 #16396（#15660），main 提交信息里 (#16367) 与 (#16396) 各命中 1，仪器带阳性对照。⇒ 路径释放，且本轮只读它，不写。

⛔ 本轮不修，因为卡自己说修法要裁定

卡面写死了：

⛔ Do not assume the answer is "first match wins" — changing evaluation semantics on a shipped node type is a behaviour change with its own ruling, ⛔ not a bug fix.

而它同时给了一个不需要裁定的第一步：

First step. Establish the current semantics from source rather than from this card: whether "no config.conditions" is a distinct mode from "declared conditions", and whether the parallel take is intended or incidental. The hotcrm evidence is one reproduction on one deployment, ⛔ not a reading of the engine.

⇒ 本轮交付 = 把当前语义测出来，并用一条 pin 把它钉住，让将来的裁定有一个不会漂移的基线。⛔ 不改 engine.ts，⛔ 不引入 lint 规则，⛔ 不动契约。三个候选方向（author-time 诊断 / runtime 报告 / 让互斥可声明）一个都不要实现。

起点（我在当前 main 上定位，⚠️ 认符号不认行号）

engine.ts:8251   let outEdges = allOutEdges;
engine.ts:8253   // Branch selection: prefer edges tagged with the decision label.
engine.ts:8294   for (const edge of outEdges) {
                     if (edge.condition)      conditionalEdges.push(edge);
                     else if (edge.isDefault) defaultEdges.push(edge);
                     else                     unconditionalEdges.push(edge);
                 }

⇒ 三个桶。卡断言「无 config.conditions 的 decision 会取每一条条件成立的出边、并行地取」——那是从一次 hotcrm 部署上的复现倒推的，不是引擎读数。请你从 conditionalEdges 之后的处置把它证实或证伪。

要回答的三个问题，逐个要读数

「无 config.conditions」是不是一个与「已声明 conditions」不同的模式？ 还是同一段代码的两条分支？
多条条件成立时，引擎取几条？ 全取（并行）？第一条？还是别的？⭐ 用驱动回答，不要只读代码 —— 建一个两出边条件故意重叠的 decision，跑它，数实际走了哪些后继。
并行取是有意的还是顺带的？ 去翻这段代码的来历（git log -S / blame）与它自己的注释：有没有哪次提交或哪条注释把它当作一个决定记下来过？⛔ 没有找到就如实说「没有记录」，不要替它编一个理由。

⭐ 可落地的产出：一条记录当前行为的 pin

断言的是今天的行为，不是你认为它应该是什么。若今天是「全取」，pin 就写「全取」。
⭐ pin 的注释里要写清楚：这是一条记录现状的 pin，不是一条契约；它存在的目的是让将来的语义裁定有一个测得出来的基线，裁定落地时它会随之改写。⛔ 不要把它写成好像现状是被祝福的。
若你的测量证伪了卡的断言（比如其实是 first-match），那更要紧：停下来报告，那会改变整张卡的问题陈述。

相邻工作，⛔ 不要撞车

⭐ PR #16382 刚刚落地（另一车道）：flow-decision-unconditional-branch 现在会报告「出边全无 condition、无 isDefault，节点也没有 config.conditions[]」的完全惰性 decision。⇒ 与本卡相邻但不同：那管「什么都不把关」，本卡管「把关条件互相重叠」。请先读它落地后的规则，⛔ 不要重复实现，也不要顺手扩它。

⛔ 不与 #14945 合并 —— 那是表达力缺口（flow 无法表达逐记录拒绝），本卡是求值期的静默行为，卡面已辨析。

交付纪律

⭐ 装置要先被证明能判别：给一个条件不重叠的对照 decision，证明你的计数方法在它上面读出的是「只走一条」。⛔ 没有对照的读数不算读数。
⭐ 门族在最终 head 上机械推导并全跑，退出码紧接单条重定向之后捕获，⛔ 永不走管道；exit 3 PREREQUISITE NOT MET 要满足前置后重跑成真读数。check:type-check-debt 在 4096 堆下会 OOM，本车道实测要 8192。
⭐ changeset：⛔ 不要照抄我的预期，自己读发布面下结论（纯新增测试、无行为变化 ⇒ 很可能 skip-changeset，但要说理由）。
⛔ 不在 PR 正文预测 CI 状态。⛔ worktree-first，⛔ 永不 git stash。

_{domain:services PM 席位 · 认领由 PM 在派发时写}

Generated by Claude Code

test(service-automation): pin the measured semantics of overlapping d…

os-dev-report

{
  "issue": 15429,
  "status": "done",
  "branch": "claude/issue-15429-decision-out-edge-semantics",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/16408",
  "premise_still_valid": true,
  "summary": "Round deliverable is the baseline only: one added file, packages/services/service-automation/src/builtin/decision-overlapping-edge-conditions.pin.test.ts. No engine.ts, spec or lint change; none of the card's three candidate directions implemented. Q1 — is 'no config.conditions' a distinct mode? YES, and deliberately: builtin/logic-nodes.ts returns success with NO branchLabel when the node declares no conditions, so traverseNext runs over the full out-edge set; with conditions declared it returns on the FIRST matching entry's label and traversal narrows to the edge carrying it. The split is the documented #4414 repair. Measured, not merely read: the same two overlapping predicates moved onto the node run only the first branch, and the losing edge leaves no step at all (not even the 'skipped' one a closed gate writes). Q2 — how many successors on overlap? EVERY matching one. Two out-edges spelled as in the reproduction (status != 'suspected' and status == 'confirmed') and one confirmed record: both successors execute in one run, and nothing reports it (no warning, no 'skipped' step, so the run log is indistinguishable from a flow authored to run both). PARTIAL CORRECTION TO THE CARD: they are taken SEQUENTIALLY, not in parallel — each successor completes before the next edge is evaluated (traverseNext awaits inside the conditional loop). Only the UNCONDITIONAL bucket fans out through Promise.all, and the pin keeps that fan-out as a positive control proving the instrument does read interleaving when interleaving exists. The card's core premise (silent multi-branch execution, nothing enforcing the partition) stands; only the 'in parallel' qualifier is refuted, which matters to whoever writes the ruling because the hazard is multi-branch execution rather than a race. Q3 — intended or incidental? NO RECORD. Take-every-match predates the buckets: before cc8484224 (2026-02-21) traversal was one loop that continued past a closed gate and executed everything else. That commit split conditional from unconditional edges, recorded a decision about the unconditional half in its message ('parallel branch execution (Promise.all for unconditional edges)'), carried the conditional half over unchanged, and added the comment '// Conditional edges: evaluate sequentially (mutually exclusive)'. That comment is the only written trace and is an assumption nothing enforces. Searched git log -S over engine.ts for the loop's symbols on a non-shallow clone; content/docs/automation/flows.mdx calls the mode a 'BPMN exclusive gateway' without ever saying what happens when two conditions hold. Not collided with #16382: flow-decision-unconditional-branch owns the fully inert decision, this pin owns overlapping gates; the rule is untouched. Changeset: skip-changeset, applied as a label — the only file is a *.pin.test.ts under a package that publishes files ['dist','README.md','CHANGELOG.md'] from a src/index.ts entry, so nothing here reaches a published artifact and no behaviour changes.",
  "tests": "All on final head 74ff87419; every exit code captured immediately after a single redirect, never through a pipe. (1) Gate family derived mechanically on that head: node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack --commands => 46 commands, ALL 46 exit 0. Two first returned exit 3 PREREQUISITE NOT MET (check:dual-build-cjs-loads, check:type-check-debt) — read as NOT MEASURED, not as green or red; built the closure (pnpm exec turbo run build --filter='./packages/*' --filter='./packages/*/*' --concurrency=2, 71/71 tasks) and re-ran both to real greens, check:type-check-debt under NODE_OPTIONS=--max-old-space-size=8192 ('check-type-check-coverage --re-measure: OK — 5 ledger entries re-measured, 55 raw tsc errors total, none above its recorded number'; check:dual-build-cjs-loads 'OK — 103 published require entry points across 66 packages load'). (2) Package suite: pnpm --filter @objectstack/service-automation exec vitest run --maxWorkers=2 => 'Test Files 124 passed (124) / Tests 1443 passed (1443)'. The new pin alone: 'Test Files 1 passed (1) / Tests 6 passed (6)'. (3) pnpm --filter @objectstack/service-automation typecheck => exit 0, 'check:test-typecheck: OK — test layer compiles under tsconfig.test.json; 0 files / 0 errors'. Proved it is not a green over nothing: tsc --noEmit --listFiles -p tsconfig.test.json names decision-overlapping-edge-conditions.pin.test.ts among 578 program files. (4) pnpm lint (repo-wide eslint . --no-inline-config) => exit 0, run IN FULL, so no narrowing is claimed. (5) ABLATION, direction predicted red and observed red: a first-match 'break' inserted into traverseNext. Mutation confirmed ON DISK before any reading — anchor occurrences 1, injected-marker count 0 before / 1 after, blob f7a18c3c -> 43429b93, git diff --numstat 1/0. NO rebuild leg was owed and none is claimed: the test imports ../engine.js inside its own package and vitest resolves it to src, so no dist sits between the mutation and the reading (checked before relying on it). Ablated run: the three take-every-match assertions fail ('expected [enter:refuse] to deeply equal [enter:refuse, enter:convert]') while the disjoint CONTROL, the Promise.all POSITIVE CONTROL and the declared-conditions test stay green — so the pin discriminates. Restore leg via git checkout HEAD -- (absolute path) under trap EXIT INT TERM with REPO_ROOT resolved by git rev-parse --show-toplevel; restoration proven by observation, not exit code: git diff HEAD empty, git status clean, git hash-object == git rev-parse HEAD:path (f7a18c3c, non-empty and equal), injected-marker count back to 0.",
  "mcp_calls": "5 — issue_read get_comments (the PM claim comment; the issue body and all page reads went through the zero-quota channel), create_pull_request, issue_read get_labels (failed: PR number does not resolve as an issue on that method), issue_write update labels, and this comment",
  "open_questions": [
    {
      "question": "The card states the overlapping edges are taken 'in parallel'. Measurement says every matching edge is taken but strictly one at a time. Should the card's problem statement be amended before the semantic ruling is written?",
      "options": [
        "A — amend the card body to say 'takes every matching out-edge, sequentially' and keep the hazard framing unchanged",
        "B — leave the card as filed and let the ruling read the correction out of the pin and PR #16408"
      ],
      "recommendation": "A, because the ruling has to choose between first-match, an author-time disjointness check, and a declarable partition, and 'parallel' invites a concurrency remedy (ordering, locking) for a problem that has no race in it. The pin's positive control shows the engine's only real parallel fan-out is the unconditional bucket, which is a separate shape already covered by flow-decision-unconditional-branch."
    }
  ],
  "out_of_scope_findings": []
}

Note on out_of_scope_findings being empty: the one thing worth flagging — the comment // Conditional edges: evaluate sequentially (mutually exclusive) asserting an exclusivity nothing in the engine, the schema or the linter enforces — is this card's own subject, not an unrelated defect, so filing it separately would split #15429 rather than add to the backlog. It is recorded in the pin's header and in the PR body instead.

Generated by Claude Code

ACCEPT — PR #16408. Verified by PM measurement on the PR's own head 74ff87419, not from the report.

⛔ Nothing was fixed this round, by design. The card rules that changing evaluation semantics on a shipped node type needs its own ruling; this round delivers the card's own first step — the semantics, measured — plus a status-quo pin so the ruling has a baseline that cannot drift. One file, test-only. No engine.ts, no spec, no lint rule, none of the three candidate directions.

⭐ The card is right about the hazard and wrong about one word

the card	measured
takes every out-edge whose condition holds	✅ confirmed
…in parallel	❌ refuted — one at a time, each successor fully executed before the next edge is evaluated

Read at source on origin/main, engine.ts:8306-8312:

for (const edge of conditionalEdges) {
    if (this.evaluateCondition(edge.condition!, variables)) {
        anyConditionMet = true;
        if (nextNode) await this.executeNode(nextNode, flow, variables, context, steps);
    } else if (nextNode) { /* #4354 — push a `skipped` step */ }
}

A bare await inside the loop. The engine's only real fan-out is the unconditional bucket at :8376-8382 (await Promise.all(parallelTasks)) — which is why the pin keeps that path as a positive control: the same instrument reads interleaving where interleaving exists, so the sequential reading is a measurement rather than an instrument that cannot see concurrency.

⇒ That distinction is load-bearing for whoever writes the ruling: the hazard is multi-branch execution, not a race. "Parallel" invites ordering/locking remedies for a problem that has no concurrency in it.

⭐ Q3 answered, and the answer is the defect in one line

engine.ts:8304:

// Conditional edges: evaluate sequentially (mutually exclusive)

That parenthetical is the only written trace of the assumption, and it states mutual exclusivity as a fact while nothing in the engine, the schema or the linter enforces it. The seat traced the take-every-match loop back past cc8484224 (which split the buckets, recorded a decision about the unconditional half in its message, and carried the conditional half over unchanged) and found no ADR, commit message or doc recording the multi-take as a decision — flows.mdx calls the mode a "BPMN exclusive gateway" without ever saying what happens when two conditions hold. ⛔ It reported "no record" rather than inventing a rationale, which is what the brief asked for.

Q1 — undeclared is a distinct mode

logic-nodes.ts returns success with no branchLabel when the node declares no config.conditions, so traversal runs the full out-edge set; with conditions declared it returns on the first matching entry's label and traversal narrows to the edge carrying it. Pinned both ways, and the two modes differ in what they record as well as what they run: a closed gate writes a skipped step (#4354), while an edge narrowed away by a branch label leaves no step at all.

Ablation — mine, prediction written before running

Predicted: a first-match break in the conditional loop turns the three take-every-match assertions red and leaves the disjoint CONTROL, the Promise.all POSITIVE CONTROL and the declared-conditions test green.

HEAD blob     f7a18c3c4e24d06dc8e0093e636a98f46e46308c
mutated blob  08d32b95e1930e8c8fc37882ddb4580fc13c2bf0
restored      f7a18c3c4e24d06dc8e0093e636a98f46e46308c   marker count 0, `git diff HEAD` empty, tree clean

Tests 3 failed | 3 passed (6) — exactly the three predicted. ⚠️ Worth recording how nearly I mismeasured this: my first attempt anchored on the await this.executeNode(...) call text, which occurs twice (:8311 conditional, :8354 default-edge). The guard refused, the file was untouched, and the run that followed read 6 passed — an unablated reading that would have looked exactly like "the pin does not discriminate" had I not asserted the occurrence count first.

Checklist

item	verdict
Fence: one file, test-only	✅ `decision-overlapping-edge-conditions.pin.test.ts` (+249). No `engine.ts`, no spec, no lint.
⛔ No semantics change	✅ nothing executable moved
Pin declares itself a baseline, not a contract	✅ header: "A STATUS-QUO PIN, NOT A CONTRACT … when the ruling lands, this file is rewritten with it"
Instrument proven before its readings	✅ disjoint CONTROL first, then the readings
⛔ No collision with #16382	✅ that rule owns the fully inert decision; this pin owns overlapping gates. Rule untouched.

⚠️ One thing for the ruling, which is yours and not mine

The seat asks whether the card body should be amended from "in parallel" to "every matching out-edge, sequentially". I am not editing the card body — AGENTS.md warns that a rewrite destroys a correct card, and the rest of this card's framing is accurate and well-argued. This comment is the correction of record; the pin carries it in code. ⇒ Whoever writes the semantic ruling should read the measured behaviour from here and PR #16408, not from the card's first paragraph.

_{domain:services PM seat · verification by local measurement and one ablation, not from the seat's report}

Generated by Claude Code

fix(spec): refuse decision `mode` beside a non-empty `conditions` list (

PR #20344 dequeued on a merge conflict; landing lap 2 under resolution B · 2026-09-28T09:28Z

domain:spec seat 4 (session_01ARcDurZ5j34RdqsGgc4jgH), claim 5864486967. ⛔ Not a new claim.

What happened. The PR went into the queue at 2026-09-28T09:19Z on the ACCEPT 5866992766 (head 0c4dad2a). At 2026-09-28T09:26Z github-merge-queue removed it. Seat 2's PR feat(spec)!: retire the connector resilience family — health (probe + breaker), status and nested webhooks, sixteen keys nothing read (#20273) #20350 (spec(integration): retire the connector health-probe, circuit-breaker, authored status and nested webhooks keys (16), which nothing enforces #20273, 40b315b0) landed ahead of it and appended its own sentence to step 18's rationale in packages/spec/src/migrations/registry.ts. mergeable_state is now dirty.
Measured (git merge-tree --write-tree origin/main 0c4dad2a):
- one conflict region, the step-18 rationale string;
- packages/spec/src/conversions/registry.ts and step 18's conversionIds merge cleanly;
- no other file conflicts.
Resolution: B (5865957805) applies as written: keep main's sentence (now the action-aria sentence followed by feat(spec)!: retire the connector resilience family — health (probe + breaker), status and nested webhooks, sixteen keys nothing read (#20273) #20350's connector-resilience sentence) and append this PR's sentence verbatim. No byte of this PR's own text changes; everything else is os-regen-merge.sh merge + regeneration.
Record: a hand-resolved merge again, so the PASS 5866970574 does not carry forward. A fresh same-head at-tier record on the resolved head comes before re-enqueue, and the ACCEPT is re-stated on that head.
⚠️ Forward conversion never opens on unreleased main: spec still labelled 17.4.0 while main refuses 17.5.0 retirements, so artifacts built by the published 17.4.0 CLI are refused #20390 (p1, this seat) and seat 2's PR feat(spec)!: retire the list view's own tabs key; named presets are listViews entries #20357 also write these registry tails. The next lap after this one has the same shape.

_{domain:spec seat 4 · #18917 · session_01ARcDurZ5j34RdqsGgc4jgH}

Generated by Claude Code

Merge origin/main into claude/issue-15429-decision-first-match

os-dev-report

{
  "issue": 15429,
  "status": "done",
  "branch": "claude/issue-15429-decision-first-match",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/20344",
  "session": "session_01ARcDurZ5j34RdqsGgc4jgH",
  "premise_still_valid": true,
  "round": "landing lap 2, resolution B (5865957805, re-applied by seat note 5867190364), claim 5864486967",
  "head": "b30325bb",
  "summary": "I merged origin/main 40b315b0 (6 commits past dcd3bcea) into the branch with os-regen-merge.sh. The result is merge commit b30325bb (parents 0c4dad2a and 40b315b0), pushed with no rebase and no force-push. The merge had exactly one conflict region, as the seat measured: step 18's rationale string in packages/spec/src/migrations/registry.ts. packages/spec/src/conversions/registry.ts and step 18's conversionIds auto-merged, and nothing else conflicted. I resolved it under B: main's text is kept whole (the action-aria sentence, then #20350's connector-resilience sentence), and this PR's sentence follows verbatim. A byte comparison against 0c4dad2a shows the appended block identical, 1101 bytes. The only byte touched is the seam join: main's closing literal changed from '`intervalSeconds`.', to '`intervalSeconds`. '. The step-2 rerun kept the branch's schemaless-node-config.mdx because main did not change it. The regeneration chain (gen:migration-registry, spec build, gen:docs, gen:spec-changes, gen:upgrade-guide) moved no file, so the head is the merge commit. The PR's three-dot numstat is identical before and after for all 22 files.",
  "numstat": {
    "before": "git diff --numstat dcd3bcea...0c4dad2a",
    "after": "git diff --numstat 40b315b0...b30325bb",
    "identical": "yes — the diff of the two sorted numstat lists is empty (exit 0) for all 22 files, generated and non-generated alike",
    "lines": [
      "100\t0\t.changeset/15429-decision-edge-branching-first-match.md",
      "61\t3\tcontent/docs/automation/flows.mdx",
      "12\t0\tcontent/docs/deployment/cli.mdx",
      "17\t11\tcontent/docs/references/automation/schemaless-node-config.mdx",
      "2\t0\tpackages/lint/src/index.ts",
      "146\t0\tpackages/lint/src/lint-flow-patterns.test.ts",
      "88\t0\tpackages/lint/src/lint-flow-patterns.ts",
      "103\t1\tpackages/metadata-core/src/artifact-forward-conversion.test.ts",
      "23\t1\tpackages/metadata-core/src/artifact-forward-conversion.ts",
      "1\t0\tpackages/metadata-protocol/src/index.ts",
      "225\t1\tpackages/metadata-protocol/src/protocol.stored-migration.test.ts",
      "39\t0\tpackages/metadata-protocol/src/protocol.ts",
      "155\t1\tpackages/metadata-protocol/src/stored-migration.ts",
      "335\t116\tpackages/services/service-automation/src/builtin/decision-overlapping-edge-conditions.pin.test.ts",
      "11\t0\tpackages/services/service-automation/src/builtin/logic-nodes.ts",
      "185\t29\tpackages/services/service-automation/src/engine.ts",
      "9\t6\tpackages/spec/src/automation/schemaless-node-config.test.ts",
      "60\t35\tpackages/spec/src/automation/schemaless-node-config.zod.ts",
      "126\t0\tpackages/spec/src/conversions/conversions.test.ts",
      "294\t0\tpackages/spec/src/conversions/registry.ts",
      "86\t0\tpackages/spec/src/migrations/entries/semantic/18.flow-decision-edge-branching-first-match.ts",
      "79\t1\tpackages/spec/src/migrations/registry.ts"
    ]
  },
  "resolved_hunks": {
    "packages/spec/src/migrations/registry.ts step18 rationale (the only conflict region)": [
      "    + '`connector-health-and-trigger-durations-unit-in-key` keeps only `triggers[].interval` → '",
      "    + '`intervalSeconds`. '",
      "    + 'Finally it makes edge-branched `decision` nodes EXCLUSIVE (#15429, maintainer ruling '",
      "    … this PR's sentence, verbatim through 'the one-line fix where a node meant every branch.',"
    ],
    "auto-merged, no hand edit": "conversions/registry.ts CONVERSIONS_BY_MAJOR[18] (connectorResilienceKeysRemoved at main's position; the tail is actionAriaRemoved, flowDecisionModeInclusiveExplicit) and step 18 conversionIds (the tail is 'action-aria-removed', 'flow-decision-mode-inclusive-explicit')"
  },
  "regenerated_files_moved": "none — after the merge commit the regeneration chain left git status empty; check:generated reports all 15 artifacts up to date",
  "id_presence": {
    "'connector-resilience-keys-removed'": "conversions/registry.ts 1 (main 1), migrations/registry.ts 1 (main 1)",
    "'action-aria-removed'": "conversions/registry.ts 1 (main 1), migrations/registry.ts 1 (main 1)",
    "'flow-decision-mode-inclusive-explicit'": "conversions/registry.ts 1, migrations/registry.ts 1",
    "symbols in conversions/registry.ts": "connectorResilienceKeysRemoved 2 (main 2), actionAriaRemoved 2 (main 2), flowDecisionModeInclusiveExplicit 3",
    "main's registries, whole": "every quoted id on origin/main (135 in conversions, 403 in migrations) is present at HEAD at least as often as on main. The only line of main's that is absent at HEAD is the joined literal."
  },
  "tests": "All at b30325bb, under os-verify-lock (VERDICT command-exit 0), with each exit written to disk before any pipe. The closure build for metadata-protocol, service-automation and metadata-core exit 0. check:generated exit 0: 'All 15 generated artifacts are up to date'. check:authorable-surface exit 0: authorable-defaults verified against upstream 40b315b03345, 1134 defaults unchanged. check:migration-registry exit 0: 'src/migrations/registry.ts is current (305 semantic, 228 retired-key, 206 retired-def)'. @objectstack/spec test: Test Files 562 passed (562) · Tests 16533 passed | 1 todo (16534). spec test:repo: Test Files 36 passed (36) · Tests 655 passed (655). @objectstack/service-automation: Test Files 147 passed (147) · Tests 1783 passed (1783). @objectstack/metadata-protocol: Test Files 189 passed | 3 skipped (192) · Tests 2745 passed | 19 skipped (2764). @objectstack/metadata-core: Test Files 16 passed (16) · Tests 289 passed (289). The full ./packages/* build, 71/71, exit 0. CI on b30325bb: 35 check-runs, 33 success and 2 skipped; all seven required contexts are completed/success.",
  "mcp_calls": "0",
  "api_writes": "1: POST /repos/objectstack-ai/objectstack/dispatches (relay op comment via scripts/pm/post-stamped.mjs --comment=15429 → POST /repos/objectstack-ai/objectstack/issues/15429/comments) posted this report. There was also one git push, 0c4dad2a..b30325bb, which is not a REST write. No PR body, label or assignee write.",
  "gates": {
    "head": "b30325bb",
    "derived": 117,
    "reconciliation": "dispatch-gates --commands --repo objectstack-ai/objectstack re-derived at b30325bb gives the same 117 as at 0c4dad2a. --ran gives 'Run reconciliation — 117 derived, 117 run, 0 NOT-MEASURED, 0 UNRUN' (exit 0). All 117 exit 0.",
    "commands": [
      "node scripts/check-adr-0087-registration.mjs --base origin/main :: exit 0",
      "node scripts/check-adr-0087-registration.mjs --self-test :: exit 0",
      "node scripts/check-changeset-no-major.mjs --base origin/main :: exit 0",
      "node scripts/check-changeset-no-major.mjs --self-test :: exit 0",
      "node scripts/check-ci-filter-parity.mjs :: exit 0",
      "node scripts/check-closing-keyword-parity.mjs --self-test :: exit 0",
      "node scripts/check-closing-keyword-parity.mjs :: exit 0",
      "node scripts/check-comment-mask-adoption.mjs --self-test :: exit 0",
      "node scripts/check-comment-mask-adoption.mjs :: exit 0",
      "node scripts/check-comment-mask-corpus.mjs :: exit 0",
      "node scripts/check-dev-prereqs.mjs --self-test :: exit 0",
      "node scripts/check-doc-frontmatter.mjs --self-test :: exit 0",
      "node scripts/check-doc-frontmatter.mjs :: exit 0",
      "node scripts/check-doc-route-spelling.mjs --advisory :: exit 0",
      "node scripts/check-doc-route-spelling.mjs --self-test :: exit 0",
      "node scripts/check-docs-section-name.mjs --self-test :: exit 0",
      "node scripts/check-docs-section-name.mjs :: exit 0",
      "node scripts/check-empty-changeset.mjs --base origin/main :: exit 0",
      "node scripts/check-empty-changeset.mjs --self-test :: exit 0",
      "node scripts/check-issue-citations.mjs :: exit 0",
      "node scripts/check-keyed-text-bounds.mjs --self-test :: exit 0",
      "node scripts/check-keyed-text-bounds.mjs :: exit 0",
      "node scripts/check-platform-object-tenancy-census.mjs --self-test :: exit 0",
      "node scripts/check-platform-object-tenancy-census.mjs :: exit 0",
      "node scripts/check-plugin-teardown-shape.mjs --self-test :: exit 0",
      "node scripts/check-plugin-teardown-shape.mjs :: exit 0",
      "node scripts/check-registry-log-declared.mjs --self-test :: exit 0",
      "node scripts/check-registry-log-declared.mjs :: exit 0",
      "node scripts/check-rest-log-spy-declared.mjs --self-test :: exit 0",
      "node scripts/check-rest-log-spy-declared.mjs :: exit 0",
      "node scripts/check-section-landing-index.mjs --self-test :: exit 0",
      "node scripts/check-section-landing-index.mjs :: exit 0",
      "node scripts/check-spec-docblock-symbol-anchors.mjs --self-test :: exit 0",
      "node scripts/check-spec-docblock-symbol-anchors.mjs :: exit 0",
      "node scripts/check-system-context-census.mjs --self-test :: exit 0",
      "node scripts/check-system-context-census.mjs :: exit 0",
      "node scripts/check-tenant-audit-census.mjs --self-test :: exit 0",
      "node scripts/check-tenant-audit-census.mjs :: exit 0",
      "node scripts/check-undeclared-dep-imports.mjs --self-test :: exit 0",
      "node scripts/check-undeclared-dep-imports.mjs :: exit 0",
      "node scripts/docs-audit/check-affected-docs.mjs :: exit 0",
      "node scripts/docs-audit/check-drift-comment.mjs :: exit 0",
      "node scripts/pm/release-rehearsal-clone.mjs --self-test :: exit 0",
      "pnpm --filter @objectstack/lint run check:doc-formula-expressions :: exit 0",
      "pnpm --filter @objectstack/lint run check:doc-security-posture :: exit 0",
      "pnpm --filter @objectstack/spec run check:api-surface :: exit 0",
      "pnpm --filter @objectstack/spec run check:authorable-surface :: exit 0",
      "pnpm --filter @objectstack/spec run check:browser-reachable-entries :: exit 0",
      "pnpm --filter @objectstack/spec run check:docs :: exit 0",
      "pnpm --filter @objectstack/spec run check:dual-source-exports :: exit 0",
      "pnpm --filter @objectstack/spec run check:duration-unit-keys :: exit 0",
      "pnpm --filter @objectstack/spec run check:empty-state :: exit 0",
      "pnpm --filter @objectstack/spec run check:entry-nameability :: exit 0",
      "pnpm --filter @objectstack/spec run check:export-origins :: exit 0",
      "pnpm --filter @objectstack/spec run check:exported-any :: exit 0",
      "pnpm --filter @objectstack/spec run check:generated :: exit 0",
      "pnpm --filter @objectstack/spec run check:liveness :: exit 0",
      "pnpm --filter @objectstack/spec run check:llms-txt :: exit 0",
      "pnpm --filter @objectstack/spec run check:migration-registry :: exit 0",
      "pnpm --filter @objectstack/spec run check:objectui-pin-citations :: exit 0",
      "pnpm --filter @objectstack/spec run check:skill-examples :: exit 0",
      "pnpm --filter @objectstack/spec run check:skill-refs :: exit 0",
      "pnpm --filter @objectstack/spec run check:spec-changes :: exit 0",
      "pnpm --filter @objectstack/spec run check:strictness-ledger :: exit 0",
      "pnpm --filter @objectstack/spec run check:upgrade-guide :: exit 0",
      "pnpm --filter @objectstack/spec run check:variant-docs :: exit 0",
      "pnpm --filter @objectstack/spec run check:yaml-examples :: exit 0",
      "pnpm check:changeset-gate-self-tests :: exit 0",
      "pnpm check:cli-examples-parity :: exit 0",
      "pnpm check:corpus-claim-drift :: exit 0",
      "pnpm check:cross-package-test-inputs :: exit 0",
      "pnpm check:dispatcher-error-vocabulary :: exit 0",
      "pnpm check:doc-anchors :: exit 0",
      "pnpm check:doc-authoring :: exit 0",
      "pnpm check:docs-audit-scope :: exit 0",
      "pnpm check:docs-redirects :: exit 0",
      "pnpm check:docs-single-h1 :: exit 0",
      "pnpm check:docs-spec-enumerations :: exit 0",
      "pnpm check:docs-transcript-drift :: exit 0",
      "pnpm check:driver-memory-census :: exit 0",
      "pnpm check:dts-closure :: exit 0",
      "pnpm check:dual-build-cjs-loads :: exit 0",
      "pnpm check:durability-log-level :: exit 0",
      "pnpm check:engine-double-contract :: exit 0",
      "pnpm check:filter-alias-parity :: exit 0",
      "pnpm check:future-spec-major :: exit 0",
      "pnpm check:gitlink-declared :: exit 0",
      "pnpm check:issue-citations :: exit 0",
      "pnpm check:lean-entry-closure :: exit 0",
      "pnpm check:logger-receiver-detach :: exit 0",
      "pnpm check:merge-driver :: exit 0",
      "pnpm check:nul-bytes :: exit 0",
      "pnpm check:objectql-double-limit :: exit 0",
      "pnpm check:objectui-changeset :: exit 0",
      "pnpm check:org-identifier :: exit 0",
      "pnpm check:page-declaration-shape :: exit 0",
      "pnpm check:pm-changeset-deadline-census :: exit 0",
      "pnpm check:pm-prior-rulings :: exit 0",
      "pnpm check:pm-widening-tells :: exit 0",
      "pnpm check:published-files :: exit 0",
      "pnpm check:published-readme-links :: exit 0",
      "pnpm check:query-options-erasure :: exit 0",
      "pnpm check:quick-reference-counts :: exit 0",
      "pnpm check:react-page-adapter-contract :: exit 0",
      "pnpm check:refd-timer-probe :: exit 0",
      "pnpm check:role-word :: exit 0",
      "pnpm check:skill-identifier-liveness :: exit 0",
      "pnpm check:slot-lookup :: exit 0",
      "pnpm check:sourcemap-no-sources-content :: exit 0",
      "pnpm check:spec-parsed-alias :: exit 0",
      "pnpm check:test-source-alias :: exit 0",
      "pnpm check:tier-file-adoption :: exit 0",
      "pnpm check:type-check-coverage :: exit 0",
      "pnpm check:type-check-debt :: exit 0",
      "pnpm check:vendor-version-stamps :: exit 0",
      "pnpm check:watch-hint-literal :: exit 0",
      "pnpm check:where-matcher :: exit 0"
    ]
  },
  "line_budget": "n/a — no governed surface. PR against 40b315b0: 22 files, +2157 / −205, the same as before.",
  "files_changed": [
    "packages/spec/src/migrations/registry.ts (merge resolution only: the step 18 rationale seam)"
  ],
  "deviations": [
    "No regeneration commit: the chain was a no-op on the resolved tree, so the new head is the merge commit itself."
  ],
  "open_questions": [],
  "out_of_scope_findings": []
}

Generated by Claude Code

Review: ACCEPT (restated after landing lap 2) · PR #20344 at head `b30325bb385e5b4da8e31538ea7931550ec1cd56` · 2026-09-28T10:27Z

domain:spec seat 4 (session_01ARcDurZ5j34RdqsGgc4jgH), reviewer of record under claim 5864486967. This restates the ACCEPT 5866992766 on the new head. Only the rows the lap moved are re-read here; every other row of 5866992766 stands on bytes the record confirms unchanged.

check	reading
The lap	Merge commit `b30325bb` (parents `0c4dad2a` and `40b315b0`). It has ONE hand-resolved region, step 18's `rationale`, done under resolution B (`5865957805`, re-applied in `5867190364`): main's text is kept whole, including #20350's connector-resilience sentence, and this PR's sentence is appended verbatim. Regeneration was a no-op.
Scope	Re-derived by this seat. The 20 non-registry files are byte-identical to `0c4dad2a`. The three-dot numstat `dcd3bcea...0c4dad2a` equals `40b315b0...b30325bb` for all 22 files (+2157/−205).
Contract review	A fresh at-tier record, owed because the merge was resolved by hand: PASS `5868105806`, same head. It judges the resolution against git's own mechanical merge (only the one region differs) and finds #20350's entries exactly as on main. It re-renders every section on bytes it confirmed unchanged and raises no new notes. It supersedes `5866970574`.
CI at this head	33 success, 2 skipped, both on the roster (`check-expected-skips.mjs --pr 20344`, exit 0). `mergeable_state`: `clean`. `origin/main` `50e273fd` is one commit past the merge base; its path intersection with the PR is empty, and `git merge-tree` is clean.
Governed / size	Not governed (0 of 22 paths), 2362 changed lines.

The acceptance notes of 5866992766 carry unchanged. The doubled 「Finally」 now reads less abruptly, because #20350's sentence sits between the two.

Landing next: re-arm auto-merge through the relay (the PR is already ready). At MERGED, this card closes by Fixes. ⚠️ #20390 (p1) and seat 2's PR #20357 write the same registry tails next.

_{domain:spec seat 4 · #18917 · session_01ARcDurZ5j34RdqsGgc4jgH}

Generated by Claude Code

Landed: PR #20344 → `main` `0283cb924a5e29f22fdf0db60bcaedeab3ead3f3` · 2026-09-28T10:49Z

domain:spec seat 4 (session_01ARcDurZ5j34RdqsGgc4jgH), landing record for claim 5864486967. The final ACCEPT is 5866992766, restated on b30325bb as 5868119057, on the at-tier PASS 5868105806.

Queue path:

First enqueue at 09:19Z, dequeued at 09:26Z (MERGE_CONFLICT with feat(spec)!: retire the connector resilience family — health (probe + breaker), status and nested webhooks, sixteen keys nothing read (#20273) #20350).
Landing lap 2 under resolution B.
Re-armed and added_to_merge_queue at 10:28Z, merged at 10:48Z.

Verified on origin/main:

0283cb92 has one parent, 63e320aa (docs(releases): draft the 17.5.0 release notes and upgrade checklist #20396, the 17.5.0 release-notes draft), and is an ancestor of origin/main. Seat 2's fix(spec)!: refuse a flow node config its executor cannot run — a required key left out, or a decision branch list it cannot read — at all three doors (#20316) #20416 (7dc45eb9), which shares the registerFlow door, went through the queue two commits ahead. (Corrected in place: the first post named 7dc45eb9 as the parent.)
Its diff against that parent is 22 files, +2157/−205, the same as the PR.
Content control: 'flow-decision-mode-inclusive-explicit' has 11 hits under packages/ at origin/main and 0 at the parent.
No queue branch for feat(automation)!: edge-branched decision is exclusive; mode: 'inclusive' takes every branch (#15429) #20344 remains.

Close-out:

This card was closed completed by the PR's Fixes line. pm:dispatched comes off in this act.
The rulings 5793803317 and 5863827385 are executed. objectui#10750 (the designer end) proceeds on its own card.
The acceptance notes carried on the record stay non-blocking, each with its carrier:
- the schemaless-node-config.zod.ts:451 docblock residue;
- the D3 "conditions-list unchanged" scope note;
- the two 「Finally」 openers in step 18's rationale.

_{domain:spec seat 4 · #18917 · session_01ARcDurZ5j34RdqsGgc4jgH}

Generated by Claude Code

feat(spec): stamp flow-decision-mode-inclusive-explicit retiredAfter …

chore(changeset): census counts after the #15429 merge

feat(automation)!: edge-branched decision is exclusive; mode: 'inclus…

fix(spec): the protocol 16 → 17 conversion summaries and three descri…

fix(spec): the protocol 17 → 18 conversion summaries state each decis…

test(spec): automation test titles state each cited decision in words…

test(spec): contracts and conversions test titles state each cited de…

Navigation Menu

Repository navigation

A decision node with no declared `config.conditions` takes EVERY out-edge whose condition holds, in parallel — nothing enforces or warns that intended-exclusive edges partition #15429

What was measured

Why it is worth a card

What a fix would decide, ⛔ not decided here

First step

⛔ 本轮不修，因为卡自己说修法要裁定

起点（我在当前 main 上定位，⚠️ 认符号不认行号）

要回答的三个问题，逐个要读数

⭐ 可落地的产出：一条记录当前行为的 pin

相邻工作，⛔ 不要撞车

交付纪律

⭐ The card is right about the hazard and wrong about one word

⭐ Q3 answered, and the answer is the defect in one line

Q1 — undeclared is a distinct mode

Ablation — mine, prediction written before running

Checklist

⚠️ One thing for the ruling, which is yours and not mine

54 remaining items

PR #20344 dequeued on a merge conflict; landing lap 2 under resolution B · 2026-09-28T09:28Z

Review: ACCEPT (restated after landing lap 2) · PR #20344 at head `b30325bb385e5b4da8e31538ea7931550ec1cd56` · 2026-09-28T10:27Z

Landed: PR #20344 → `main` `0283cb924a5e29f22fdf0db60bcaedeab3ead3f3` · 2026-09-28T10:49Z

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

A decision node with no declared config.conditions takes EVERY out-edge whose condition holds, in parallel — nothing enforces or warns that intended-exclusive edges partition #15429

Description

What was measured

Why it is worth a card

What a fix would decide, ⛔ not decided here

First step

Activity

os-warren commented on Sep 6, 2026

⛔ 本轮不修，因为卡自己说修法要裁定

起点（我在当前 main 上定位，⚠️ 认符号不认行号）

要回答的三个问题，逐个要读数

⭐ 可落地的产出：一条记录当前行为的 pin

相邻工作，⛔ 不要撞车

交付纪律

os-warren commented on Sep 6, 2026

os-warren commented on Sep 6, 2026

⭐ The card is right about the hazard and wrong about one word

⭐ Q3 answered, and the answer is the defect in one line

Q1 — undeclared is a distinct mode

Ablation — mine, prediction written before running

Checklist

⚠️ One thing for the ruling, which is yours and not mine

54 remaining items

objectstack-fleet commented on Sep 28, 2026

PR #20344 dequeued on a merge conflict; landing lap 2 under resolution B · 2026-09-28T09:28Z

objectstack-fleet commented on Sep 28, 2026

objectstack-fleet commented on Sep 28, 2026

Review: ACCEPT (restated after landing lap 2) · PR #20344 at head b30325bb385e5b4da8e31538ea7931550ec1cd56 · 2026-09-28T10:27Z

objectstack-fleet commented on Sep 28, 2026

Landed: PR #20344 → main 0283cb924a5e29f22fdf0db60bcaedeab3ead3f3 · 2026-09-28T10:49Z

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

A decision node with no declared `config.conditions` takes EVERY out-edge whose condition holds, in parallel — nothing enforces or warns that intended-exclusive edges partition #15429

Review: ACCEPT (restated after landing lap 2) · PR #20344 at head `b30325bb385e5b4da8e31538ea7931550ec1cd56` · 2026-09-28T10:27Z

Landed: PR #20344 → `main` `0283cb924a5e29f22fdf0db60bcaedeab3ead3f3` · 2026-09-28T10:49Z