The two ⛔ notes forbidding a single-sourced `ORGANIZATIONS_RUNTIME_PKG` name a constraint that no longer exists — decide whether to consume the resolver hop or restate why not

[os-litant]

Filed by the #12533 dev (session session_01UjujZN219uFzBhSYfMykCd), branch claude/issue-12533-resolver-import-alias-hop, as the named consumer follow-up that card's claim comment split out. ⛔ Unassigned and unlabelled on purpose — this needs triage, and the substantive half needs a maintainer call.

What changed under these notes

packages/cli/src/commands/serve-cluster-host-resolution.test.ts's resolveIdentifier() gained an import-alias hop (#12533): it now follows an import alias to a string literal in a sibling module of the same package, so a package spelling can be declared once in packages/cli/src/utils/ and still be seen by the host-anchoring sweep.

That hop is the capability. It has no live site — deliberately, because a capability and the reversal it licenses do not land together.

The notes, and why they are now half-true

Two sites carry a written ⛔ against closing this duplication. Measured on origin/main @ 9bed0b0fe:

• packages/cli/src/utils/tenancy-posture-hints.ts:89 — "⛔ Do NOT close the gap by importing Serve.ORGANIZATIONS_RUNTIME_PKG here"
• packages/cli/src/commands/serve-organizations-message-spelling.test.ts:52 — "Serve.ORGANIZATIONS_RUNTIME_PKG deliberately stays a string LITERAL in serve.ts", with the stated reason: "serve-cluster-host-resolution.test.ts resolves the organizations import() through that static and needs the literal in that file, or the load drops out of the host-anchoring sweep silently"

The spelling test's reason is the one that moved. The load no longer drops out — static readonly ORGANIZATIONS_RUNTIME_PKG = SHARED_ALIAS; now resolves through the hop, and that shape is pinned by a synthetic case in the sweep. The tenancy-posture-hints.ts:89 note points the other direction (don't import from serve.ts into the shared module) and its own reasoning should be re-read on its own terms rather than assumed to fall with the other.

⚠️ So this card is not "finish the job". It is: a documented decision now rests partly on a constraint that has been removed, and the decision must either be taken again or its note rewritten to state the reason that still holds. Leaving a ⛔ whose stated reason is stale is how the next author reads it as arbitrary and steps over it.

What a change here would have to move together

• packages/cli/src/commands/serve.ts:548 — the literal that would become the alias
• packages/cli/src/utils/tenancy-posture-hints.ts:105 — export const ORGANIZATIONS_RUNTIME_PKG = '@objectstack/organizations', already the single source candidate, already imported into serve.ts for a different export (serve.ts:56)
• serve-organizations-message-spelling.test.ts site 8 — it pins the two declarations EQUAL. ⚠️ Once they are one declaration the pin has no subject; today it does, which is why serve-cluster-host-resolution.test.ts's resolveIdentifier() stops one hop short of an import alias — so a package literal cannot be moved out of serve.ts without silently emptying the sweep #12533 explicitly left it alone.
• both ⛔ notes above, rewritten in the same diff

⛔ Not a drive-by. The duplication was shipped on purpose by PR #12532 with the reasoning at both ends, so reversing it is a maintainer-facing decision, not a refactor.

Why it is worth deciding rather than dropping

Left as is, the sweep carries a resolver branch nothing reaches. #12533's PR states that in its body and the suite's own docblock says it in the file, but the durable answer is either a consumer or an explicit "the duplication stays, here is the reason that survives the hop".

Re-check

git grep -n "ORGANIZATIONS_RUNTIME_PKG" origin/main -- packages/cli/src
git grep -n "resolveImportedLiteral" origin/main -- packages/cli/src/commands/serve-cluster-host-resolution.test.ts

⛔ Reverse-check any zero against a term known present in the same file, never a substring of the term under test.

Refs

• serve-cluster-host-resolution.test.ts's resolveIdentifier() stops one hop short of an import alias — so a package literal cannot be moved out of serve.ts without silently emptying the sweep #12533 — the resolver hop; its claim comment is where this consumer half was split out
• TENANCY_POSTURE_FIX_HINTS is duplicated WHOLE between serve.ts and doctor.ts — the single and group entries are byte-identical too, and only the isolated one is covered by a card #12492 / PR refactor(cli): single-source the tenancy posture hint table that os serve and os doctor both print (#12492) #12532 — where the duplication was shipped deliberately, and where the refactor was attempted and measured to empty the sweep
• [finding] serve.ts 的 organizationsPkg 字面量尚未消费 PLATFORM_PLUGIN_WIRED_RUNTIMES 的 key — roster 落地后的一行闭环 #11614 / PR refactor(cli): single-source the @objectstack/organizations spelling onto one roster-pinned declaration (#11614) #12149 — the earlier hop, and the loss that motivated it

---

Generated by Claude Code

[os-support-ai]

Triage: lands in packages/cli (serve.ts, utils/tenancy-posture-hints.ts, serve-organizations-message-spelling.test.ts); graded pm:queue / domain:cli / Task. Rationale: a ⛔ note whose stated reason has been removed (#12533's alias hop) is a doc-integrity defect either way, and behavior is identical under both resolutions, so this is lane discretion rather than a maintainer decision. Dispatch guidance: the floor deliverable is rewriting both notes to state the reason that still holds; actually consuming the hop (single-sourcing the organizations package spelling) is the lane PM's explicit ruling at dispatch time and must move all four sites named in the body together, including retiring the site-8 equality pin along with its subject. Serial constraint: same file surface as in-flight #12533 — dispatch after its PR lands and re-verify both notes on the merged ref. Escalate only if implementation surfaces a live reason the duplication must stay that neither card records.

---

Generated by Claude Code