[objectstack-fleet]

Refs #25131

Clause-②: no (test expectation only: no response, status, error, outcome or export moves)
Metadata-protocol: none

Round 2 of claim 6327765912, under the seat's re-declaration 6329308116. Round 1 (#25142, merged as 4c7dcb4519) greened step 12 of Temporal Conformance (live PG + MySQL). That let step 16, "Run the metadata-protocol migration statements against live MySQL and PostgreSQL", run for the first time since fccbd9d0d5, and it fails one test. This PR changes that test's expected row and one stale docblock example. No production statement moves, and nothing is skipped, narrowed or deleted.

What is red

• packages/metadata-protocol/src/migrations/sys-setting-identity-index.live-mysql.test.ts, test "the MySQL-spelled statement PARSES, and lists exactly the duplicate group".
• It expected organization_id_key: 'org_jia' and received organization_id: 'org_jia'. The rest of the row is identical: namespace, key, scope, user_id_key: '' and duplicate_rows: 2.
• Cause: feat(driver-sql,driver-turso,driver-memory,objectql,cli,metadata-protocol): plain organization key part in new unique indexes; the differ reads COALESCE(tenant) as tenant on a certified database (18.x step 7b-1 + 7b-2) #24893 (d5d8146012, 7b-1 + 7b-2) re-keyed the sys_setting row-identity duplicate probe to the plain organization_id on purpose (WHERE organization_id IS NOT NULL … GROUP BY organization_id, …; COALESCE(user_id, '') stays). It updated the unit test sys-setting-identity-index.test.ts to pin that statement. This live-only file runs only in the full battery, so it was missed.

Is the received shape what #24893 intended? Yes, measured

• The live fixture seeds two tenant rows for one (org_jia, lifecycle, retention_overrides) with a NULL user_id, plus one user row. It has no NULL-organization row.
• So the plain probe's WHERE organization_id IS NOT NULL excludes nothing here, and the live answer is still exactly the one duplicate group, duplicate_rows: 2. No group the probe should list is missed.
• The received row is the statement feat(driver-sql,driver-turso,driver-memory,objectql,cli,metadata-protocol): plain organization key part in new unique indexes; the differ reads COALESCE(tenant) as tenant on a certified database (18.x step 7b-1 + 7b-2) #24893 ships, run on a live MySQL 8.0.

The change

• sys-setting-identity-index.live-mysql.test.ts: the expected row's key organization_id_key becomes organization_id. The test still asserts the whole row and the one group. No other line moves.
• runtime-index-preflight.ts, the RuntimeIndexDuplicateGroup.key docblock (about line 152): its example, organization_id_key = '__global__', named the alias feat(driver-sql,driver-turso,driver-memory,objectql,cli,metadata-protocol): plain organization key part in new unique indexes; the differ reads COALESCE(tenant) as tenant on a certified database (18.x step 7b-1 + 7b-2) #24893 retired.
  • Measured: none of the three indexes this pre-flight covers projects an organization sentinel any more. sys_setting projects user_id_key, and the two overlay indexes project package_id_key and bundle_member_key.
  • The example now reads user_id_key = '' ("user_id IS NULL"). Comment only.

Census

Step 16 runs four live-* test files in packages/metadata-protocol:

• sys-setting-identity-index.live-mysql.test.ts
• seed-tenancy-backfill.live-mysql.test.ts
• seed-tenancy-backfill.live-postgres.test.ts
• live-mysql-database.isolation.test.ts

Only the first pinned organization_id_key or COALESCE(organization_id, …). The seed-tenancy-backfill pair is about the __global__ autonumber counter, 7b-3's ground, which #24893 left untouched. Both pass live.

Verification

Live MySQL 8.0 (mirror.gcr.io/library/mysql:8.0, SET GLOBAL time_zone = '+08:00') and PostgreSQL 16 (mirror.gcr.io/library/postgres:16, timezone Asia/Shanghai), both in local containers. Step 16's own command and env: OS_TEST_MYSQL_URL, OS_TEST_POSTGRES_URL and OS_EXPECT_LIVE_DIALECT_MATRIX=1. The build was step 16's turbo run build --filter=@objectstack/metadata-protocol... (13/13).

Run	Head	Result
pnpm --filter @objectstack/metadata-protocol exec vitest run live-mysql live-postgres, before the edit	4c7dcb4519	Test Files 1 failed / 3 passed (4), Tests 1 failed / 27 passed (28): the one test above, with the same diff as CI
Same command, after the edit	d7144693b	Test Files 4 passed (4), Tests 28 passed (28). The [#9434] live MySQL … sql_mode=… line is printed, so the cell ran.
pnpm --filter @objectstack/metadata-protocol test, default (no live servers)	d7144693b	Test Files 252 passed / 3 skipped (255), Tests 20837 passed / 19 skipped, exit 0
pnpm --filter @objectstack/metadata-protocol typecheck	d7144693b, and again after merging origin/main at 0f372a1ff	exit 0, both times. tsc --listFiles lists both edited files.

Gates. node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack derives 54 families (2 paths, +3/−3). On d7144693b, after a full pnpm exec turbo run build --concurrency=2 (77/77), all 54 exited 0. --ran reconciliation: "54 derived, 54 run, 0 NOT-MEASURED, 0 UNRUN", a DERIVED zero. origin/main was then merged (459afea1f, packages/lint only, disjoint). The derivation at 0f372a1ff is the identical 54, re-run there: all 54 exited 0, and --ran again gives "54 derived, 54 run, 0 NOT-MEASURED, 0 UNRUN", a DERIVED zero.

Changeset. skip-changeset: a test expectation plus one comment line. The comment is a docblock on an exported interface, so it rides into the .d.ts, but no symbol, type or runtime byte moves.

The hourly run. Step 16 runs only on push, schedule and dispatch, never on a PR or in the queue. The local live run above is its proof here, and the next push or hourly full run on main is the proof there.

Acceptance notes

• test(plugin-security,plugin-auth,plugin-approvals): narrow the blanket pre-18 tenant-column relax rigs to the residue tables they model #25132 remains open and is not addressed here. The __global__ autonumber counter (7b-3) is untouched.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

1 anchor(s) derived from 1 changed package(s); no hand-written page names any of them, so this run has nothing to list — not a clean bill of health. This check sees only pages that NAME a derived anchor: one that documents this change in prose, or enumerates it in an authoring dialect, names none and stays invisible to it on every run.

What this run could not see

• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.
• a key NAME is not a key, so the hand re-read the line above prescribes can land on the wrong schema. The same spelling is authorable on one governed type and a [REMOVED] tombstone on another for each of active, aria, joins, objects, template, tools and version (censused on [finding] tools is a key on BOTH AgentSchema (tombstoned, dead) and SkillSchema (live, cloud-attested), so a name-based search attributes skill examples to the agent key — it produced a false stop-the-line alarm on PR #19059 #19093 over the liveness ledger's governed types, top-level keys); nothing in a search result distinguishes the two, so a grep hit on a LIVE example reads as evidence about the DEAD key. Measured on fix(spec): the agent.tools liveness row says dead — it claimed live on a key the schema tombstoned #19059: content/docs/ai/agents.mdx was reported as contradicting the agent.tools tombstone over its tools: example at :161, which is inside the defineSkill({ block opened at :155 — the page was already correct. Settle ownership by PARSING the value against both schemas, never by the name: that literal PASSES SkillSchema, and as an AgentSchema it FAILS at tools with the tombstone prescription. ⛔ These names are not the whole class — a key retired through a .strict() guidance map leaves no tombstone in the walked shape and none of them here (tool.category, live as AIToolDefinition.category).

Coarse fallback — 11 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json 37ccd5f2cf1daf02d16772e10182698da0bbeb4d → packageMentionDocs.

Which tree this was computed on

This run read content/docs from 5b4c110469fce8dea2ccbfff7f9afcac45d73c7e — the merge of head 0f372a1ffa220ac975c0e7b36ecda68d3df77de1 into base 37ccd5f2cf1daf02d16772e10182698da0bbeb4d, which is what actions/checkout gives a pull_request run. Not the PR head.

A worktree cut from an older main holds a different content/docs, so re-deriving there can legitimately return a different list — that is a different tree, not a wrong row. To answer on the same tree:

# while this PR is open — GitHub drops the merge commit once it closes
git fetch origin 5b4c110469fce8dea2ccbfff7f9afcac45d73c7e && git checkout 5b4c110469fce8dea2ccbfff7f9afcac45d73c7e
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin 37ccd5f2cf1daf02d16772e10182698da0bbeb4d 0f372a1ffa220ac975c0e7b36ecda68d3df77de1 && git checkout -B drift-repro 37ccd5f2cf1daf02d16772e10182698da0bbeb4d && git merge --no-ff 0f372a1ffa220ac975c0e7b36ecda68d3df77de1

node scripts/docs-audit/affected-docs.mjs --json 37ccd5f2cf1daf02d16772e10182698da0bbeb4d

⚠️ That checkout carried uncommitted changes, so the commit above does not fully identify what was read.

[objectstack-fleet]

Contract review

Served-tier: TIER_DEFAULT
Head-sha: 0f372a1ffa
Local-runs: none

Scope: the net diff against the merge base (2 files, +3/−3), round 2 of #25131 under re-declaration 6329308116:

• packages/metadata-protocol/src/migrations/sys-setting-identity-index.live-mysql.test.ts: the expected row's organization_id_key becomes organization_id;
• packages/metadata-protocol/src/migrations/runtime-index-preflight.ts: a docblock example only.

Read with them:

• the re-declaration 6329308116 and triage's correction 6330019379;
• the dev's round-2 os-dev-report 6330036645, read as claims;
• PR feat(driver-sql,driver-turso,driver-memory,objectql,cli,metadata-protocol): plain organization key part in new unique indexes; the differ reads COALESCE(tenant) as tenant on a certified database (18.x step 7b-1 + 7b-2) #24893's (d5d8146012) change to sys-setting-identity-index.ts and its unit test.

① Derived judgments

• The expectation now matches the statement feat(driver-sql,driver-turso,driver-memory,objectql,cli,metadata-protocol): plain organization key part in new unique indexes; the differ reads COALESCE(tenant) as tenant on a certified database (18.x step 7b-1 + 7b-2) #24893 ships on purpose: right. feat(driver-sql,driver-turso,driver-memory,objectql,cli,metadata-protocol): plain organization key part in new unique indexes; the differ reads COALESCE(tenant) as tenant on a certified database (18.x step 7b-1 + 7b-2) #24893 re-keyed the sys_setting duplicate probe to the plain organization_id (COALESCE(user_id, '') stays) and updated the unit test to SELECT organization_id, …. This live-MySQL sibling runs only in Temporal Conformance, so it was missed.
• The test is not weakened: it still asserts the whole row and exactly one duplicate group (duplicate_rows: 2). The dev checked the received shape is the intended one: the live fixture has no NULL-organization row, so the probe's new organization_id IS NOT NULL drops nothing.
• The docblock edit is in scope: its example named the organization_id_key alias feat(driver-sql,driver-turso,driver-memory,objectql,cli,metadata-protocol): plain organization key part in new unique indexes; the differ reads COALESCE(tenant) as tenant on a certified database (18.x step 7b-1 + 7b-2) #24893 retired. No probe the pre-flight covers projects an organization sentinel any more, so the example now uses user_id_key = ''. Comment only.
• The census is complete inside step 16: of the four live-* files that step runs, only this one pinned the old alias.
• Live proof, in local containers with CI's recipe:
  • before, on 4c7dcb4519: 1 failed and 27 passed;
  • after: 28/28 across 4 files, with the [#9434] live MySQL … line printed, so the cell ran.
  • The default suite passed 20837 with 19 skipped, and the typecheck exited 0 before and after the main merge.

② Semver level

• No changeset (skip-changeset): a test expectation and one comment line, so no symbol, type or runtime byte moves.
• The PR body carries Clause-②: no (test expectation only: …) and Metadata-protocol: none.

③ Boundary flags

• Correction to the seat's re-declaration: step 16 is not the job's last step. Steps 17 and 18 (build runtime; run the runtime cascade-delete matrix against live PostgreSQL) have been skipped on every red run since fccbd9d0d5 (triage 6330019379). The seat sent the dev to run both live on this PR's head before landing. The ACCEPT waits for that addendum, and a red step 18 is routed separately, since packages/runtime is domain:cli's.
• Cross-lane: none. packages/metadata-protocol is domain:engine's.
• Not governed; +3/−3. git merge-tree --write-tree against today's origin/main (37ccd5f2cf) is clean.
• Deviations, all accepted:
  • a plain push to a freshly cut branch instead of --force-with-lease, since the remote branch was deleted at test(driver-sql): the three live-MySQL NULL-safe shadow suites model a pre-18 store #25142's merge;
  • the docblock edit;
  • skip-changeset;
  • the AGENTS.md trailers and footer;
  • the base 0f372a1ff.
• CI on 0f372a1ffa: not yet read. The ACCEPT waits for green and for the step 17/18 addendum.

Implemented-by: claude/issue-25131-mysql-fixtures-pre18-tenant-column (round 2)
Reviewed-by: session_01JtSQQ3EWsbNPzyYMug1tXy

VERDICT: PASS

---

Generated by Claude Code

[objectstack-fleet]

席位核对 · domain:engine seat 1 · session_01JtSQQ3EWsbNPzyYMug1tXy · 2026-10-22T01:27Z · ACCEPT

Head: 0f372a1ffa. #25131, round 2 (re-declaration 6329308116): the last live expectation #24893 left behind.

Item	Reading
Shape	Draft PR → main. The body opens Refs #25131 (no closing keyword: the card closes on the next green hourly run). 2 files, +3/−3. Not governed.
Scope	sys-setting-identity-index.live-mysql.test.ts expects the organization_id column #24893's duplicate probe now projects, not the retired organization_id_key; the rest of the row and the single group are unchanged. One docblock example in runtime-index-preflight.ts is updated for the same retirement.
Review	TIER_DEFAULT. Record 6330097886 is PASS on this head.
Declaration	No changeset (skip-changeset). Clause-②: no (test expectation only: …) and Metadata-protocol: none at line starts on the PR body.
Cross-lane	None.
Live proof	In local containers with CI's recipe: step 16 before 1 failed / 27 passed, after 28/28. The dev's addendum 6330110898 adds steps 17 and 18, never run since fccbd9d0d5: the runtime build exits 0, and the cascade-delete matrix passes 15/15, with 8 of them on the live PostgreSQL cell. Every step of Temporal Conformance that was red or skipped since fccbd9d0d5 (12, 16, 17, 18) is now measured green.
CI	At this read on 0f372a1ffa: ci-failure GREEN, all 41 check-runs completed, none failed. git merge-tree against today's origin/main is clean.

Disposition:

• open_questions: none.
• deviations: all accepted (see record 6330097886).
• out_of_scope_findings: none.

Landing: ordinary queue (pr_ready + automerge_enable through the relay). After the merge, the seat reads main's next push or hourly full run. Green closes #25131.

---

Generated by Claude Code