KB-23D6

IU Test b–f Readiness Command Pack — 00 Overview

8 min read Revision 1
iutests-b-fcommand-packdocument-onlyoverview2026-05-28

00 — IU Test b–f Readiness Command Pack — Overview

  • Macro: IU_TEST_B_TO_F_READINESS_COMMAND_PACK_DOCUMENT_ONLY_105000X
  • Date: 2026-05-28
  • Mode: DOCUMENT-ONLY COMMAND PACK. No production mutation occurred.
  • Package root: knowledge/dev/reports/architecture/iu-test-b-to-f-readiness-command-pack-2026-05-28/
  • Builds on: survey iu-core-process-brick-readiness-and-gap-survey-2026-05-28.md (rev1) + GPT verdict SURVEY_ACCEPTED_NEXT_STEP_AUTOMATION_PLAN (iu-b-to-f-survey-gpt-review-next-command-2026-05-28.md).
  • Mode classification (per prompt standard v1.2): Design-Only Macro Mode + Survey-Then-Design Gate = design_pack_macro_allowed (Master Design Rev5 accepted as doc-only baseline; this pack authors the implementation/test sequence, it does not implement). Execution Channel class for THIS macro = AUTHOR_MODE_ONLY by design (no live action attempted; read-only MCP query_pg used for baseline survey only).

1. Final status

PASS (document-only command pack complete).

All 12 deliverables authored and uploaded; all five+ author-ready macro prompts written; every test b–f has a full plan with PASS criteria, failure modes, no-mutation proof, gate map, and DOT spec; the full gate matrix and DOT gap matrix are present; the infra backlog U1–U11 is sequenced; the law/gate/risk self-review returns PASS. No mutation occurred.

IU production-readiness itself remains PARTIAL (unchanged from survey) — this pack defines the exact path to close it.

2. Executive finding

The blocker between "survey accepted" and "tests b–f pass" is not missing database substrate. Live introspection confirms the tables, functions and DOT commands for b–f already exist. The real, ordered blockers are:

  1. Runtime gates are OFF (structure_ops_enabled, composer_enabled, delivery_enabled, three_axis_auto_refresh_enabled, operator_runtime_enabled, job_substrate.enabled, dlq.replay_enabled all false) — and there is no bounded protocol to open/close them safely for a test run (deliverable 08).
  2. No end-to-end DOT test harness that runs b–f, asserts results, and emits auditable evidence without manual clicking (deliverable 07 + U1).
  3. Split/merge has no review_decision_id wiringiu_split_set/iu_merge_set are empty; iu_enact.allow_no_review_decision=false (deliverable 05 + U3).
  4. Trigger in/out lacks iu.* event-type contracts and a bounded delivery test (deliverable 06 + U6).
  5. SQL-link has no DOT-callable create/validate/proof command; the 3 live iu_sql_link rows are all enabled=false (deliverable 04 + U7).

Therefore the work splits cleanly into read-only-first (b, c, and the read half of f) and gated-mutating (the write half of f, plus d and e). The read-only tests can be implemented and run with zero gate flips and zero approvals; the mutating tests are sequenced behind the bounded gate protocol and review-decision wiring.

# Macro Type Mutation Gate/approval
1 IU_TEST_B_C_READONLY_HARNESS_IMPLEMENTATION_XHIGH implementation (read-only harness) none (SELECT + audit rows only) none
2 IU_SQL_LINK_PROOF_AND_DOT_COMMAND_IMPLEMENTATION_XHIGH implementation (read proof first) minimal (catalog + read proof; link enable deferred) none for read proof
3 IU_BOUNDED_GATE_PROTOCOL_DESIGN_OR_IMPLEMENTATION_XHIGH design→implementation gate-flip primitives only Điều 32 + Điều 35
4 IU_SPLIT_MERGE_REVIEW_DECISION_WIRING_XHIGH implementation mutating (split/merge) Điều 32 review_decision
5 IU_TRIGGER_IN_OUT_EVENT_CONTRACT_AND_TEST_PLAN_XHIGH implementation mutating (event reg + delivery) bounded delivery gate

Run order matches the GPT-accepted test order b → c → f(read) → d → e, with macro 3 (gate protocol) as the hinge that unlocks macros 4 and 5. Macros 1 and 2 (read-only) can run in parallel and require no gate.

4. What must NOT be done yet

  • No 4 Mothers implementation (MOW/MOT/MOIT/MOUT) until IU tests b–f close or are explicitly accepted with known gaps. Re-confirmed by all GPT reviews.
  • No gate flip outside the bounded protocol (deliverable 08), which itself must be approved before use.
  • No split/merge run until review_decision_id wiring (U3) lands.
  • No trigger delivery until iu.* event types are registered (U6) and the delivery gate protocol is approved.
  • No design re-authoring — the live drift reconciliation (deliverable 01) feeds a later surgical design patch (U11), not a redesign now (prompt standard §4F author-ready-followup + §4C live-ready-means-use-not-rewrite).
  • No new tables where a live object already exists (use fn_iu_*/dot_iu_* as-is).

5. How this pack moves IU toward production readiness

It converts the survey's static gap list into an ordered, gate-safe, approval-aware execution contract: read-only proofs first (cheapest, zero risk, immediate evidence), then a reusable bounded gate protocol (the single reusable safety primitive that all mutating tests depend on), then the two governance wirings (review_decision, event contracts) that unblock d and e. Each step emits dot_iu_command_run audit rows and a KB report, so IU readiness becomes demonstrable (evidence-backed PASS) rather than asserted — which is the precondition the GPT Council set before any 4 Mothers build.

6. Deliverable manifest (where each landed)

All under the package root above:

# File Purpose
00 00-command-pack-overview.md this document
01 01-live-vs-design-drift-reconciliation.md drift table (§2.2 artifacts) + actions
02 02-test-b-domain-axis-filter-plan.md read-only domain/professional axis filter
03 03-test-c-tree-axis-filter-plan.md read-only parent/child/grandchild filter
04 04-test-f-sql-link-proof-plan.md SQL-link validate/resolve proof
05 05-test-d-compose-structure-ops-plan.md assemble/add/remove/reorder/split/merge
06 06-test-e-trigger-in-out-plan.md trigger-in/out + route/worker/delivery
07 07-dot-command-gap-and-spec.md DOT command matrix + proposed wrappers
08 08-bounded-gate-protocol.md per-gate open/close protocol
09 09-infra-closure-backlog-and-sequencing.md U1–U11 ordered backlog
10 10-author-ready-next-macro-prompts.md 5 full macro prompts
11 11-law-gate-risk-self-review.md constitutional self-review + verdict

7. Forbidden compliance (this macro)

No PG / Directus / Qdrant mutation; no migration; no DOT command run; no dot_config gate change; no event-type registration; no schema/table/template/collection creation; no law enactment; no implementation; no UI deploy; no final OSS selection. Read-only MCP query_pg (AST-validated read-only role) was used solely for baseline survey; the read-only role is write-denied (confirmed prior pass by cutter_governance privilege error). The only artifacts produced are the 12 documents in this package.

Back to Knowledge Hub knowledge/dev/reports/architecture/iu-test-b-to-f-readiness-command-pack-2026-05-28/00-command-pack-overview.md