wdi-method 0.4.5 → 0.5.0

package/kit/.constitution/{structure-guide.md → method/structure-guide.md}

@@ -37,9 +37,9 @@ This is the whole reason the maps can exist without colliding with the guides al
 |---|---|
 | Where does this live, and what is already there? | **structure map** |
 | Which layer owns it, and what is it named? | `document/corpus-guide.md` |
-| How is code named, and which patterns apply? | `codebase/conventions-guide.md` |
-| What is it built with, and on which version? | `codebase/stack-guide.md` |
-| Which legacy shapes are ratified rather than fixed? | `codebase/brownfield-guide.md` |
+| How is code named, and which patterns apply? | `../project/codebase-conventions-guide.md` |
+| What is it built with, and on which version? | `../project/codebase-stack-guide.md` |
+| Which legacy shapes are ratified rather than fixed? | `../project/codebase-brownfield-guide.md` |
 
 - A structure map MUST NOT restate a naming rule, a layer rule, or a version. It MUST reference the
   guide that owns it.

package/kit/.constitution/method/why/README.md

@@ -0,0 +1,169 @@
+---
+status: Reference
+---
+
+# The WDI Method — orientation
+
+**Opened when:** you have never seen this method before, or you have and want the shape back in one reading.
+
+This file **explains**. It does not bind — `../document/*-guide.md` does, and where the two disagree the guide
+wins and the disagreement is a defect to report.
+
+Five minutes. Three files in this folder: this one for orientation, `artifact-map.md` for *"where does this
+file go"*, `rationale.md` for *"why is it like this"* — open that one before changing a rule, so you know
+what you are about to break.
+
+## What the method is
+
+Two methods joined. **WDI** owns the flow, the gates, and the artifacts nobody else writes. **BMad** owns
+the writing skills where it has one. Every BMad skill is invoked through a WDI wrapper, never directly: the
+wrapper is what checks position, verifies the result against the guide, and lands the memlog.
+
+The whole thing rests on one sentence:
+
+> Owner time is spent at five points. Between them, the agents work alone.
+
+## Five gates
+
+A gate is named for **what is decided there**, never for the work before it.
+
+| Gate | Decides | How often | Budget |
+|---|---|---|---|
+| **G1 Problem** | What the problem is, whose it is, why it earns work | once | 20' |
+| **G2 Product** | What is built, and how it feels to use | once per PRD | 45' |
+| **G3 Blueprint** | The whole portrait: which use cases, their entities, tables, endpoints, screens, and the invariants binding them | once per **product** | 45' |
+| **G4 Component** | How one Product Component is built, and what the choice costs | once per **component** | 20–30' |
+| **G5 Release** | Whether it is done and proven | once per wave | 10' |
+
+**Only G4 changes shape.** The other four are always the same, whatever the settings — and that is what lets
+the whole system be held in one head.
+
+## Two settings, and they control different things
+
+| Setting | Where | Controls |
+|---|---|---|
+| `mode` | globally in `index.yaml`, per component in `components.yaml` | **Document depth**, and only that |
+| `risk_accepted` | per component | **Review intensity**, and only that |
+
+`mode` takes `catalog` · `outline` · `guarded` · `deep`, and the default is `catalog`. A component at
+`catalog` **skips G4 entirely** — its control moved to G3, where its use cases, tables, endpoints, screens,
+domain model, and C4 were all approved.
+
+`risk_accepted` takes `low` · `medium` · `high`, and its direction reads off the name: `high` means *"I
+accept a lot of risk here"*, so its review is the lightest.
+
+Keeping them apart is what lets one component be **thin on purpose and reviewed the hardest**. Why that
+matters is in `rationale.md`; what each value demands is in `../document/delivery-flow-guide.md`.
+
+## The run, first time through
+
+| # | Step | Run | Gate |
+|---|---|---|---|
+| 0 | Set up | `wdi-init` intent `setup` — registry scaffolded, global `mode` set, existing documents reported, structure maps derived | — |
+| 1 | Discovery and brief | `wdi-problem` | **G1** |
+| 2 | PRD, one per initiative | `wdi-product` intent `prd` | **G2** |
+| 2b | UX — only when the interface is a large part of the promise | `wdi-ux` | with G2 |
+| 3 | Birth the components, set `mode` and `risk_accepted` | `wdi-init` intent `component` | — (tail of G2) |
+| 4 | Blueprint | `wdi-blueprint` intent `catalog`, then `platform` | **G3** |
+| 5 | One component's depth | `wdi-component` — as deep as its `mode`; **skipped at `catalog`** | **G4** |
+| 6 | Build | `wdi-build` — opens the wave, runs `bmad-spec`, ships each story, closes the wave | **G5** |
+
+After step 6 the next component enters at **step 5**, not at the beginning. Steps 0–4 happen once in the
+life of the product.
+
+`SPEC.md` and story files are **not read by humans**. The human review surface stops at the PRD, `.what/`,
+and `.how/`.
+
+## The run, every time after
+
+| Situation | Run |
+|---|---|
+| The next component is being taken on | `wdi-init` intent `mode` or `risk` if either needs changing → `wdi-component` → **G4** → `wdi-build` → **G5** |
+| That component is at `mode: catalog` | straight to `wdi-build`. G4 is skipped |
+| A promise changes where a PRD already exists | `wdi-product` intent `update` — never a second PRD for the same area |
+| A new initiative with a different reader | `wdi-product` intent `prd` → `wdi-init` intent `component` if it births components |
+| A small fix touching no `FR`, `UC`, `AD-N`, or domain model | Fast Path: `bmad-build` directly. It **stops and becomes a wave `S`** the moment an `FR` is touched |
+| A bug, a failing test, unexpected behaviour | `wdi-systematic-debugging`, **before** any fix is proposed |
+| A planning assumption turned out void | `wdi-decision` — it wraps `bmad-correct-course`, proposes, and changes nothing itself |
+| An estimate or a task list is needed | `wdi-report` intent `estimate` |
+| You do not know where you are | `wdi-help` |
+
+## Fifteen skills
+
+Named for the **gate they serve**, so *"which skill do I run"* is answered by *"which gate am I at"*.
+
+**Moment-bound** — running them outside their point is wrong:
+
+| Skill | Its moment |
+|---|---|
+| `wdi-init` intent `setup` | before G1, once per project |
+| `wdi-problem` | G1 |
+| `wdi-product` | G2 |
+| `wdi-init` intent `component` | tail of G2, and whenever a new PRD births a component |
+| `wdi-blueprint` | G3 |
+| `wdi-component` | G4 |
+| `wdi-build` | G5, one wave per run |
+
+**Anytime** — run the moment the trigger appears, without waiting for a gate:
+
+| Skill | Its trigger |
+|---|---|
+| `wdi-decision` | A decision worth remembering · a void assumption · an accepted decision to carry into documents |
+| `wdi-question` | Something that cannot be decided now |
+| `wdi-log` | A meeting finished, or a non-technical fact now binds |
+| `wdi-help` | "Where am I, what next" |
+| `wdi-reconcile` | Any time. Read-only — it reports, it never edits |
+| `wdi-review` | Over any document, any time |
+| `wdi-systematic-debugging` | A bug, a failed test, a failed build, unexpected behaviour |
+| `wdi-report` | An estimate at the start · progress periodically · before a client update |
+| `wdi-init` intents `mode` · `risk` · `structure` | Any time |
+| `wdi-ux` | Any time after a PRD exists, if UX is being used |
+
+## Who writes what — WDI and BMad
+
+| Artifact | Written by | Wrapped in |
+|---|---|---|
+| Product brief | `bmad-product-brief` | `wdi-problem` |
+| PRD | `bmad-prd` | `wdi-product` |
+| UX | `bmad-ux` | `wdi-ux` |
+| Spine + C4 | `bmad-architecture` | `wdi-blueprint` |
+| **UC catalogue · actors · entities · business rules** | **nothing in BMad** | `wdi-blueprint` writes it itself |
+| **SRS and all of `.what/<pc>/`** | **nothing in BMad** | `wdi-component` writes it itself |
+| **SDD and all of `.how/<pc>/`** | **nothing in BMad** | `wdi-component` writes it itself |
+| `SPEC.md` + stories | `bmad-spec` | `wdi-build` |
+| Code | `bmad-build` · `bmad-build-auto` | `wdi-build` |
+| Retrospective | `bmad-retrospective` | `wdi-build` |
+| Document review | `bmad-review` | `wdi-review` |
+| Course correction | `bmad-correct-course` | `wdi-decision` |
+
+**The bold rows are why this method exists.** BMad stops at the promise and starts again at the mechanism,
+and every behaviour in between had no author. Three consequences stick to those artifacts and are handled
+deliberately: no `doc_standards` fires a review, no memlog is born on its own, and no template enforces
+itself.
+
+## Where things live
+
+| The thing in your hand | Its folder |
+|---|---|
+| How we work — a rule, a guide, a template | `.constitution/` |
+| What currently holds — a decision, a question, a registry, a map | `.control/` |
+| What is promised — the brief, a PRD, a use case, a business rule | `.what/` |
+| How it is built — the spine, C4, an inventory, an SDD, a contract | `.how/` |
+| A skill run's working output | `_bmad-output/` |
+| Scratch that empties when the task closes | `.work/` |
+| The application | `src/` · `web/` |
+
+The test that settles anything ambiguous: **is this file still correct after its wave has passed?** Yes →
+the corpus. No → `_bmad-output/`. In doubt, `../document/corpus-guide.md`.
+
+## Model choice
+
+| Point | Model |
+|---|---|
+| Decisions — proposing a slicing, wording a `DEC-`, preparing a gate | `opus@high` |
+| Writing, derivation, a review-fix pass | `sonnet@high` |
+| Code review panel | Two different CLI families, and never the family that wrote the code |
+
+In a derivation pass, quality comes from the input rather than the model. Running a "find the gap" lens with
+the most careful model produces the most gaps, and each one becomes an open question — a cost nobody sees
+until the question list has stopped being readable.

package/kit/.constitution/method/{artifact-map.md → why/artifact-map.md}

@@ -6,11 +6,11 @@ status: Reference
 
 **Opened when:** someone asks *"where does this file go"*, or *"does this document exist at my `mode`"*.
 
-This file **explains**. It does not bind — `document/*-guide.md` does, and where the two disagree the guide
+This file **explains**. It does not bind — `../document/*-guide.md` does, and where the two disagree the guide
 wins and the disagreement is a defect to report.
 
 It answers three questions and nothing else: which files exist at each `mode`, who owns each one, and how
-the units of work line up. The **rules** about depth live in `document/delivery-flow-guide.md`; what is
+the units of work line up. The **rules** about depth live in `../document/delivery-flow-guide.md`; what is
 here is the map.
 
 ## The one thing to read first
@@ -106,7 +106,7 @@ Repealed: `_platform/architecture/` (one file does not earn a folder) and `<pc>/
 ## Who owns each file
 
 A skill lands the output of the layer it owns, and landing is part of producing it — never a follow-up
-someone else performs. `document/corpus-guide.md` holds the binding version of this table.
+someone else performs. `../document/corpus-guide.md` holds the binding version of this table.
 
 | Owner | Writes |
 |---|---|
@@ -137,7 +137,7 @@ Task** — it travels as a label, because one `FR` can be delivered by stories i
 satisfy part of two `FR`.
 
 The binding version of all of this, including why a wave MAY cross components and what has to be true
-first, is in `document/delivery-flow-guide.md`. It is not restated here.
+first, is in `../document/delivery-flow-guide.md`. It is not restated here.
 
 ## What needs no template, and why
 
@@ -151,5 +151,5 @@ Stated so the next completeness audit does not report it again:
 | `SPEC.md` · story files | They belong to `bmad-spec`. WDI does not write their shape, and MUST NOT |
 | Registry `*.yaml` | Their shape is the comment block at the head of each file, plus the validator |
 
-Everything else in this map has a template in `document/templates/` — 27 of them, and every row above is
+Everything else in this map has a template in `../document/templates/` — 27 of them, and every row above is
 covered by one.

package/kit/.constitution/method/{rationale.md → why/rationale.md}

@@ -6,7 +6,7 @@ status: Reference
 
 **Opened when:** you are about to change a rule and need to know what you would break.
 
-This file **explains**. It does not bind — `document/*-guide.md` does, and where the two disagree the guide
+This file **explains**. It does not bind — `../document/*-guide.md` does, and where the two disagree the guide
 wins and the disagreement is a defect to report. **A rule MUST NOT be born here.** If this file notices a
 rule that should exist, it says so as a finding and the rule is written in its guide.
 
@@ -18,11 +18,11 @@ not finished. Six complaints, their causes, and where each is answered now:
 
 | Complaint | Cause | Answered in |
 |---|---|---|
-| Open questions in floods — ids reached `OQ-146`; one review produced 200 findings, then another produced 90 | Three rules combined into a generator: the `edge-case-hunter` lens mandatory over a whole SRS at once; no route for *"take the assumption, record it, continue"*; no budget and no severity class | `document/decision-guide.md`, `.control/questions/` |
-| Ceremony drift into the PRD; three corrections ended "reported but not fixed" | Correcting a *sentence* of an `FR` was treated as changing its *promise* — both had to go through the PRD skill | `document/prd-guide.md` |
+| Open questions in floods — ids reached `OQ-146`; one review produced 200 findings, then another produced 90 | Three rules combined into a generator: the `edge-case-hunter` lens mandatory over a whole SRS at once; no route for *"take the assumption, record it, continue"*; no budget and no severity class | `../document/decision-guide.md`, `.control/questions/` |
+| Ceremony drift into the PRD; three corrections ended "reported but not fixed" | Correcting a *sentence* of an `FR` was treated as changing its *promise* — both had to go through the PRD skill | `../document/prd-guide.md` |
 | Gates failing repeatedly over unanswered questions | No rule said that only a **blocking** question holds a gate | `.control/questions/blocking.md` |
-| The spine had to be updated every story | Invariants and inventories were mixed, and the tax was paid before the story, when the information was thinnest | `document/architecture-guide.md` |
-| A large PRD made `.what` over-think — 41 of 56 use cases marked `critical` (73%) | Not the PRD's length. The stage's scope was the whole corpus at once, plus a `critical` definition with two elastic criteria that passed everything | `document/delivery-flow-guide.md` |
+| The spine had to be updated every story | Invariants and inventories were mixed, and the tax was paid before the story, when the information was thinnest | `../document/architecture-guide.md` |
+| A large PRD made `.what` over-think — 41 of 56 use cases marked `critical` (73%) | Not the PRD's length. The stage's scope was the whole corpus at once, plus a `critical` definition with two elastic criteria that passed everything | `../document/delivery-flow-guide.md` |
 | Too slow for a small-to-medium application | A composite of the five above | all of it |
 
 One cause was **not** the method: **model choice.** A "find the gap" lens with no upper bound, run by the

package/kit/.constitution/project/constitution.md

@@ -0,0 +1,63 @@
+---
+status: Accepted
+---
+
+# Constitution — {product}
+
+Articles 1, 2, and 5. They are **yours**: seeded once at install, never overwritten by `update`, and
+never carried into the package by `promote`. Rewrite Articles 2 and 5 for the product this repo is;
+Article 1 needs no edit — it cites `product.name` in `.control/registry/index.yaml`.
+
+The method's own articles — 3 Layers, 4 Lifecycle, 6 Decisions, 7 Non-technical facts — live in
+[`../method/constitution.md`](../method/constitution.md) and are replaced on every update. The
+numbering is shared across the two files and has gaps in each.
+
+An agent working here MUST be able to act on the contents of this repo alone.
+
+## Article 1 — Scope
+
+This repo covers the product named at `product.name` in
+`.control/registry/index.yaml`. One product, one repo. A second product MUST
+get a repo of its own.
+
+`product.client` in the same file names the client if there is one, and stays
+empty if there is not. The product brief at G1 uses `product.name` as its
+title. Neither this file nor the brief is a second source of the name.
+
+An agent working here MUST NOT demand that sibling organisation repositories
+be open in the same session.
+
+## Article 2 — Content boundary, `.work/`, and cross-repo references
+
+All three are governed by [`../method/repo-guide.md`](../method/repo-guide.md). Its rules MUST NOT be
+repeated here — one rule, one place.
+
+What is particular to this repo, and therefore lives here:
+
+- Name any extra boundary this product has (a public-repo rule, a ban on `3p.md`,
+  a `.work/` resting state). If there is none, delete this bullet list and the
+  sentence above it.
+- `3p.md` MUST NOT be created in a product repo. Operational engagement memory
+  lives outside.
+
+## Article 5 — The method arrives from WDI Method
+
+This is the **consumer** article. Use it in every product repo.
+
+`.constitution/` guides and templates (except this file's Articles 1, 2, and 5,
+`codebase/*-guide.md`, and any extra file this repo added), the `wdi-*` skills,
+and `_bmad/custom/*.toml` arrive from the public WDI Method package via
+`npx wdi-method install` / `update`.
+
+- A method file MUST NOT be invented or patched here to improve the method. If a
+  rule is wrong, it is fixed in the WDI Method package, then brought here with
+  `update`.
+- `wdi-method update` MUST overwrite method files and MUST NOT touch `.what/`,
+  `.how/`, `.control/` product state, this file's Articles 1–2 and 5,
+  `codebase/*-guide.md` once `Accepted`, extra constitution files this repo
+  added, or `_bmad/custom/*.user.toml`.
+- A rule particular to this repo MUST be written out in full in this file or a
+  sibling, and MUST NOT be replaced by a pointer into another repository.
+
+A prefix in `.claude/skills/` names the **method**, not the owner: `bmad-*` is
+BMad's, `wdi-*` is this method's.

package/kit/assets/bmad-custom/bmad-advanced-elicitation.toml

@@ -1,15 +1,15 @@
-# Skill ini tidak punya key `persistent_facts`, jadi aturan rumah tidak bisa disuntikkan lewat
-# jalur biasa. `preferences` adalah tuas setaranya: kalimat literal yang dipatuhi perefine tiap
-# sesi.
+# This skill has no `persistent_facts` key, so house rules cannot be injected through the usual
+# path. `preferences` is the equivalent lever: literal sentences a reviewer follows every
+# session.
 #
-# Hanya aturan rumah yang mengubah CARA sebuah kritik mendarat yang pantas di sini — metode
-# lengkapnya tetap di .constitution/document/delivery-flow-guide.md.
+# Only a house rule that changes HOW a critique lands belongs here — the full method stays in
+# .constitution/method/document/delivery-flow-guide.md.
 
 [workflow]
 
 preferences = [
-  "Honor AGENTS.md dan .constitution/document/ sebagai aturan yang sedang berlaku di repo ini.",
-  "Nilai artefak G1-G4 terutama dari Decision Summary-nya: satu halaman, bahasa bisnis, tanpa jargon yang belum punya entri Glossary. Ringkasan yang tidak bisa dibaca habis dalam time budget gate-nya adalah temuan itu sendiri.",
-  "Kata normatif hanya lima — MUST, MUST NOT, SHOULD, SHOULD NOT, MAY — dalam bahasa Inggris dan huruf besar; prosanya tetap Bahasa Indonesia. Tandai sebagai cacat: penanda imperatif lain dalam bahasa apa pun, istilah metode yang diterjemahkan, dan karangan padanan Indonesia untuk istilah teknis yang lazim ditulis dalam bahasa Inggris.",
-  "Pushback MUST menyebut keputusan yang sedang dipertaruhkan dan ongkosnya bila salah, bukan sekadar menilai tulisannya.",
+  "Honor AGENTS.md and .constitution/method/document/ as the rules currently in force in this repo.",
+  "Judge a G1-G4 artifact chiefly by its Decision Summary: one page, business language, no jargon that lacks a Glossary entry. A summary that cannot be read in full within its gate's time budget is the finding itself.",
+  "There are only five normative words — MUST, MUST NOT, SHOULD, SHOULD NOT, MAY — in English and uppercase; the surrounding prose follows the product's chosen `doc_language`. Flag as a defect: any other imperative marker in any language, a translated method term, or an invented equivalent for a technical term the industry writes in English.",
+  "Pushback MUST name the decision actually at stake and its cost if wrong, not merely grade the writing.",
 ]

package/kit/assets/bmad-custom/bmad-architecture.toml

@@ -1,31 +1,33 @@
-# Kelas A — altitude initiative saja.
-# Spine memuat invariant yang mencegah unit yang dibangun terpisah jadi divergen.
-# Struktur (stack, tree, data shape) diperlakukan sebagai seed, bukan kontrak.
-#
-# Urutan mengikat: spine lebih dulu, bmad-spec membacanya sebagai companion — bukan sebaliknya.
-
-[workflow]
-
-spine_template     = "{project-root}/.constitution/document/templates/architecture-spine.md"
-spine_output_path  = "{project-root}/.how/_platform"
-run_folder_pattern = "architecture"
-
-# architecture-guide.md MUST NOT dipasang sebagai doc_standards, meski itu tempat lazimnya
-# bagi guide dokumen. Skill ini mengecualikan spine dari doc_standards secara sengaja —
-# customize.toml bawaannya menyebut "never to the spine", dan langkah 5 SKILL.md mengulanginya:
-# "apply {workflow.doc_standards} polish to that prose only, never to the spine".
-# Dipasang di sana, aturan AD-N cuma memoles deck dan C4 — dokumen yang tidak punya AD-N —
-# sementara spine ditulis tanpa guide-nya. Karena itu ia jadi fakta yang dipegang sejak
-# aktivasi, bukan poles di akhir.
-#
-# doc_standards sengaja tidak diisi. Default BMad `skill:bmad-review lenses=structure,prose`
-# tetap berlaku untuk rendering, dan array ini append — mengosongkannya tidak mencabut default.
-#
-# Akibatnya spine tidak punya review otomatis. `bmad-review` atasnya MUST dipanggil manual
-# sebelum G4 — lihat tabel "Siapa memanggil bmad-review" di 05-gates.md.
-
-persistent_facts = [
-  "file:{project-root}/.constitution/document/architecture-guide.md",
-  "file:{project-root}/.constitution/document/decision-guide.md",
-  "Seluruh panggilan memlog.py MUST memakai --path {project-root}/.control/memlog/spine.md. --workspace MUST NOT dipakai.",
-]
+# Class A — initiative altitude only.
+# The spine holds the invariants that keep separately built units from diverging.
+# Structure (stack, tree, data shapes) is treated as a seed, not a contract.
+#
+# Order binds: the spine comes first, and bmad-spec reads it as a companion — never the other
+# way round.
+
+[workflow]
+
+spine_template     = "{project-root}/.constitution/method/document/templates/architecture-spine.md"
+spine_output_path  = "{project-root}/.how/_platform"
+run_folder_pattern = "architecture"
+
+# architecture-guide.md MUST NOT be installed as doc_standards, even though that is the usual
+# place for a document guide. This skill deliberately excludes the spine from doc_standards —
+# its own default customize.toml says "never to the spine", and SKILL.md step 5 repeats it:
+# "apply {workflow.doc_standards} polish to that prose only, never to the spine".
+# Installed there, the AD-N rule would only polish the deck and C4 — documents that carry no
+# AD-N — while the spine itself gets written with no guide at all. So it is installed as a
+# persistent fact, held from activation, rather than as end-of-run polish.
+#
+# doc_standards is deliberately left empty. BMad's own default
+# `skill:bmad-review lenses=structure,prose` still applies to rendering, and this array appends —
+# emptying it does not remove the default.
+#
+# The consequence is that the spine gets no automatic review. `bmad-review` on it MUST be called
+# manually before G4 — see the "Who calls bmad-review" table in 05-gates.md.
+
+persistent_facts = [
+  "file:{project-root}/.constitution/method/document/architecture-guide.md",
+  "file:{project-root}/.constitution/method/document/decision-guide.md",
+  "Every memlog.py call MUST use --path {project-root}/.control/memlog/spine.md. --workspace MUST NOT be used.",
+]

package/kit/assets/bmad-custom/bmad-build-auto.toml

@@ -1,10 +1,10 @@
-# Satu iterasi loop tanpa interaksi manusia. Dispatch lewat {spec_folder} + {story_id}.
+# One loop iteration with no human in it. Dispatched through {spec_folder} + {story_id}.
 
 [workflow]
 
 persistent_facts = [
-  "Repo ini berjalan di stories mode. sprint-status.yaml tidak ada dan MUST NOT dibuat.",
-  "Disiplin test-first mengikat: test MUST gagal lebih dulu, dan MUST gagal karena alasan yang diharapkan.",
-  "Worker MUST NOT mengubah apa pun di luar story yang didispatch. Perubahan yang dinilai perlu di luar itu MUST dinyatakan di hasil akhir, tidak pernah dilakukan diam-diam.",
-  "Saat test atau build gagal dan sebabnya belum diketahui, wdi-systematic-debugging MUST dijalankan sebelum perbaikan apa pun diusulkan. Percobaan perbaikan ketiga yang gagal adalah tanda untuk berhenti dan melapor, bukan untuk mencoba keempat.",
+  "This repo runs in stories mode. sprint-status.yaml does not exist and MUST NOT be created.",
+  "Test-first discipline binds: a test MUST fail first, and MUST fail for the expected reason.",
+  "A worker MUST NOT change anything outside the story it was dispatched. A change it judges necessary beyond that MUST be stated in the final report, never made silently.",
+  "When a test or build fails and the cause is not yet known, wdi-systematic-debugging MUST run before any fix is proposed. A third failed fix attempt is the signal to stop and report, not to try a fourth.",
 ]

package/kit/assets/bmad-custom/bmad-build.toml

@@ -1,52 +1,52 @@
-# G5 — satu-satunya pelaksana, dan wdi-build yang mendispatch-nya. Pelaksana MUST NOT merakit
-# rantai sendiri.
-#
-# Memasok disiplin test-first yang bmad-build tidak punya: ia mengaudit test SETELAH implementasi
-# (Matrix Test Audit di step-03), tetapi tidak pernah mewajibkan sebuah test gagal lebih dulu.
-#
-# Aturan merge: string mengganti, list menambah, tabel merge per key.
-# .claude/skills/bmad-build/customize.toml MUST NOT diedit — ia ditimpa tiap update.
-
-[workflow]
-
-# Mengganti handoff step-03 bawaan. Semua yang dibawa bawaan tetap ada (subagent tanpa konteks
-# percakapan, spec sebagai satu-satunya sumber kebenaran, memuat berkas `context:`, laporan
-# penutup); siklus test-first inilah yang ditambahkan override ini.
-#
-# Sengaja memuat METODE saja. Pernyataan ulang tujuan, daftar berkas, acceptance criteria, dan
-# aturan gaya rumah TIDAK masuk dispatch — semuanya sudah ada di spec, dan step-03 melarang
-# menggandakannya di sini.
-
-implementation_handoff = """
-Launch a subagent with no prior conversation context, with this prompt:
-
-> Read {spec_file} fully and implement it — the spec is the sole source of truth. Load every file listed in its frontmatter `context:` before you start.
->
-> Work test-first, one task at a time:
->
-> 1. Write the smallest test that expresses the next unmet acceptance criterion.
-> 2. Run it. Confirm it fails, and confirm it fails for the reason you expect — a test that errors on a typo or a missing import has proven nothing.
-> 3. Write the minimum code that makes it pass. No extra scope, no "while I'm here" refactoring.
-> 4. Run the test plus the surrounding suite. Both must be green before you start the next task.
->
-> If a criterion genuinely cannot be expressed as a test — pure configuration, a docs change, a visual-only tweak — say so explicitly in your report and state how you verified it instead. Never invent a test that asserts nothing in order to satisfy the cycle, and never write the implementation first and the test afterwards to make it look followed.
->
-> When done, report what you changed, how you verified it, and anything left incomplete or risky. Include the actual command output for the final test run — a claim of green without the output does not count as verification.
-"""
-
-persistent_facts = [
-  "Repo ini berjalan di stories mode. sprint-status.yaml tidak ada dan MUST NOT dibuat; status story tinggal di frontmatter story file.",
-  "Satu Logical Component MUST terdaftar di components.yaml sebelum story yang menyentuhnya dikerjakan.",
-  "Saat ada yang rusak dan sebabnya belum diketahui, wdi-systematic-debugging MUST dijalankan sebelum perbaikan apa pun diusulkan.",
-]
-
-# Tidak dioverride, dan itu disengaja:
-#
-# - review_layers — tiga layer bawaan (blind-hunter, edge-case-hunter, verification-gap) plus
-#   Review Panel dua keluarga CLI yang dijalankan bmad-code-review sesudahnya sudah menutup
-#   kebutuhan review repo ini. Menambah layer keempat memakan token tanpa menutup celah.
-#
-# - on_complete — pembaruan registry setelah story selesai adalah tugas wdi-build dan skill pemilik
-#   lapisannya,
-#   bukan bmad-build. Menaruhnya di sini membuat dua penulis untuk satu berkas.
-#   Checklist penutup story ada di 05-gates.md.
+# G5 — the only executor, dispatched by wdi-build. An executor MUST NOT assemble its own chain.
+#
+# Supplies the test-first discipline bmad-build lacks on its own: it audits tests AFTER
+# implementation (Matrix Test Audit in step-03), but never requires a test to fail first.
+#
+# Merge rule: a string replaces, a list appends, a table merges per key.
+# .claude/skills/bmad-build/customize.toml MUST NOT be edited — it is overwritten on every update.
+
+[workflow]
+
+# Replaces the default step-03 handoff. Everything the default carries stays (a subagent with no
+# conversation context, the spec as the sole source of truth, loading the files listed in
+# `context:`, a closing report); this override adds the test-first cycle on top.
+
+#
+# Deliberately carries METHOD only. Restating the goal, the file list, the acceptance criteria,
+# and house style rules does NOT belong in the dispatch — all of it is already in the spec, and
+# step-03 forbids duplicating it here.
+
+implementation_handoff = """
+Launch a subagent with no prior conversation context, with this prompt:
+
+> Read {spec_file} fully and implement it — the spec is the sole source of truth. Load every file listed in its frontmatter `context:` before you start.
+>
+> Work test-first, one task at a time:
+>
+> 1. Write the smallest test that expresses the next unmet acceptance criterion.
+> 2. Run it. Confirm it fails, and confirm it fails for the reason you expect — a test that errors on a typo or a missing import has proven nothing.
+> 3. Write the minimum code that makes it pass. No extra scope, no "while I'm here" refactoring.
+> 4. Run the test plus the surrounding suite. Both must be green before you start the next task.
+>
+> If a criterion genuinely cannot be expressed as a test — pure configuration, a docs change, a visual-only tweak — say so explicitly in your report and state how you verified it instead. Never invent a test that asserts nothing in order to satisfy the cycle, and never write the implementation first and the test afterwards to make it look followed.
+>
+> When done, report what you changed, how you verified it, and anything left incomplete or risky. Include the actual command output for the final test run — a claim of green without the output does not count as verification.
+"""
+
+persistent_facts = [
+  "This repo runs in stories mode. sprint-status.yaml does not exist and MUST NOT be created; story status lives in the story file's own frontmatter.",
+  "A Logical Component MUST be registered in components.yaml before the story touching it is worked.",
+  "When something breaks and the cause is not yet known, wdi-systematic-debugging MUST run before any fix is proposed.",
+]
+
+# Not overridden, and deliberately so:
+#
+# - review_layers — the three default layers (blind-hunter, edge-case-hunter,
+#   verification-gap) plus the two-CLI-family Review Panel that bmad-code-review runs afterward
+#   already cover this repo's review needs. A fourth layer would spend tokens without closing a
+#   gap.
+#
+# - on_complete — updating the registry once a story is done belongs to wdi-build and the skill
+#   that owns that layer, not to bmad-build. Putting it here would give one file two writers.
+#   The story-closing checklist lives in 05-gates.md.

package/kit/assets/bmad-custom/bmad-code-review.toml

@@ -1,12 +1,13 @@
-# Review Panel: dua peninjau dari dua keluarga CLI berbeda, dan tidak boleh dari keluarga yang
-# menjalankan implementasinya.
-#
-# PERINGATAN: lensa adversarial dan edge-case di sini dibaca dari salinan bundel review-prompts/,
-# BUKAN dari skill bmad-review. Override pada bmad-review tidak akan terlihat di jalur ini.
-
-[workflow]
-
-persistent_facts = [
-  "file:{project-root}/.constitution/codebase/conventions-guide.md",
-  "Temuan yang berulang lintas story MUST dicatat lewat bmad-project-context intent record, bukan diperbaiki diam-diam tiap kali muncul.",
-]
+# Review Panel: two reviewers from two different CLI families, and neither may come from the
+# family that ran the implementation.
+#
+# WARNING: the adversarial and edge-case lenses here are read from the bundled copy of
+# review-prompts/, NOT from the bmad-review skill. An override on bmad-review will not be visible
+# on this path.
+
+[workflow]
+
+persistent_facts = [
+  "file:{project-root}/.constitution/project/codebase-conventions-guide.md",
+  "A finding that repeats across stories MUST be recorded through bmad-project-context intent record, not fixed silently every time it recurs.",
+]