slipwai 1.1.1.dev4__tar.gz → 1.2.0.dev19__tar.gz

{slipwai-1.1.1.dev4 → slipwai-1.2.0.dev19}/NOTICE

@@ -18,9 +18,11 @@ notices:
 - event-modeling — current adaptation under the root MIT licence; its NOTICE
   records the CC0-1.0 starting point
 - find-skills — MIT; copyright Vercel, Inc.
+- frontend-design — Apache-2.0; copyright Anthropic, PBC
 - improve-codebase-architecture — MIT; copyright Matt Pocock
 - reduce-system-complexity — MIT; copyright Adam Bulmer
 - technical-writing — MIT; copyright Adam Bulmer
+- web-interface-guidelines — MIT (references/guidelines.md); copyright Vercel, Inc.
 - wtf — MIT; copyright Adam Bulmer
 
 The applicable licence notice and exact provenance for each adaptation remain

{slipwai-1.1.1.dev4 → slipwai-1.2.0.dev19}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: slipwai
-Version: 1.1.1.dev4
+Version: 1.2.0.dev19
 Summary: One-shot scaffolder for product monorepos: a walking skeleton, a verify gate, a path to production, and a repository set up for coding agents.
 Project-URL: Homepage, https://github.com/ROBCOATVG/slipwai
 Project-URL: Repository, https://git.treyco.dev/ROBCOATVG/slipwai
@@ -18,9 +18,11 @@ License-File: assets/toolkit/skills/codebase-design/LICENSE
 License-File: assets/toolkit/skills/diagrams/NOTICE
 License-File: assets/toolkit/skills/event-modeling/NOTICE
 License-File: assets/toolkit/skills/find-skills/LICENSE
+License-File: assets/toolkit/skills/frontend-design/LICENSE
 License-File: assets/toolkit/skills/improve-codebase-architecture/LICENSE
 License-File: assets/toolkit/skills/reduce-system-complexity/LICENSE
 License-File: assets/toolkit/skills/technical-writing/LICENSE
+License-File: assets/toolkit/skills/web-interface-guidelines/LICENSE
 License-File: assets/toolkit/skills/wtf/LICENSE
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown
@@ -197,8 +199,8 @@ Spec Kit is deliberately absent until you ask for it. From inside the generated
 ```
 
 That installs Spec Kit. It then asks which coding agent to project `skills/`, `commands/` and `agents/`
-into — any of [36 harnesses](docs/spec-kit.md). It then offers optional extensions as a checkbox menu, of
-which CodeGraph is the first.
+into — any of [36 harnesses](docs/spec-kit.md). It then offers optional extensions as a checkbox menu:
+CodeGraph, UI/UX Pro Max and the UX gates.
 
 `--integration <name>` and `--extension <key>` skip those questions. `--extension` also adds one later. See
 [Extensions](docs/extensions.md).
@@ -215,6 +217,11 @@ make demo        # the whole thing in containers, printing the addresses once it
 Then open an agent session and type `/drive`. It walks the ladder from principles to a demo the actor can
 see, and enters at the first stage whose artifact is missing. See [The delivery loop](docs/delivery-loop.md).
 
+`/drive` stops when it needs a person: for a product decision, and at every demo. `/cruise` runs the same
+ladder and answers those stops itself. It decides as the product owner, runs each demo as the actor, and
+writes every answer down where you can read it and overturn it. It keeps going until the specification is
+satisfied, and it stops only for you. It ships switched off. See [Cruise](docs/cruise.md).
+
 ---
 
 ## What you get
@@ -226,8 +233,8 @@ Everything below is in a generated repository from its first commit.
 
 | | |
 |---|---|
-| **[The delivery loop](docs/delivery-loop.md)** | `/drive`'s ten-stage ladder on top of Spec Kit. It is resumable, because it reads artifacts rather than conversation memory. Fifteen workflow commands, thirteen of them in `standard`. Story splitting and example mapping are first-class stages with real heuristics behind them. Hooks apply the method even to a session that never typed `/drive` |
-| **[Up to 49 skills](docs/skills.md)** | The part of the delivery catalogue this project can use, owned by the project: TDD, testing, hexagonal architecture, DDD, ubiquitous language, API and BFF design, observability, secure OAuth/OIDC, refactoring, debugging and more. Code examples are rendered in the languages this project's services are actually written in |
+| **[The delivery loop](docs/delivery-loop.md)** | `/drive`'s ten-stage ladder on top of Spec Kit. It is resumable, because it reads artifacts rather than conversation memory. Seventeen workflow commands, fifteen of them in `standard`. Story splitting and example mapping are first-class stages with real heuristics behind them. Hooks apply the method even to a session that never typed `/drive` |
+| **[Up to 51 skills](docs/skills.md)** | The part of the delivery catalogue this project can use, owned by the project: TDD, testing, hexagonal architecture, DDD, ubiquitous language, API and BFF design, observability, secure OAuth/OIDC, refactoring, debugging and more. Code examples are rendered in the languages this project's services are actually written in |
 | **[The global event model](docs/event-model.md)** | *Event profile.* One cumulative model for the whole system, in `model.yaml`. `make model` renders the timeline, the per-segment diagrams and a self-contained browsable page. `make check-model` fails when the code and the model disagree |
 | **[The read side](docs/what-you-get.md#the-read-side)** | *Event profile.* The machinery a view is *maintained* with, finished in every backend rather than left as the greenfield half: a unit of work on the event store; a `CheckpointStore` port with memory, SQLite and Postgres adapters behind a contract suite of its own; a catch-up runner that advances the checkpoint inside the view's own transaction; a rebuild; whatever each framework already schedules a pass with; and a tag index derived from the log, so a conditional append can hold a boundary one stream cannot |
 | **[Gates](docs/verification.md)** | `make verify` runs native checks, tests, architecture direction, agent and Spec Kit drift, constitution coverage, and the event model. It is identical locally and in CI, and never needs Docker. Integration, adversarial, mutation and audit checks sit deliberately outside it |
@@ -236,17 +243,18 @@ Everything below is in a generated repository from its first commit.
 | **[Generated documentation](docs/what-you-get.md#documentation-written-for-this-project)** | Thirteen pages written for this project's actual shape rather than copied, and an index built from the files that shipped |
 | **[Services and bounded contexts](docs/services.md)** | One list of applications, each with its own language, framework and answers. `add-service` and `add-frontend` grow it. Contexts are found rather than declared |
 | **[A path to production](docs/aws-target.md)** | `--target aws` or `--target azure` gives you `infra/` in OpenTofu, one deployable per application released blue/green, a pipeline from every push to `main` through staging to production, and a one-command rollback. The same promise on either cloud. [The Azure page](docs/azure-target.md) is where they are compared |
-| **[Extensions](docs/extensions.md)** | Optional dev tooling, adopted with `./init --extension <key>`. CodeGraph is the first. None of them changes the generated skeleton's code |
+| **[Extensions](docs/extensions.md)** | Optional dev tooling, adopted with `./init --extension <key>`: a code index, a design-system generator, and objective UX gates in `make verify`. None of them changes the generated skeleton's code |
 
 ### The commands a project gets
 
-Fifteen, of which thirteen are in `standard`. A repository that adopted the method gets four more. None of
+Seventeen, of which fifteen are in `standard`. A repository that adopted the method gets four more. None of
 them is copied: each is generated against this project's profile, languages, frameworks and services, so
 the paths and toolchains named in them are real.
 
 | Command | What it does |
 |---|---|
 | `/drive` | Takes one slice from wherever it stands to a demo the actor can see |
+| `/cruise` | Runs `/drive` on its own, deciding as the product owner and demoing as the actor, until the specification is satisfied. Stops only for a person |
 | `/whats-next` | Says what is next — one slice, one stage, one command. Reads the disk and changes nothing |
 | `/where-are-we` | Shows the progress board: what works, what is in progress, what is still to come |
 | `/gaps` | Reviews an artifact for holes, before they become rewritten tests |
@@ -261,14 +269,15 @@ the paths and toolchains named in them are real.
 | `/constitution-coverage` | Checks the constitution against the floor this project depends on, or prints the floor |
 | `/drive-settings` | Shows or changes how `/drive` hands implementation to a delegate |
 | `/model-delegation-settings` | Shows or changes which model runs each stage of `/drive` |
+| `/cruise-settings` | Shows or changes how `/cruise` runs: who decides, how it releases, what it demos with, when it parks |
 
 [The delivery loop](docs/delivery-loop.md) describes each one in full, and the ladder they sit on.
 
-#### The two that change settings
+#### The three that change settings
 
-Run either with no arguments to see the current setting. Pass arguments to change it. Both refuse a value
-they do not know and write nothing, and both ask you to commit the settings file on its own — the choice is
-versioned with the project.
+Run any of them with no arguments to see the current setting. Pass arguments to change it. Each refuses a
+value it does not know and writes nothing, and each asks you to commit the settings file on its own — the
+choice is versioned with the project.
 
 **`/drive-settings`** — how much one implementation delegate is handed, and how many failing tests a cycle
 opens with. It writes `.specify/drive.json`.
@@ -296,6 +305,20 @@ what a role runs on. Changing it also rewrites the agent types, so the harnesses
 
 The change takes effect at the next stage `/drive` runs. Nothing already running is interrupted.
 
+**`/cruise-settings`** — how `/cruise` runs `/drive` with nobody at the wheel. It writes `.specify/cruise.json`.
+
+```
+/cruise-settings                            # show every setting
+/cruise-settings enabled=true               # switch it on; it ships switched off
+/cruise-settings release=park               # stop before every push and ask a person
+/cruise-settings max_hours=8                # give a run a budget
+```
+
+`enabled` switches the command on. `decide` says who answers a product question. `release` says whether every
+merge stays behind a flag or waits for a person. `hand` says what a demo is driven with, a browser first.
+The rest bound the run: when it counts as stuck, how many iterations or hours it may take, how often a parked
+run looks for a reason to resume. [Cruise](docs/cruise.md) explains each one.
+
 ---
 
 ## The answers you give
@@ -313,7 +336,7 @@ Beyond the profile, each role is a separate question, answered independently:
 | HTTP transport | `--http` | `none`, `fastify`, `fastapi`, `net-http`, `quarkus-rest`, `spring-web` |
 | Staff authentication | `--auth` | `none`, `keycloak` (local only), `cognito` (AWS), `entra` (Azure), `auth0` (either cloud) |
 | Customer authentication | `--users` | `none`, `keycloak` (local only), `cognito` (AWS), `auth0` (either cloud) |
-| Dev tooling | `--extension` | `codegraph` |
+| Dev tooling | `--extension` | `codegraph`, `uipro`, `ux-gates` |
 
 The default is `event-modelling` with `typescript`, the `react-vite` frontend, a Postgres event store, and
 the HTTP transport that backend has.
@@ -351,8 +374,9 @@ is which, what each answer brings, and how a project answers a question again la
 | Document | Covers |
 |---|---|
 | [What a generated repository gets for free](docs/what-you-get.md) | The tour: the layout, the generated documentation, the agent harnesses, and how to run it |
-| [The delivery loop](docs/delivery-loop.md) | The `/drive` ladder and the diagram behind it, the fifteen commands, story splitting and example mapping with worked examples, the Spec Kit hooks, the preset layer, and the constitution floor |
-| [The skill catalogue](docs/skills.md) | The 49 skills grouped by what they are for, why a project is given only the ones whose subject it has, how examples are rendered in your own languages, and where to edit them |
+| [The delivery loop](docs/delivery-loop.md) | The `/drive` ladder and the diagram behind it, the seventeen commands, story splitting and example mapping with worked examples, the Spec Kit hooks, the preset layer, and the constitution floor |
+| [Cruise](docs/cruise.md) | `/cruise`: the same ladder with nobody at the wheel — the agent decides as the product owner and runs each demo as the actor, writes every answer where a person can overturn it, and stops only for a human; how to start, watch and stop a run, the settings, and the limits |
+| [The skill catalogue](docs/skills.md) | The 51 skills grouped by what they are for, why a project is given only the ones whose subject it has, how examples are rendered in your own languages, and where to edit them |
 | [The global event model](docs/event-model.md) | Why the model is global, what `make model` renders, the status ladder `make check-model` enforces, and how the browsable page is published |
 | [Gates](docs/verification.md) | What `make verify` runs, what is deliberately outside it, and why the split falls where it does |
 | [Bootstrap Spec Kit](docs/spec-kit.md) | `./init`, the 36 agent integrations, how Spec Kit is obtained, and the constitution floor gated from both sides |

{slipwai-1.1.1.dev4 → slipwai-1.2.0.dev19}/README.md

@@ -170,8 +170,8 @@ Spec Kit is deliberately absent until you ask for it. From inside the generated
 ```
 
 That installs Spec Kit. It then asks which coding agent to project `skills/`, `commands/` and `agents/`
-into — any of [36 harnesses](docs/spec-kit.md). It then offers optional extensions as a checkbox menu, of
-which CodeGraph is the first.
+into — any of [36 harnesses](docs/spec-kit.md). It then offers optional extensions as a checkbox menu:
+CodeGraph, UI/UX Pro Max and the UX gates.
 
 `--integration <name>` and `--extension <key>` skip those questions. `--extension` also adds one later. See
 [Extensions](docs/extensions.md).
@@ -188,6 +188,11 @@ make demo        # the whole thing in containers, printing the addresses once it
 Then open an agent session and type `/drive`. It walks the ladder from principles to a demo the actor can
 see, and enters at the first stage whose artifact is missing. See [The delivery loop](docs/delivery-loop.md).
 
+`/drive` stops when it needs a person: for a product decision, and at every demo. `/cruise` runs the same
+ladder and answers those stops itself. It decides as the product owner, runs each demo as the actor, and
+writes every answer down where you can read it and overturn it. It keeps going until the specification is
+satisfied, and it stops only for you. It ships switched off. See [Cruise](docs/cruise.md).
+
 ---
 
 ## What you get
@@ -199,8 +204,8 @@ Everything below is in a generated repository from its first commit.
 
 | | |
 |---|---|
-| **[The delivery loop](docs/delivery-loop.md)** | `/drive`'s ten-stage ladder on top of Spec Kit. It is resumable, because it reads artifacts rather than conversation memory. Fifteen workflow commands, thirteen of them in `standard`. Story splitting and example mapping are first-class stages with real heuristics behind them. Hooks apply the method even to a session that never typed `/drive` |
-| **[Up to 49 skills](docs/skills.md)** | The part of the delivery catalogue this project can use, owned by the project: TDD, testing, hexagonal architecture, DDD, ubiquitous language, API and BFF design, observability, secure OAuth/OIDC, refactoring, debugging and more. Code examples are rendered in the languages this project's services are actually written in |
+| **[The delivery loop](docs/delivery-loop.md)** | `/drive`'s ten-stage ladder on top of Spec Kit. It is resumable, because it reads artifacts rather than conversation memory. Seventeen workflow commands, fifteen of them in `standard`. Story splitting and example mapping are first-class stages with real heuristics behind them. Hooks apply the method even to a session that never typed `/drive` |
+| **[Up to 51 skills](docs/skills.md)** | The part of the delivery catalogue this project can use, owned by the project: TDD, testing, hexagonal architecture, DDD, ubiquitous language, API and BFF design, observability, secure OAuth/OIDC, refactoring, debugging and more. Code examples are rendered in the languages this project's services are actually written in |
 | **[The global event model](docs/event-model.md)** | *Event profile.* One cumulative model for the whole system, in `model.yaml`. `make model` renders the timeline, the per-segment diagrams and a self-contained browsable page. `make check-model` fails when the code and the model disagree |
 | **[The read side](docs/what-you-get.md#the-read-side)** | *Event profile.* The machinery a view is *maintained* with, finished in every backend rather than left as the greenfield half: a unit of work on the event store; a `CheckpointStore` port with memory, SQLite and Postgres adapters behind a contract suite of its own; a catch-up runner that advances the checkpoint inside the view's own transaction; a rebuild; whatever each framework already schedules a pass with; and a tag index derived from the log, so a conditional append can hold a boundary one stream cannot |
 | **[Gates](docs/verification.md)** | `make verify` runs native checks, tests, architecture direction, agent and Spec Kit drift, constitution coverage, and the event model. It is identical locally and in CI, and never needs Docker. Integration, adversarial, mutation and audit checks sit deliberately outside it |
@@ -209,17 +214,18 @@ Everything below is in a generated repository from its first commit.
 | **[Generated documentation](docs/what-you-get.md#documentation-written-for-this-project)** | Thirteen pages written for this project's actual shape rather than copied, and an index built from the files that shipped |
 | **[Services and bounded contexts](docs/services.md)** | One list of applications, each with its own language, framework and answers. `add-service` and `add-frontend` grow it. Contexts are found rather than declared |
 | **[A path to production](docs/aws-target.md)** | `--target aws` or `--target azure` gives you `infra/` in OpenTofu, one deployable per application released blue/green, a pipeline from every push to `main` through staging to production, and a one-command rollback. The same promise on either cloud. [The Azure page](docs/azure-target.md) is where they are compared |
-| **[Extensions](docs/extensions.md)** | Optional dev tooling, adopted with `./init --extension <key>`. CodeGraph is the first. None of them changes the generated skeleton's code |
+| **[Extensions](docs/extensions.md)** | Optional dev tooling, adopted with `./init --extension <key>`: a code index, a design-system generator, and objective UX gates in `make verify`. None of them changes the generated skeleton's code |
 
 ### The commands a project gets
 
-Fifteen, of which thirteen are in `standard`. A repository that adopted the method gets four more. None of
+Seventeen, of which fifteen are in `standard`. A repository that adopted the method gets four more. None of
 them is copied: each is generated against this project's profile, languages, frameworks and services, so
 the paths and toolchains named in them are real.
 
 | Command | What it does |
 |---|---|
 | `/drive` | Takes one slice from wherever it stands to a demo the actor can see |
+| `/cruise` | Runs `/drive` on its own, deciding as the product owner and demoing as the actor, until the specification is satisfied. Stops only for a person |
 | `/whats-next` | Says what is next — one slice, one stage, one command. Reads the disk and changes nothing |
 | `/where-are-we` | Shows the progress board: what works, what is in progress, what is still to come |
 | `/gaps` | Reviews an artifact for holes, before they become rewritten tests |
@@ -234,14 +240,15 @@ the paths and toolchains named in them are real.
 | `/constitution-coverage` | Checks the constitution against the floor this project depends on, or prints the floor |
 | `/drive-settings` | Shows or changes how `/drive` hands implementation to a delegate |
 | `/model-delegation-settings` | Shows or changes which model runs each stage of `/drive` |
+| `/cruise-settings` | Shows or changes how `/cruise` runs: who decides, how it releases, what it demos with, when it parks |
 
 [The delivery loop](docs/delivery-loop.md) describes each one in full, and the ladder they sit on.
 
-#### The two that change settings
+#### The three that change settings
 
-Run either with no arguments to see the current setting. Pass arguments to change it. Both refuse a value
-they do not know and write nothing, and both ask you to commit the settings file on its own — the choice is
-versioned with the project.
+Run any of them with no arguments to see the current setting. Pass arguments to change it. Each refuses a
+value it does not know and writes nothing, and each asks you to commit the settings file on its own — the
+choice is versioned with the project.
 
 **`/drive-settings`** — how much one implementation delegate is handed, and how many failing tests a cycle
 opens with. It writes `.specify/drive.json`.
@@ -269,6 +276,20 @@ what a role runs on. Changing it also rewrites the agent types, so the harnesses
 
 The change takes effect at the next stage `/drive` runs. Nothing already running is interrupted.
 
+**`/cruise-settings`** — how `/cruise` runs `/drive` with nobody at the wheel. It writes `.specify/cruise.json`.
+
+```
+/cruise-settings                            # show every setting
+/cruise-settings enabled=true               # switch it on; it ships switched off
+/cruise-settings release=park               # stop before every push and ask a person
+/cruise-settings max_hours=8                # give a run a budget
+```
+
+`enabled` switches the command on. `decide` says who answers a product question. `release` says whether every
+merge stays behind a flag or waits for a person. `hand` says what a demo is driven with, a browser first.
+The rest bound the run: when it counts as stuck, how many iterations or hours it may take, how often a parked
+run looks for a reason to resume. [Cruise](docs/cruise.md) explains each one.
+
 ---
 
 ## The answers you give
@@ -286,7 +307,7 @@ Beyond the profile, each role is a separate question, answered independently:
 | HTTP transport | `--http` | `none`, `fastify`, `fastapi`, `net-http`, `quarkus-rest`, `spring-web` |
 | Staff authentication | `--auth` | `none`, `keycloak` (local only), `cognito` (AWS), `entra` (Azure), `auth0` (either cloud) |
 | Customer authentication | `--users` | `none`, `keycloak` (local only), `cognito` (AWS), `auth0` (either cloud) |
-| Dev tooling | `--extension` | `codegraph` |
+| Dev tooling | `--extension` | `codegraph`, `uipro`, `ux-gates` |
 
 The default is `event-modelling` with `typescript`, the `react-vite` frontend, a Postgres event store, and
 the HTTP transport that backend has.
@@ -324,8 +345,9 @@ is which, what each answer brings, and how a project answers a question again la
 | Document | Covers |
 |---|---|
 | [What a generated repository gets for free](docs/what-you-get.md) | The tour: the layout, the generated documentation, the agent harnesses, and how to run it |
-| [The delivery loop](docs/delivery-loop.md) | The `/drive` ladder and the diagram behind it, the fifteen commands, story splitting and example mapping with worked examples, the Spec Kit hooks, the preset layer, and the constitution floor |
-| [The skill catalogue](docs/skills.md) | The 49 skills grouped by what they are for, why a project is given only the ones whose subject it has, how examples are rendered in your own languages, and where to edit them |
+| [The delivery loop](docs/delivery-loop.md) | The `/drive` ladder and the diagram behind it, the seventeen commands, story splitting and example mapping with worked examples, the Spec Kit hooks, the preset layer, and the constitution floor |
+| [Cruise](docs/cruise.md) | `/cruise`: the same ladder with nobody at the wheel — the agent decides as the product owner and runs each demo as the actor, writes every answer where a person can overturn it, and stops only for a human; how to start, watch and stop a run, the settings, and the limits |
+| [The skill catalogue](docs/skills.md) | The 51 skills grouped by what they are for, why a project is given only the ones whose subject it has, how examples are rendered in your own languages, and where to edit them |
 | [The global event model](docs/event-model.md) | Why the model is global, what `make model` renders, the status ladder `make check-model` enforces, and how the browsable page is published |
 | [Gates](docs/verification.md) | What `make verify` runs, what is deliberately outside it, and why the split falls where it does |
 | [Bootstrap Spec Kit](docs/spec-kit.md) | `./init`, the 36 agent integrations, how Spec Kit is obtained, and the constitution floor gated from both sides |

slipwai-1.2.0.dev19/VERSION

@@ -0,0 +1 @@
+1.2.0.dev19

{slipwai-1.1.1.dev4 → slipwai-1.2.0.dev19}/assets/frontends/react-vite/app/src/styles/base.css

@@ -144,8 +144,8 @@ a {
 }
 
 :focus-visible {
-  outline: 2px solid var(--colour-accent);
-  outline-offset: 2px;
+  outline: var(--focus-ring);
+  outline-offset: var(--focus-ring-offset);
 }
 
 button,

{slipwai-1.1.1.dev4 → slipwai-1.2.0.dev19}/assets/frontends/react-vite/app/src/styles/tokens.css

@@ -53,6 +53,9 @@
   --radius-large: 10px;
   --border: 1px solid var(--colour-border);
   --shadow: 0 1px 2px rgb(22 25 29 / 8%);
+  /* The one ring every focusable element shows: a token so no rule below carries a literal width. */
+  --focus-ring: 2px solid var(--colour-accent);
+  --focus-ring-offset: 2px;
 
   /* The reading column. Wider than this and prose stops being readable. */
   --measure: 68ch;

{slipwai-1.1.1.dev4 → slipwai-1.2.0.dev19}/assets/toolkit/docs/event-model/README.md

@@ -7,7 +7,9 @@ when the model and the repository disagree. Everything else in this directory is
 
 ```bash
 make model         # regenerate the diagram and the browsable page (PNG=1 also writes a raster copy)
-make check-model   # validate the model and prove the committed diagram is current — runs in CI
+make check-model   # validate the model and its links to the code — runs in CI, inside make verify
+make model-drawio  # write the committed draw.io canvas, docs/event-model/model.drawio
+make check-drawio  # fail when that canvas is missing or stale — runs in CI, inside make verify
 ```
 
 ## Why one model rather than one per feature
@@ -483,15 +485,54 @@ Generated artifacts:
 | `segments/model-N.svg` | the timeline in readable pieces — what the README embeds |
 | `slices/<id>.svg` | one slice with the events it consumes drawn in |
 | `model.png` | raster copy, only with `PNG=1` |
+| `model.drawio` | **the committed canvas** — the same timeline as one editable draw.io page, written by `make model-drawio` and held current by `make check-drawio`. See below |
 
-`model.mmd` and `model.svg` are committed so the diagram is reviewable and renders where Mermaid 11.16 is
-not available. **GitHub's Markdown renderer runs a Mermaid older than 11.15**, so a fenced `eventmodeling`
+Everything `make model` writes is generated on demand and not committed: it takes a headless browser to
+draw, so nothing that runs on every commit could prove a committed copy current, and a picture that can
+quietly stop matching the model is worse than none. The pages workflow redraws it all on every model
+change. **GitHub's Markdown renderer runs a Mermaid older than 11.15**, so a fenced `eventmodeling`
 block will not draw there — check what it runs today by putting a fenced `info` diagram in a comment, which
-renders as the version string. Reference the SVG instead:
+renders as the version string. Where the README needs a picture, `make model` embeds the segment SVGs
+inside its markers, or links the published page once `render.page` is set.
 
-```markdown
-![Global event model](docs/event-model/model.svg)
-```
+## The committed canvas
+
+`make model-drawio` writes the whole timeline as one draw.io page, `docs/event-model/model.drawio`, and that
+file **is committed**. It is the one rendering with that property, and the reason is what it takes to draw:
+a `.drawio` file is plain XML, so writing one is arithmetic and a string — no browser, no account, no
+credential, no network. That is what lets `make check-drawio` regenerate it in memory and compare it on every
+commit, inside `make verify`, which nothing driving a headless browser could ever do.
+
+| | Mermaid (`make model`) | draw.io (`make model-drawio`) |
+|---|---|---|
+| Produces | slice diagrams, README segments, the whole-timeline SVG, the browsable page | one editable canvas |
+| Lives | generated on demand, not committed | committed |
+| Needs | a headless browser | Node, and nothing else |
+| Currency | redrawn by the pages workflow | **`check-drawio`, inside `verify`** |
+| Good for | reading, embedding, GitHub | working on, annotating, presenting |
+
+Neither replaces the other, and both draw the same model: a box that sits fourth on the canvas sits fourth in
+`model.svg`, in the same swimlane, with the same arrows and the same colour. That is not a coincidence to
+maintain — which lane a box is in and which arrows exist are answered once, in `scripts/event-model/model.ts`,
+and both renderers ask it; the palette is one table both read.
+
+What the canvas looks like: one column per box, left to right in timeline order; rows are the notation's
+bands — UI and automation on top, commands and read models in the middle, events underneath — with a lane
+per actor or stream where the model groups them; a caption per slice above the bands; a legend above that.
+A step within a slice is a solid arrow. A read of an earlier event is dashed, and one that spans more than
+a column detours below the bands through a corridor rather than crossing every box in between, with every
+reader of one event sharing that event's corridor. Under each box that reads anything is a caption naming
+every event it reads, so a reader learns where an arrow came from without following it.
+
+Open it in draw.io — [app.diagrams.net](https://app.diagrams.net), the desktop app, or the VS Code
+extension — to present it, walk a stakeholder through it, or lay annotations over it. **Do not edit the
+committed file**: `make check-drawio` compares it byte for byte with what the model produces, so a hand
+edit fails the gate on the next commit. Copy it to annotate, and put what you learn back into `model.yaml`,
+where the next `make model-drawio` will draw it. A model with no slices has no canvas, and a canvas left
+behind by a model that was emptied is stale like any other.
+
+Fit-to-window on a model of any size shows boxes and no text: at a few percent zoom every label is
+sub-pixel. That is the viewer, not the file — zoom in.
 
 ## Publishing it
 
@@ -535,6 +576,11 @@ not a project devDependency, so `npm install` stays fast and `make check-model`
 — needs no browser at all. The local prefix is so the renderer can patch mermaid's swimlane bug
 ([mermaid-js/mermaid#7925](https://github.com/mermaid-js/mermaid/issues/7925)) before the first diagram.
 
+**`check-drawio` needs Node and no browser.** It runs the same TypeScript pipeline `make model` does, minus
+the rendering, so `make verify` in an event-profile project installs `scripts/event-model`'s three
+dependencies on first run — about a second once the tree is there — whichever language the services are
+written in. CI is given a Node for it where the project has none of its own.
+
 **That download does not exist on linux/arm64.** Google ships no Chrome build for it, so on an ARM
 sandbox or container `make model` fails at the browser fetch. Nothing gates on this — `check-model` needs
 no browser — but to actually regenerate the diagram there, supply a browser yourself and tell the

{slipwai-1.1.1.dev4 → slipwai-1.2.0.dev19}/assets/toolkit/docs/event-model/model.yaml

@@ -2,7 +2,9 @@
 # The global event model. One file, for the life of the project.
 #
 # This is the source of truth for the diagram, not the diagram itself: `make model` generates
-# model.mmd, model.svg, one view per slice, and model.html from what is here. Never edit those.
+# model.mmd, model.svg, one view per slice, and model.html from what is here, and `make model-drawio`
+# writes model.drawio — the one rendering that is committed, and that `make check-drawio` holds
+# current on every commit. Never edit any of those.
 #
 # It ships empty on purpose. A slice named here is a permanent commitment — event schemas are never
 # rewritten and stream identity sets the concurrency ceiling — so the starter makes none. Your first

{slipwai-1.1.1.dev4 → slipwai-1.2.0.dev19}/assets/toolkit/scripts/agents/benchmark.py

@@ -55,14 +55,14 @@ OVERVIEW = "benchmark.md"
 # The ladder, in order, plus the stages an adopted repository adds; a stage outside it is accepted and sorted last.
 LADDER = (
     "ground", "principles", "specify", "event-model", "split", "example-map", "gaps", "release-constraint", "plan",
-    "tasks", "pin", "implement", "converge", "demo", "adversary", "mutation",
+    "tasks", "pin", "implement", "converge", "demo", "adversary", "mutation", "skipper", "hand", "bosun",
 )
 OUTCOMES = ("accepted", "behaviour", "implementation")
 # Converge passes beyond which the overview says something: one pass to find work and one to confirm it
 # closed is the shape of a slice that converged, so the third is the first that is worth reading about.
 REPEATED = 3
 COUNTS = ("gaps", "findings", "seams", "verify_failures", "split")
-WORDS = ("mutation_score", "outcome", "model", "agent", "note", "delegate", "cycle")
+WORDS = ("mutation_score", "outcome", "model", "agent", "note", "delegate", "cycle", "driver")
 COMMENT = (
     "What each stage of /drive cost this slice and how well it did, one entry per stage run, appended by "
     "scripts/agents/benchmark.py at the stage's start and end. Tokens come from the harness's own transcript or are "