@adia-ai/adia-ui-factory 0.8.0 → 0.8.2

package/skills/adia-project/evals/routing-corpus.json

@@ -0,0 +1,171 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "name": "adia-project routing accuracy corpus",
+  "version": "1.0.0",
+  "purpose": "Routing-eval corpus for adia-project (adia-ui-factory plugin). Each phrase declares whether adia-project SHOULD be the routing target. Scored by the estate's TF-IDF routing-eval runner against the skill's description (heuristic token overlap).",
+  "scoring_notes": "Heuristic signal, not ground truth. Treat misroutes as a prompt to tighten the skill description, never as a reason to keyword-stuff it. Real harness routing is LLM-driven.",
+  "license": "internal",
+  "scope": "adia-project routing — does this phrase activate adia-project?",
+
+  "minimums_per_spec": {
+    "trigger_phrases": 14,
+    "adversarial_phrases": 6,
+    "task_shapes_covered": "new-app, add-surface, add-page, add-component, inventory",
+    "adversarial_fraction": "30%"
+  },
+
+  "phrases": [
+    {
+      "id": "project-newapp-01",
+      "phrase": "scaffold a new SPA app for this adia-ui project",
+      "should_route_to_project": true,
+      "expected_shape": "new-app",
+      "rationale": "The exact 'scaffold an app' trigger phrase named in the description, canonical new-app mode."
+    },
+    {
+      "id": "project-newapp-02",
+      "phrase": "set up an SSR app with the Next framework — four-axis dirs and all",
+      "should_route_to_project": true,
+      "expected_shape": "new-app",
+      "rationale": "Names the SSR one-shot (bin/adia-scaffold ssr <name> --framework next) and the four-axis layout directly."
+    },
+    {
+      "id": "project-newapp-03",
+      "phrase": "help me decide the project shape — single-surface, rollup, or shared-foundation — for this fresh app",
+      "should_route_to_project": true,
+      "expected_shape": "new-app",
+      "rationale": "Names all three project shapes verbatim; this is the shape-decision gate adia-project owns before scaffolding."
+    },
+
+    {
+      "id": "project-addsurface-01",
+      "phrase": "we're already a rollup — add a new surface under app/ for the billing section",
+      "should_route_to_project": true,
+      "expected_shape": "add-surface",
+      "rationale": "Mode and shape are already decided (rollup); this is the add-surface mode growing an existing shape, adia-project's on-disk-structure job."
+    },
+    {
+      "id": "project-addsurface-02",
+      "phrase": "grow the shared-foundation with a new sibling app under app/shared",
+      "should_route_to_project": true,
+      "expected_shape": "add-surface",
+      "rationale": "Names shared-foundation and app/shared directly — add-surface mode with shape already settled."
+    },
+
+    {
+      "id": "project-addpage-01",
+      "phrase": "add a new page to the settings rollup",
+      "should_route_to_project": true,
+      "expected_shape": "add-page",
+      "rationale": "The exact 'add a page' trigger phrase from the description, add-page mode."
+    },
+    {
+      "id": "project-addpage-02",
+      "phrase": "decide whether this new page should be trio or DUO form",
+      "should_route_to_project": true,
+      "expected_shape": "add-page",
+      "rationale": "Names the page-form gate (trio vs DUO) verbatim from the skill body."
+    },
+    {
+      "id": "project-addpage-03",
+      "phrase": "scaffold a page in the reports surface with --duo",
+      "should_route_to_project": true,
+      "expected_shape": "add-page",
+      "rationale": "Direct bin/adia-scaffold page invocation with the --duo flag named in the skill body."
+    },
+
+    {
+      "id": "project-addcomponent-01",
+      "phrase": "scaffold a new custom element folder in our app for the badge widget",
+      "should_route_to_project": true,
+      "expected_shape": "add-component",
+      "rationale": "add-component mode: a consumer-app custom element folder via bin/adia-scaffold component, not a framework primitive."
+    },
+    {
+      "id": "project-addcomponent-02",
+      "phrase": "add a component folder via adia-scaffold for the app's local <stat-card> element",
+      "should_route_to_project": true,
+      "expected_shape": "add-component",
+      "rationale": "Explicit adia-scaffold component invocation, app-local element."
+    },
+    {
+      "id": "project-addcomponent-03",
+      "phrase": "check that the new component's folder name matches its tag per convention",
+      "should_route_to_project": true,
+      "expected_shape": "add-component",
+      "rationale": "Verifies the add-component gate: components/<tag>/<tag>.{js,css} exists, folder = tag."
+    },
+
+    {
+      "id": "project-inventory-01",
+      "phrase": "audit this existing app's structure — score it against the structure rubric",
+      "should_route_to_project": true,
+      "expected_shape": "inventory",
+      "rationale": "The exact 'audit its structure' trigger phrase from the description, inventory mode."
+    },
+    {
+      "id": "project-inventory-02",
+      "phrase": "score whether our project shape still matches how the app has actually grown",
+      "should_route_to_project": true,
+      "expected_shape": "inventory",
+      "rationale": "Inventory mode: shape declared & matched is one of the rubric's named gates."
+    },
+    {
+      "id": "project-inventory-03",
+      "phrase": "check this rollup for duplicated cross-surface code",
+      "should_route_to_project": true,
+      "expected_shape": "inventory",
+      "rationale": "Names a specific structure-rubric gate ('no duplicated cross-surface code') verbatim."
+    },
+
+    {
+      "id": "project-adv-01",
+      "phrase": "add a surface for the new claims-review tool — we haven't decided mode or shell yet",
+      "should_route_to_project": false,
+      "expected_alternative": "adia-orient (adia-ui-factory plugin)",
+      "rationale": "Adversarial — the unevidenced handoff the audit flagged: 'add a surface' appears in both adia-orient's and adia-project's descriptions, but with mode/shape/shell still undecided this is cold-start orientation, not structure-laying. adia-project's add-surface mode presumes the shape is already picked."
+    },
+    {
+      "id": "project-adv-02",
+      "phrase": "scaffold a plain React app, no AdiaUI components involved",
+      "should_route_to_project": false,
+      "expected_alternative": "none",
+      "rationale": "Adversarial — 'scaffold an app' surface-matches the trigger phrase, but with no adia-ui in scope this is generic framework tooling; no AdiaUI skill should claim it."
+    },
+    {
+      "id": "project-adv-03",
+      "phrase": "restructure the framework monorepo's package layout under packages/web-components",
+      "should_route_to_project": false,
+      "expected_alternative": "adia-author (adia-ui-forge plugin)",
+      "rationale": "Adversarial — adia-project lays out a CONSUMER app's on-disk structure; restructuring the framework monorepo itself is forge-side authoring territory."
+    },
+    {
+      "id": "project-adv-04",
+      "phrase": "build the actual screen content for the new page we just scaffolded",
+      "should_route_to_project": false,
+      "expected_alternative": "adia-compose",
+      "rationale": "Adversarial — the description's own NOT-fence: 'the UI inside a screen' is adia-compose's job once adia-project has laid out where the page lives."
+    },
+    {
+      "id": "project-adv-05",
+      "phrase": "wire the host's SPA registration and CSS cascade for this app",
+      "should_route_to_project": false,
+      "expected_alternative": "adia-host",
+      "rationale": "Adversarial — host bootstrap/registration is explicitly deferred to adia-host in the description's NOT-fence."
+    },
+    {
+      "id": "project-adv-06",
+      "phrase": "wire up the state/data pattern for the new surface",
+      "should_route_to_project": false,
+      "expected_alternative": "adia-data",
+      "rationale": "Adversarial — state/data pattern choice is explicitly deferred to adia-data in the references section, distinct from structure/scaffolding."
+    }
+  ],
+
+  "evaluator_notes": {
+    "current_state": "20 cases (14 trigger + 6 adversarial). Trigger cases: 3 new-app, 2 add-surface, 3 add-page, 3 add-component, 3 inventory (all 5 modes covered).",
+    "promotion_criteria": "Promote to a CI gate (warn → hard-fail) when F1 >= 0.85 on this corpus across 3+ consecutive runs without description edits.",
+    "review_cadence": "Re-run on every adia-project description edit. Add 1-2 new cases per checkpoint reflecting newly-observed routing failure modes.",
+    "known_limitations": "project-adv-01 encodes the audit's flagged unevidenced handoff: 'add a surface' appears verbatim in both adia-orient's and adia-project's descriptions. The heuristic TF-IDF scorer cannot see whether mode/shape were already decided (that context lives in conversation state, not the phrase text alone), so this case is likely to be the corpus's most fragile — an LLM router resolves it on context adia-project's own description doesn't carry; the phrase text alone may false-positive toward adia-project until the description gains an explicit 'mode/shape already decided' qualifier."
+  }
+}

package/skills/adia-shells/SKILL.md

@@ -3,8 +3,9 @@ name: adia-shells
 description: >-
   Picks and composes an adia-ui page-chrome shell (@adia-ai/web-modules):
   admin (app frame), chat (LLM), editor (canvas+panes), simple
-  (marketing/error/landing), or embed. Use on "use a shell",
-  "sidebar + topbar layout", "embed this surface", or shell markup debugging.
+  (marketing/error), or embed. Use when asked to "use a
+  shell", "sidebar + topbar layout", "embed this surface", or on shell markup
+  debugging.
   NOT for screens inside it (adia-compose) or host/SSR wiring (adia-host).
 disable-model-invocation: false
 user-invocable: true
@@ -36,7 +37,7 @@ A composed shell is done when all five gates hold and the surface renders (`adia
 | Gate | Check | Enforcement |
 | --- | --- | --- |
 | Cluster registered | the barrel import is present; JS-bearing children resolve | self-verified |
-| Canonical nesting | parent→child structure matches the shell's reference (e.g. `admin-page` only inside `admin-scroll`; `admin-page-header` wraps `<header-ui>`, not a raw `<header>`) | self-verified against the reference |
+| Canonical nesting | parent→child structure matches the shell's reference (e.g. `admin-page` only inside `admin-scroll`; shell children never wrapped in `<col-ui>`/`<row-ui>` — the grid reads tag selectors) | self-verified against the reference |
 | No legacy shapes | no retired data-attribute forms | mechanized: `adia-lint` `LEGACY-SHELL` |
 | No native-primitive leak | controls are `*-ui`, not raw `<button>`/`<input>` | mechanized: `adia-lint` `NATIVE-PRIMITIVE` |
 | One route owner | SSR uses the framework outlet, not `<router-ui>` | self-verified |
@@ -44,6 +45,25 @@ A composed shell is done when all five gates hold and the surface renders (`adia
 The plugin's `bin/adia-lint` mechanizes the two marked gates on write; the other three are
 checked against the per-shell reference before declaring done.
 
+## Deliverable — the ShellComposition record
+
+Composing a shell emits this record — it materializes the pick table's "decide on a cited
+signal" demand and the five-gate table's results, in the same row order:
+
+```text
+Shell:                     admin | chat | editor | simple | embed | none  — signal: <file / dep / marker, or the user's explicit words>
+Barrel import:             <e.g. '@adia-ai/web-modules/shell'>
+Cluster registered:        pass | fail  — reference: <shell-<name>.md section checked>
+Canonical nesting:         pass | fail  — reference: <shell-<name>.md section checked>
+No legacy shapes:          pass | fail  — adia-lint: <LEGACY-SHELL output>
+No native-primitive leak:  pass | fail  — adia-lint: <NATIVE-PRIMITIVE output>
+One route owner:           pass | fail  — reference: <shell-<name>.md section checked>
+Reference consulted:       <shell-<name>.md>
+```
+
+The two mechanized gates cite `adia-lint`'s actual output; the three self-verified gates cite
+the reference section checked against, never an assumption.
+
 ## Shared conventions (every shell; the per-shell reference carries the specifics)
 
 - **Register by cluster barrel**, not piecemeal: `import '@adia-ai/web-modules/shell'` (or

package/skills/adia-shells/evals/routing-corpus.json

@@ -0,0 +1,167 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "name": "adia-shells routing accuracy corpus",
+  "version": "1.0.0",
+  "purpose": "Routing-eval corpus for adia-shells (adia-ui-factory plugin). Each phrase declares whether adia-shells SHOULD be the routing target. Scored by the estate's TF-IDF routing-eval runner against the skill's description (heuristic token overlap).",
+  "scoring_notes": "Heuristic signal, not ground truth. Treat misroutes as a prompt to tighten the skill description, never as a reason to keyword-stuff it. Real harness routing is LLM-driven.",
+  "license": "internal",
+  "scope": "adia-shells routing — does this phrase activate adia-shells?",
+
+  "minimums_per_spec": {
+    "trigger_phrases": 14,
+    "adversarial_phrases": 6,
+    "task_shapes_covered": "pick-admin, pick-chat, pick-editor, pick-simple, pick-embed, shell-debugging, shell-selection-generic",
+    "adversarial_fraction": "30%"
+  },
+
+  "phrases": [
+    {
+      "id": "shells-admin-01",
+      "phrase": "I need a sidebar + topbar layout with a command palette for this app",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-admin",
+      "rationale": "The exact 'sidebar + topbar layout' trigger phrase named in the description — this is the fence-test the audit flagged: it must land on adia-shells (picking page-chrome), not adia-compose (which builds screen content, not the frame around it)."
+    },
+    {
+      "id": "shells-admin-02",
+      "phrase": "set up the full app frame — sidebar(s) plus topbar plus command palette — for the dashboard",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-admin",
+      "rationale": "Restates the admin-shell decision-table signal ('full app frame — sidebar(s) + topbar + command palette + pages') in the requester's own words."
+    },
+    {
+      "id": "shells-chat-01",
+      "phrase": "build an LLM conversation surface with a thread and a composer",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-chat",
+      "rationale": "Matches chat-shell's decision-table signal: an LLM conversation surface (thread + composer)."
+    },
+    {
+      "id": "shells-chat-02",
+      "phrase": "wire up the chat-shell's thread and composer chrome",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-chat",
+      "rationale": "Names chat-shell directly and asks for its chrome, not the conversation content inside it."
+    },
+    {
+      "id": "shells-editor-01",
+      "phrase": "I need a design tool with a center canvas and resizable side panes",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-editor",
+      "rationale": "Matches editor-shell's decision-table signal: a design tool — center canvas + resizable side panes."
+    },
+    {
+      "id": "shells-editor-02",
+      "phrase": "add focus mode to the editor-shell's canvas + panes layout",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-editor",
+      "rationale": "Focus mode is an editor-shell chrome feature named in the skill body, not screen content."
+    },
+    {
+      "id": "shells-simple-01",
+      "phrase": "scaffold a minimal centered chrome for the marketing landing page",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-simple",
+      "rationale": "Matches simple-shell's decision-table signal: marketing/error/landing/auth — minimal centered chrome."
+    },
+    {
+      "id": "shells-simple-02",
+      "phrase": "pick a shell for the auth error page — it just needs minimal centered chrome",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-simple",
+      "rationale": "Explicit shell-selection ask against the simple-shell signal."
+    },
+    {
+      "id": "shells-embed-01",
+      "phrase": "embed this surface into a host page — size and center the light-DOM element",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-embed",
+      "rationale": "The exact 'embed this surface' trigger phrase named in the description, matched to embed-shell's decision-table row."
+    },
+    {
+      "id": "shells-embed-02",
+      "phrase": "wire the embed-shell cluster for the widget we're dropping into a third-party site",
+      "should_route_to_shells": true,
+      "expected_shape": "pick-embed",
+      "rationale": "Names embed-shell directly; cluster registration is shell-selection/composition territory."
+    },
+    {
+      "id": "shells-debug-01",
+      "phrase": "the sidebar's [collapsed] attribute doesn't collapse the admin-shell layout — help me debug the shell markup",
+      "should_route_to_shells": true,
+      "expected_shape": "shell-debugging",
+      "rationale": "The exact 'shell markup debugging' trigger phrase named in the description, tied to the shared state-is-an-attribute convention."
+    },
+    {
+      "id": "shells-debug-02",
+      "phrase": "adia-lint flagged a legacy <aside data-sidebar> shape in our shell — fix the markup",
+      "should_route_to_shells": true,
+      "expected_shape": "shell-debugging",
+      "rationale": "Legacy data-attribute shapes are called out explicitly in the skill body as retired shell-tier vocabulary; fixing them is shell markup debugging."
+    },
+    {
+      "id": "shells-generic-01",
+      "phrase": "use a shell for this new app frame",
+      "should_route_to_shells": true,
+      "expected_shape": "shell-selection-generic",
+      "rationale": "The literal 'use a shell' trigger phrase from the description with no shell yet named — canonical cold-start-within-shells ask."
+    },
+    {
+      "id": "shells-generic-02",
+      "phrase": "which shell fits an app that needs both a big canvas and side panes",
+      "should_route_to_shells": true,
+      "expected_shape": "shell-selection-generic",
+      "rationale": "A decision-cue question aimed at the pick-the-shell table without naming a shell — still adia-shells' job, not a specific-shell shape."
+    },
+
+    {
+      "id": "shells-adv-01",
+      "phrase": "compose the screens that go inside the admin-shell's page area",
+      "should_route_to_shells": false,
+      "expected_alternative": "adia-compose",
+      "rationale": "Adversarial — the audit's core fence question, resolved the other direction: screens INSIDE an already-chosen shell are adia-compose's job per the description's own NOT-fence."
+    },
+    {
+      "id": "shells-adv-02",
+      "phrase": "wire the SSR host so the framework's route outlet renders inside the shell",
+      "should_route_to_shells": false,
+      "expected_alternative": "adia-host",
+      "rationale": "Adversarial — host/SSR route-outlet wiring is explicitly carved out to adia-host in both the description and the shared-conventions section (NEVER mount <router-ui> under SSR — the framework outlet owns the route)."
+    },
+    {
+      "id": "shells-adv-03",
+      "phrase": "the admin-shell's sidebar-collapse behavior regressed — fix the bug in packages/web-modules source",
+      "should_route_to_shells": false,
+      "expected_alternative": "adia-author (adia-ui-forge plugin)",
+      "rationale": "Adversarial — a shell BUG in the framework monorepo itself (editing packages/web-modules source) is forge-side authoring, not factory-side shell selection/composition."
+    },
+    {
+      "id": "shells-adv-04",
+      "phrase": "wire the state/data hydration for the shell's sidebar list",
+      "should_route_to_shells": false,
+      "expected_alternative": "adia-data",
+      "rationale": "Adversarial — data/state pattern choice is adia-data's territory; adia-shells only reflects state as an attribute the consumer already owns."
+    },
+    {
+      "id": "shells-adv-05",
+      "phrase": "add the LLM streaming client to power the chat-shell's composer",
+      "should_route_to_shells": false,
+      "expected_alternative": "adia-llm",
+      "rationale": "Adversarial — the chat LLM client/proxy contract is explicitly deferred to adia-llm in the shell's own references section, even though it's chat-shell-adjacent."
+    },
+    {
+      "id": "shells-adv-06",
+      "phrase": "scaffold a brand new app's on-disk structure before we even pick a shell",
+      "should_route_to_shells": false,
+      "expected_alternative": "adia-project",
+      "rationale": "Adversarial — app structure/scaffolding precedes shell selection and belongs to adia-project; no shell has been named or implied."
+    }
+  ],
+
+  "evaluator_notes": {
+    "current_state": "20 cases (14 trigger + 6 adversarial). Each of the 7 task shapes (5 named shells + shell-debugging + shell-selection-generic) has 2 trigger cases.",
+    "promotion_criteria": "Promote to a CI gate (warn → hard-fail) when F1 >= 0.85 on this corpus across 3+ consecutive runs without description edits.",
+    "review_cadence": "Re-run on every adia-shells description edit. Add 1-2 new cases per checkpoint reflecting newly-observed routing failure modes.",
+    "known_limitations": "The audit's core open question — whether 'sidebar + topbar layout' reaches adia-shells or gets pulled toward adia-compose — is encoded as shells-admin-01/02 (positive) and shells-adv-01 (the compose-side adversarial); the heuristic TF-IDF scorer may not separate 'layout' (shell chrome) from 'screen'/'form'/'dashboard' (compose content) as cleanly as an LLM router would, since both descriptions share UI-layout vocabulary."
+  }
+}

package/skills/adia-verify/SKILL.md

@@ -3,7 +3,7 @@ name: adia-verify
 description: >-
   Browser-QA gate for adia-ui surfaces — renders headless with zero
   console/page errors, non-zero bounding boxes, and a screenshot actually
-  read, plus AdiaUI a11y checks. Use before shipping a surface, on
+  read, plus AdiaUI a11y checks. Use when shipping a surface, on
   "verify/QA this page", or when "tests pass" is the only evidence. NOT for
   composing/fixing UI (adia-compose) or structural lint (adia-lint hook).
 disable-model-invocation: false
@@ -27,7 +27,27 @@ Probe shape, the failure classes only this gate catches (0×0 host, empty-page-w
 
 ## Tooling boundary
 
-The advisory `adia-lint` PostToolUse hook mechanizes the structural slice on every write (shadow DOM, raw color/px, `::slotted`, native-primitive leaks, legacy shell shapes, SSR traps) and never blocks — fix its findings rather than re-deriving its rules. The framework's `audit:shell-composition` / `audit:native-primitive-leak` run in the @adia-ai framework repo; they are not shipped in this plugin and cannot be invoked from a consumer app. Everything else — the render, the screenshot read, a11y — is self-verified: no shipped script enforces it.
+The advisory `adia-lint` PostToolUse hook mechanizes the structural slice on every write (shadow DOM, raw color/px, `::slotted`, native-primitive leaks, legacy shell shapes, SSR traps, and the shell-nesting/LLM-key/genui-validate rules) and never blocks — fix its findings rather than re-deriving its rules. The framework's `audit:shell-composition` / `audit:native-primitive-leak` run in the @adia-ai framework repo; they are not shipped in this plugin and cannot be invoked from a consumer app.
+
+The render gate itself IS shipped: `node "${CLAUDE_PLUGIN_ROOT}/bin/adia-probe.mjs" <url> --selector <key-element> [--json]` runs gates 1–2 mechanically (console/page errors, bounding boxes, the `deviceScaleFactor: 2` capture — set at context creation, where it actually takes effect) and emits the VerifyProof below with verdict `pass-pending-read`. The screenshot READ and the a11y judgment slice stay yours — the probe's `imageRead` slot exists precisely so a report can't silently skip them. Requires Playwright in the consumer app (`npm i -D playwright`).
+
+## Deliverable — the VerifyProof
+
+The QA pass returns this record (the probe emits the mechanized half via `--json`; the reader completes it):
+
+```text
+VerifyProof
+url:            <the probed page>
+consoleErrors:  pass | fail — <each error verbatim; empty = pass>
+boundingBoxes:  pass | fail — <selector: w×h per key element; any 0×0 = fail>
+screenshot:     <path> @ deviceScaleFactor 2
+imageRead:      <what the pixels actually show — REQUIRED prose, never "looks fine">
+a11y:           pass | fail — region role/label · .open-driven overlays · heading roles · keyboard path · AA contrast
+structure:      adia-lint clean on every written file | <findings>
+verdict:        ship | hold — <one line naming the blocker if hold>
+```
+
+A VerifyProof with an empty `imageRead` or a bare "looks fine" is incomplete — the slot carries what was SEEN, so a later reader can challenge it.
 
 ## Verify rubric
 
@@ -39,5 +59,6 @@ A surface ships only when all pass:
 | Screenshot read `[gate]` | the `deviceScaleFactor: 2` capture was looked at, after the latest structural change |
 | Accessible `[gate]` | region role + label; `.open`-driven overlays; real heading roles; keyboard path; AA contrast |
 | Structurally clean `[gate]` | `adia-lint` reports no smells on every file written for the surface |
+| Data wired `[gate]` | the surface's Wiring Record (adia-data §Deliverable) exists with every round-trip row observed — a screen with unrecorded state wiring isn't done |
 
 Neighboring work routes out: composing or fixing the surface → `adia-compose` · shell chrome → `adia-shells` · hydration/state → `adia-data` · generated-markup validation → `adia-genui`.

package/skills/adia-verify/evals/routing-corpus.json

@@ -0,0 +1,171 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "name": "adia-verify routing accuracy corpus",
+  "version": "1.0.0",
+  "purpose": "Routing-eval corpus for adia-verify (adia-ui-factory plugin). Each phrase declares whether adia-verify SHOULD be the routing target. Scored by the estate's TF-IDF routing-eval runner against the skill's description (heuristic token overlap).",
+  "scoring_notes": "Heuristic signal, not ground truth. Treat misroutes as a prompt to tighten the skill description, never as a reason to keyword-stuff it. Real harness routing is LLM-driven.",
+  "license": "internal",
+  "scope": "adia-verify routing — does this phrase activate adia-verify (the consumer-app browser-QA exit gate), and does it correctly stay OUT of adia-release's release pre-flight gates, which is the one known central-corpus misroute for any 'verify'-flavored phrase?",
+
+  "minimums_per_spec": {
+    "trigger_phrases": 14,
+    "adversarial_phrases": 6,
+    "task_shapes_covered": "browser-gate, qa-request, a11y-check, debug-symptom, anti-pattern-tests-pass",
+    "adversarial_fraction": "30%"
+  },
+
+  "phrases": [
+    {
+      "id": "verify-gate-01",
+      "phrase": "verify this surface renders with zero console errors before we ship it",
+      "should_route_to_verify": true,
+      "expected_shape": "browser-gate",
+      "rationale": "Canonical pre-ship browser-render gate phrasing — matches the skill's own 'before shipping a surface' trigger."
+    },
+    {
+      "id": "verify-gate-02",
+      "phrase": "run the browser gate on the new settings page",
+      "should_route_to_verify": true,
+      "expected_shape": "browser-gate",
+      "rationale": "Direct 'browser gate' terminology from the skill's own name for its check."
+    },
+    {
+      "id": "verify-gate-03",
+      "phrase": "check that bounding boxes are non-zero on the dashboard before merging",
+      "should_route_to_verify": true,
+      "expected_shape": "browser-gate",
+      "rationale": "Non-zero bounding box is the gate's literal pass condition."
+    },
+
+    {
+      "id": "verify-qa-01",
+      "phrase": "verify/QA this page before we ship it",
+      "should_route_to_verify": true,
+      "expected_shape": "qa-request",
+      "rationale": "Verbatim 'verify/QA this page' phrase from the skill description."
+    },
+    {
+      "id": "verify-qa-02",
+      "phrase": "is this screen done? run the exit gate on it",
+      "should_route_to_verify": true,
+      "expected_shape": "qa-request",
+      "rationale": "'Is this screen done' — the exit-gate framing this skill owns."
+    },
+    {
+      "id": "verify-qa-03",
+      "phrase": "QA the checkout flow before it goes live for users",
+      "should_route_to_verify": true,
+      "expected_shape": "qa-request",
+      "rationale": "Consumer-app pre-launch QA request; deliberately avoids the word 'release' to not collide with adia-release."
+    },
+
+    {
+      "id": "verify-a11y-01",
+      "phrase": "run an accessibility pass on the profile page",
+      "should_route_to_verify": true,
+      "expected_shape": "a11y-check",
+      "rationale": "A11y check is one of the two gate halves this skill owns."
+    },
+    {
+      "id": "verify-a11y-02",
+      "phrase": "check AA contrast and the keyboard path on the new modal",
+      "should_route_to_verify": true,
+      "expected_shape": "a11y-check",
+      "rationale": "AA contrast + keyboard path are named rubric rows."
+    },
+    {
+      "id": "verify-a11y-03",
+      "phrase": "does this overlay have a real heading role and an aria-label on its region?",
+      "should_route_to_verify": true,
+      "expected_shape": "a11y-check",
+      "rationale": "Region role + heading role are named a11y-gate items."
+    },
+
+    {
+      "id": "verify-debug-01",
+      "phrase": "the screen renders completely blank, help me figure out why",
+      "should_route_to_verify": true,
+      "expected_shape": "debug-symptom",
+      "rationale": "'Empty page' symptom — a named failure class only this gate catches."
+    },
+    {
+      "id": "verify-debug-02",
+      "phrase": "this element reports a 0x0 bounding box, why isn't it sized",
+      "should_route_to_verify": true,
+      "expected_shape": "debug-symptom",
+      "rationale": "0x0 host is the skill's named failure class."
+    },
+    {
+      "id": "verify-debug-03",
+      "phrase": "the page looks broken visually but the console shows no errors",
+      "should_route_to_verify": true,
+      "expected_shape": "debug-symptom",
+      "rationale": "Clean-console-but-visually-broken is the anti-pattern this gate exists to catch."
+    },
+
+    {
+      "id": "verify-anti-01",
+      "phrase": "our unit tests pass, but I want someone to actually look at the rendered page before we call it done",
+      "should_route_to_verify": true,
+      "expected_shape": "anti-pattern-tests-pass",
+      "rationale": "Directly invokes the skill's named anti-pattern: 'tests pass, ship it'."
+    },
+    {
+      "id": "verify-anti-02",
+      "phrase": "confirm the surface actually renders and isn't just clipped — don't take the test suite's word for it",
+      "should_route_to_verify": true,
+      "expected_shape": "anti-pattern-tests-pass",
+      "rationale": "Clipped-content-despite-passing-tests is the gate's reason to exist."
+    },
+
+    {
+      "id": "verify-adv-01",
+      "phrase": "run the pre-flight verification gates before we cut the v0.8 release",
+      "should_route_to_verify": false,
+      "expected_alternative": "adia-release (forge plugin)",
+      "rationale": "Adversarial — the known central-corpus misroute: 'verify' language describing release pre-flight gates (F-N1, tag/publish checkpoints), not a consumer-app browser QA pass."
+    },
+    {
+      "id": "verify-adv-02",
+      "phrase": "run a dogfood sweep across every demo page in the framework repo",
+      "should_route_to_verify": false,
+      "expected_alternative": "adia-dogfood (forge plugin)",
+      "rationale": "Adversarial — framework-side demo sweep (visual probe, attr-quote typos, shell composition) is adia-dogfood's territory, not the consumer-app exit gate."
+    },
+    {
+      "id": "verify-adv-03",
+      "phrase": "the QA pass found the button is unstyled and misplaced — go fix the screen",
+      "should_route_to_verify": false,
+      "expected_alternative": "adia-compose (adia-ui-factory plugin)",
+      "rationale": "Adversarial — verify only reports the gap; fixing the composed screen routes to adia-compose (generator, not the gate)."
+    },
+    {
+      "id": "verify-adv-04",
+      "phrase": "the gate found the data isn't hydrating on load — fix the state wiring",
+      "should_route_to_verify": false,
+      "expected_alternative": "adia-data (adia-ui-factory plugin)",
+      "rationale": "Adversarial — fixing a hydration/state defect the gate surfaced belongs to adia-data, not to re-running verify."
+    },
+    {
+      "id": "verify-adv-05",
+      "phrase": "run the structural lint check on the files from my last commit",
+      "should_route_to_verify": false,
+      "expected_alternative": "adia-lint hook (advisory PostToolUse hook, not a skill)",
+      "rationale": "Adversarial — structural smells (shadow DOM, raw color/px, native-primitive leaks) are the adia-lint hook's mechanized slice; verify only checks that hook reports clean, it doesn't run the lint itself."
+    },
+    {
+      "id": "verify-adv-06",
+      "phrase": "write unit tests for the new date-parsing utility function",
+      "should_route_to_verify": false,
+      "expected_alternative": null,
+      "rationale": "Adversarial — generic unit-test authoring with no browser-render or a11y surface; no AdiaUI skill should activate."
+    }
+  ],
+
+  "evaluator_notes": {
+    "current_state": "20 cases (14 trigger + 6 adversarial). Each of the 5 task shapes has 2-3 trigger cases; adv-01 specifically targets the documented known misroute (verify-flavored phrase landing on adia-release).",
+    "promotion_criteria": "Promote to a CI gate (warn → hard-fail) when F1 ≥ 0.85 on this corpus across 3+ consecutive runs without description edits.",
+    "review_cadence": "Re-run on every adia-verify description edit. Add 1-2 new cases per checkpoint reflecting newly-observed routing failure modes.",
+    "known_limitations": "verify-adv-01 encodes the single highest-value adversarial for this skill: 'verify' as a bare word is release-gate-shaped in the central corpus, and the heuristic scorer may weight token overlap toward adia-release's own 'pre-flight gates' language. The description's browser/a11y-specific vocabulary (bounding boxes, console.error, aria-label) is the intended disambiguator."
+  }
+}

package/skills/adia-verify/references/verification.md

@@ -13,9 +13,16 @@ Render the surface in a real (headless) browser and assert four things, in order
 
 Gates 1 and 2 are independent, not redundant: `customElements.whenDefined(name)` never rejects, so a `Promise.all([...whenDefined]).then(bootstrap)` gate with one never-imported tag hangs forever — shell chrome still renders (tag-keyed CSS), the console stays **clean**, and the page is empty. Only the bounding-box and screenshot checks see it.
 
-Minimal probe shape (Playwright; works against a Vite static host or an SSR dev URL):
+The shipped probe runs this mechanically and emits the VerifyProof:
+`node "${CLAUDE_PLUGIN_ROOT}/bin/adia-probe.mjs" <url> --selector <el> --json`.
+Its logic, for when you need a bespoke variant (Playwright; works against a
+Vite static host or an SSR dev URL):
 
 ```js
+// deviceScaleFactor must be set at CONTEXT creation — scale:'device' on a
+// default context captures at 1x and silently fails gate 3's 2x claim.
+const context = await browser.newContext({ deviceScaleFactor: 2 });
+const page = await context.newPage();
 const errors = [];
 page.on('console', m => m.type() === 'error' && errors.push(m.text()));
 page.on('pageerror', e => errors.push(String(e)));