Move the scenario runner out of the MCP server

The validate-ui-scenario skill runs `runScenario`, which drives VS Code through
`test/automation` and writes an evidence bundle. None of that is MCP: the runner
loads no MCP module at runtime, and the SDK import it inherited was type-only, so
TypeScript already elided it. It only lived under `test/mcp` because that is where
the evidence pipeline was first written.

That matters now: deleting the MCP server would take the skill with it.

Move the six files that have nothing to do with MCP into a new `test/scenario`
package, and leave `test/mcp` as one of its consumers alongside the skill. The MCP
evidence tools move to `test/mcp/src/evidenceTools.ts`, where the server-specific
schemas belong. Deleting `test/mcp` now removes only MCP code.

Drop the step banner along with it. `showOverlay` appended a banner to the DOM of
the product under test, which can shift layout and influence focus, so the runner
always opted out via VSCODE_EVIDENCE_CLEAN_CAPTURE. With the runner as the only
caller that opinion is unanimous, so the overlay and its opt-out both go and the
capture is unconditionally faithful. Step titles are still rendered onto the
finished recording by renderEvidenceChapters.

The new package emits declarations, matching `test/automation`, so the MCP server
keeps real types rather than silently degrading to `any`.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485

Author: bryanchen-d

.github/skills/validate-ui-scenario/SKILL.md

@@ -11,15 +11,15 @@ Use this to reproduce a reported bug, to show that a fix works, or to attach a r
 test-plan item. For deterministic regression coverage that runs on every build, write a smoke test
 instead (see the `smoke-tests` skill) — this skill is for one-off, issue-derived validation.
 
-A scenario is a small JavaScript file run by `test/mcp/out/runScenario.js`. Nothing else has to be
+A scenario is a small JavaScript file run by `test/scenario/out/runScenario.js`. Nothing else has to be
 configured: the runner launches VS Code, records video and a trace, captures a screenshot at every
 step boundary, writes the report, and captions the recording with each step and its result.
 
 ## Prepare
 
 ```bash
 npm install                        # once
-npm --prefix test/mcp run compile  # after any change under test/mcp
+npm --prefix test/scenario run compile  # after any change under test/scenario
 ```
 
 Add `ffmpeg` and `ffprobe` to `PATH` to get the caption band on the video. Without them the run still
@@ -130,7 +130,7 @@ Each step receives a `context` with `app`, `workbench`, `code`, `page`, and `ski
 ## Run it
 
 ```bash
-node test/mcp/out/runScenario.js <scenario.cjs> --build "<app-root>"
+node test/scenario/out/runScenario.js <scenario.cjs> --build "<app-root>"
 ```
 
 Exit code `0` means every step passed, `1` means the run failed or was aborted, `2` a usage error.
@@ -149,7 +149,7 @@ Evidence is written to `.build/vscode-playwright-mcp/evidence/<run-id>/`:
 The caption band is added **above** the recorded frame rather than drawn over it, so no recorded
 pixel is hidden and the recording keeps its original length. Each caption carries the step number
 and id, its status, the step title, and the validation detail the step reported. Re-render after
-editing a manifest with `node test/mcp/out/renderEvidenceChapters.js <run-dir>`.
+editing a manifest with `node test/scenario/out/renderEvidenceChapters.js <run-dir>`.
 
 ## What makes evidence trustworthy
 

build/npm/dirs.ts

@@ -59,6 +59,7 @@ export const dirs = [
 	'test/integration/browser',
 	'test/monaco',
 	'test/smoke',
+	'test/scenario',
 	'test/mcp',
 	'.vscode/extensions/vscode-selfhost-import-aid',
 	'.vscode/extensions/vscode-selfhost-test-provider',

eslint.config.js

@@ -2329,10 +2329,22 @@ export default defineConfig(
 						'*' // node modules
 					]
 				},
+				{
+					'target': 'test/scenario/**',
+					'restrictions': [
+						'test/automation',
+						'test/scenario/**',
+						'@vscode/*',
+						'@parcel/*',
+						'@playwright/*',
+						'*' // node modules
+					]
+				},
 				{
 					'target': 'test/mcp/**',
 					'restrictions': [
 						'test/automation',
+						'test/scenario',
 						'test/mcp/**',
 						'@vscode/*',
 						'@parcel/*',

test/mcp/package.json

@@ -5,10 +5,10 @@
   "main": "./out/main.js",
   "private": true,
   "scripts": {
-    "compile": "cd ../automation && npm run compile && cd ../mcp && node ../../node_modules/typescript/bin/tsc6",
-    "watch-automation": "cd ../automation && npm run watch",
+    "compile": "cd ../scenario && npm run compile && cd ../mcp && node ../../node_modules/typescript/bin/tsc6",
+    "watch-scenario": "cd ../scenario && npm run watch",
     "watch-mcp": "node ../../node_modules/typescript/bin/tsc6 --watch --preserveWatchOutput",
-    "watch": "npm-run-all2 -lp watch-automation watch-mcp",
+    "watch": "npm-run-all2 -lp watch-scenario watch-mcp",
     "start-stdio": "echo 'Starting vscode-automation-mcp... For customization and troubleshooting, see ./test/mcp/README.md' && npm ci && npm run -s compile && node ./out/stdio.js"
   },
   "dependencies": {

test/mcp/src/automation.ts

@@ -4,11 +4,11 @@
  *--------------------------------------------------------------------------------------------*/
 
 import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
-import { ApplicationService } from './application';
+import { ApplicationService, EvidenceService } from '../../scenario';
 import { applyAllTools } from './automationTools/index.js';
 import type { Server } from '@modelcontextprotocol/sdk/server/index.js';
 import { z } from 'zod';
-import { applyEvidenceStartTool, applyEvidenceTools, EvidenceService } from './evidence.js';
+import { applyEvidenceStartTool, applyEvidenceTools } from './evidenceTools.js';
 
 export async function getServer(appService: ApplicationService): Promise<Server> {
 	const server = new McpServer({

test/mcp/src/automationTools/activityBar.ts

@@ -4,7 +4,7 @@
  *--------------------------------------------------------------------------------------------*/
 
 import { McpServer, RegisteredTool } from '@modelcontextprotocol/sdk/server/mcp.js';
-import { ApplicationService } from '../application';
+import { ApplicationService } from '../../../scenario';
 
 /**
  * Activity Bar Tools

test/mcp/src/automationTools/chat.ts

@@ -4,7 +4,7 @@
  *--------------------------------------------------------------------------------------------*/
 
 import { McpServer, RegisteredTool } from '@modelcontextprotocol/sdk/server/mcp.js';
-import { ApplicationService } from '../application';
+import { ApplicationService } from '../../../scenario';
 import { z } from 'zod';
 
 /**

test/mcp/src/automationTools/core.ts

@@ -5,7 +5,7 @@
 
 import { McpServer, RegisteredTool } from '@modelcontextprotocol/sdk/server/mcp.js';
 import { z } from 'zod';
-import { ApplicationService, assertNoProfileOverrides } from '../application';
+import { ApplicationService, assertNoProfileOverrides } from '../../../scenario';
 
 /**
  * Core Application Management Tools

test/mcp/src/automationTools/debug.ts

@@ -4,7 +4,7 @@
  *--------------------------------------------------------------------------------------------*/
 
 import { McpServer, RegisteredTool } from '@modelcontextprotocol/sdk/server/mcp.js';
-import { ApplicationService } from '../application';
+import { ApplicationService } from '../../../scenario';
 import { z } from 'zod';
 
 /**

test/mcp/src/automationTools/editor.ts

@@ -4,7 +4,7 @@
  *--------------------------------------------------------------------------------------------*/
 
 import { McpServer, RegisteredTool } from '@modelcontextprotocol/sdk/server/mcp.js';
-import { ApplicationService } from '../application';
+import { ApplicationService } from '../../../scenario';
 import { z } from 'zod';
 
 /**