refactor: Remove LLM suggestion feature in favor of deterministic shell-quote parser

This commit simplifies the command pattern extraction implementation by:
- Removing all LLM-based command suggestion functionality
- Always using the shell-quote parser for deterministic command pattern extraction
- Eliminating the commandSuggestionsEnabled setting and related UI components
- Removing unnecessary complexity from the codebase

The shell-quote parser provides consistent and predictable results without
requiring LLM calls, making the feature more reliable and performant.

Author: hannesrudolph

command_whitelist_ui_location_summary.md

@@ -0,0 +1,67 @@
+# Command Whitelisting UI Location Summary
+
+## Overview
+
+The command whitelisting feature has been successfully moved from VS Code's native settings to the Roo Code plugin's settings interface. This consolidates all auto-approval settings in one convenient location.
+
+## Previous Location (REMOVED)
+
+- **VS Code Settings**: `Preferences > Settings > Extensions > Roo Code`
+- **Setting Name**: `roo-code.commandWhitelist`
+- **Access**: Required navigating through VS Code's settings UI or editing `settings.json`
+
+## New Location (CURRENT)
+
+The command whitelisting feature is now located in:
+
+### Access Path
+
+1. Open the Roo Code extension panel in VS Code
+2. Click on the **Settings** icon (gear icon) in the top toolbar
+3. Navigate to the **Auto Approve** section
+4. Find the **Execute** subsection
+
+### UI Components
+
+Within the Auto Approve > Execute section, you'll find:
+
+1. **Enable/Disable Toggle**
+
+    - Label: "Auto-approve command execution"
+    - Controls whether commands can be auto-approved
+
+2. **Command Patterns List**
+    - Label: "Command patterns"
+    - Description: "Add command patterns that can be auto-approved (e.g., 'npm test', 'git status')"
+    - Features:
+        - Add new patterns using the input field
+        - Remove patterns with the × button
+        - Patterns support wildcards (\*)
+        - Empty list means no commands are auto-approved
+
+### Example Patterns
+
+- `npm test` - Auto-approves exact command
+- `npm *` - Auto-approves any npm command
+- `git status` - Auto-approves git status command
+- `*` - Auto-approves all commands (use with caution)
+
+## Migration
+
+- Existing command whitelist settings from VS Code settings are automatically migrated to the new location on first launch
+- The old VS Code setting (`roo-code.commandWhitelist`) is removed from `package.json`
+- Users don't need to manually transfer their settings
+
+## Benefits
+
+1. **Centralized Settings**: All auto-approval settings (read, write, execute) are now in one place
+2. **Better UX**: No need to navigate VS Code's complex settings structure
+3. **Visual Consistency**: Matches the UI pattern of other auto-approve settings
+4. **Easier Discovery**: Users can find all related settings together
+
+## Technical Details
+
+- Setting is stored in the global state using key: `commandWhitelist`
+- Synchronized across VS Code instances
+- Supports the same pattern matching as before
+- Maintains backward compatibility through automatic migration

packages/types/src/global-settings.ts

@@ -48,7 +48,6 @@ export const globalSettingsSchema = z.object({
 	alwaysAllowFollowupQuestions: z.boolean().optional(),
 	followupAutoApproveTimeoutMs: z.number().optional(),
 	alwaysAllowUpdateTodoList: z.boolean().optional(),
-	disableLlmCommandSuggestions: z.boolean().optional(),
 	allowedCommands: z.array(z.string()).optional(),
 	deniedCommands: z.array(z.string()).optional(),
 	allowedMaxRequests: z.number().nullish(),

src/core/prompts/tools/__tests__/execute-command.spec.ts

@@ -8,61 +8,25 @@ describe("getExecuteCommandDescription", () => {
 		supportsComputerUse: false,
 	}
 
-	it("should include suggestions section when disableLlmCommandSuggestions is false", () => {
-		const args: ToolArgs = {
-			...baseArgs,
-			settings: {
-				disableLlmCommandSuggestions: false,
-			},
-		}
-
-		const description = getExecuteCommandDescription(args)
-
-		// Check that the description includes the suggestions parameter
-		expect(description).toContain("<suggestions>")
-		expect(description).toContain("- suggestions: (optional) Command patterns for the user to allow/deny")
-		expect(description).toContain("Suggestion Guidelines")
-		// Check for chained command guidance
-		expect(description).toContain("For chained commands")
-		expect(description).toContain("cd backend && npm install")
-	})
-
-	it("should include suggestions section when disableLlmCommandSuggestions is not set", () => {
+	it("should not include suggestions section", () => {
 		const args: ToolArgs = {
 			...baseArgs,
 			settings: {},
 		}
 
 		const description = getExecuteCommandDescription(args)
 
-		// Check that the description includes the suggestions parameter
-		expect(description).toContain("<suggestions>")
-		expect(description).toContain("- suggestions: (optional) Command patterns for the user to allow/deny")
-		expect(description).toContain("Suggestion Guidelines")
-	})
-
-	it("should exclude suggestions section when disableLlmCommandSuggestions is true", () => {
-		const args: ToolArgs = {
-			...baseArgs,
-			settings: {
-				disableLlmCommandSuggestions: true,
-			},
-		}
-
-		const description = getExecuteCommandDescription(args)
-
 		// Check that the description does NOT include the suggestions parameter
 		expect(description).not.toContain("<suggestions>")
 		expect(description).not.toContain("- suggestions: (optional) Command patterns for the user to allow/deny")
 		expect(description).not.toContain("Suggestion Guidelines")
+		expect(description).not.toContain("For chained commands")
 	})
 
-	it("should include basic command and cwd parameters regardless of settings", () => {
+	it("should include basic command and cwd parameters", () => {
 		const args: ToolArgs = {
 			...baseArgs,
-			settings: {
-				disableLlmCommandSuggestions: true,
-			},
+			settings: {},
 		}
 
 		const description = getExecuteCommandDescription(args)
@@ -71,5 +35,21 @@ describe("getExecuteCommandDescription", () => {
 		expect(description).toContain("- command: (required)")
 		expect(description).toContain("- cwd: (optional)")
 		expect(description).toContain("execute_command")
+		expect(description).toContain("/test/path")
+	})
+
+	it("should include usage examples", () => {
+		const args: ToolArgs = {
+			...baseArgs,
+			settings: {},
+		}
+
+		const description = getExecuteCommandDescription(args)
+
+		// Check that usage examples are included
+		expect(description).toContain("Usage:")
+		expect(description).toContain("<execute_command>")
+		expect(description).toContain("Example: Requesting to execute npm run dev")
+		expect(description).toContain("Example: Requesting to execute ls in a specific directory")
 	})
 })

src/core/prompts/tools/execute-command.ts

@@ -1,88 +1,27 @@
 import { ToolArgs } from "./types"
 
 export function getExecuteCommandDescription(args: ToolArgs): string | undefined {
-	const disableLlmSuggestions = args.settings?.disableLlmCommandSuggestions ?? false
-
-	const baseDescription = `## execute_command
+	return `## execute_command
 Description: Request to execute a CLI command on the system. Use this when you need to perform system operations or run specific commands to accomplish any step in the user's task. You must tailor your command to the user's system and provide a clear explanation of what the command does. For command chaining, use the appropriate chaining syntax for the user's shell. Prefer to execute complex CLI commands over creating executable scripts, as they are more flexible and easier to run. Prefer relative commands and paths that avoid location sensitivity for terminal consistency, e.g: \`touch ./testdata/example.file\`, \`dir ./examples/model1/data/yaml\`, or \`go test ./cmd/front --config ./cmd/front/config.yml\`. If directed by the user, you may open a terminal in a different directory by using the \`cwd\` parameter.
 
 Parameters:
 - command: (required) The CLI command to execute. This should be valid for the current operating system. Ensure the command is properly formatted and does not contain any harmful instructions.
-- cwd: (optional) The working directory to execute the command in (default: ${args.cwd})`
-
-	if (disableLlmSuggestions) {
-		return (
-			baseDescription +
-			`
-
-Usage:
-<execute_command>
-<command>Your command here</command>
-<cwd>Working directory path (optional)</cwd>
-</execute_command>
-
-Example: Requesting to execute npm run dev
-<execute_command>
-<command>npm run dev</command>
-</execute_command>
-
-Example: Requesting to execute ls in a specific directory
-<execute_command>
-<command>ls -la</command>
-<cwd>/home/user/projects</cwd>
-</execute_command>`
-		)
-	}
-
-	return (
-		baseDescription +
-		`
-- suggestions: (optional) Command patterns for the user to allow/deny for future auto-approval. Use <suggest> tags.
-
-**Suggestion Guidelines:**
-- Suggestions use prefix matching (case-insensitive)
-- For simple commands: Include the base command (e.g., "npm", "git") and optionally a more specific pattern
-- For chained commands (using &&, ||, ;, |): Include patterns for EACH individual command in the chain (NOT the full chain)
-  - Example: For "cd backend && npm install", suggest: "cd", "npm install", "npm"
-- Include 2-4 relevant patterns total
-- Only suggest "*" (allow all) if explicitly requested by the user
+- cwd: (optional) The working directory to execute the command in (default: ${args.cwd})
 
 Usage:
 <execute_command>
 <command>Your command here</command>
 <cwd>Working directory path (optional)</cwd>
-<suggestions>
-<suggest>pattern 1</suggest>
-<suggest>pattern 2</suggest>
-</suggestions>
 </execute_command>
 
 Example: Requesting to execute npm run dev
 <execute_command>
 <command>npm run dev</command>
-<suggestions>
-<suggest>npm run</suggest>
-<suggest>npm</suggest>
-</suggestions>
-</execute_command>
-
-Example: Requesting to execute a chained command
-<execute_command>
-<command>cd backend && npm install</command>
-<suggestions>
-<suggest>cd</suggest>
-<suggest>npm install</suggest>
-<suggest>npm</suggest>
-</suggestions>
 </execute_command>
 
 Example: Requesting to execute ls in a specific directory
 <execute_command>
 <command>ls -la</command>
 <cwd>/home/user/projects</cwd>
-<suggestions>
-<suggest>ls</suggest>
-</suggestions>
 </execute_command>`
-	)
 }