fix: update additional missed <task> and <feedback> usages

- src/services/command/built-in-commands.ts: Change <task> to <user_message> in init command
- src/core/tools/ExecuteCommandTool.ts: Change <feedback> to <user_message> for running command feedback
- src/core/mentions/processUserContentMentions.ts: Update comment to reflect new tag
- src/core/tools/__tests__/readFileTool.spec.ts: Update mock implementations

Note: <task> tags in fetch_instructions tool remain unchanged as they are
XML tool parameters (parameter name = 'task'), not user content wrappers.

src/core/mentions/processUserContentMentions.ts

@@ -30,13 +30,10 @@ export async function processUserContentMentions({
 	// Process userContent array, which contains various block types:
 	// TextBlockParam, ImageBlockParam, ToolUseBlockParam, and ToolResultBlockParam.
 	// We need to apply parseMentions() to:
-	// 1. All TextBlockParam's text (first user message with task)
+	// 1. All TextBlockParam's text (first user message)
 	// 2. ToolResultBlockParam's content/context text arrays if it contains
-	// "<feedback>" (see formatToolDeniedFeedback, attemptCompletion,
-	// executeCommand, and consecutiveMistakeCount >= 3) or "<answer>"
-	// (see askFollowupQuestion), we place all user generated content in
-	// these tags so they can effectively be used as markers for when we
-	// should parse mentions).
+	// "<user_message>" - we place all user generated content in this tag
+	// so it can effectively be used as a marker for when we should parse mentions.
 	return Promise.all(
 		userContent.map(async (block) => {
 			const shouldProcessMentions = (text: string) => text.includes("<user_message>")

src/core/tools/ExecuteCommandTool.ts

@@ -337,8 +337,7 @@ export async function executeCommandInTerminal(
 				[
 					`Command is still running in terminal from '${terminal.getCurrentWorkingDirectory().toPosix()}'.`,
 					result.length > 0 ? `Here's the output so far:\n${result}\n` : "\n",
-					`The user provided the following feedback:`,
-					`<feedback>\n${text}\n</feedback>`,
+					`<user_message>\n${text}\n</user_message>`,
 				].join("\n"),
 				images,
 			),

src/core/tools/__tests__/readFileTool.spec.ts

@@ -86,14 +86,8 @@ const { toolResultMock, imageBlocksMock } = vi.hoisted(() => {
 vi.mock("../../prompts/responses", () => ({
 	formatResponse: {
 		toolDenied: vi.fn(() => "The user denied this operation."),
-		toolDeniedWithFeedback: vi.fn(
-			(feedback?: string) =>
-				`The user denied this operation and provided the following feedback:\n<feedback>\n${feedback}\n</feedback>`,
-		),
-		toolApprovedWithFeedback: vi.fn(
-			(feedback?: string) =>
-				`The user approved this operation and provided the following context:\n<feedback>\n${feedback}\n</feedback>`,
-		),
+		toolDeniedWithFeedback: vi.fn((feedback?: string) => `<user_message>\n${feedback}\n</user_message>`),
+		toolApprovedWithFeedback: vi.fn((feedback?: string) => `<user_message>\n${feedback}\n</user_message>`),
 		rooIgnoreError: vi.fn(
 			(path: string) =>
 				`Access to ${path} is blocked by the .rooignore file settings. You must try to continue in the task without using this file, or ask the user to update the .rooignore file.`,

src/services/command/built-in-commands.ts

@@ -11,11 +11,11 @@ const BUILT_IN_COMMANDS: Record<string, BuiltInCommandDefinition> = {
 	init: {
 		name: "init",
 		description: "Analyze codebase and create concise AGENTS.md files for AI assistants",
-		content: `<task>
+		content: `<user_message>
 Please analyze this codebase and create an AGENTS.md file containing:
 1. Build/lint/test commands - especially for running a single test
 2. Code style guidelines including imports, formatting, types, naming conventions, error handling, etc.
-</task>
+</user_message>
 
 <initialization>
   <purpose>