feat (core): Implement tracker related SI changes

packages/core/src/prompts/promptProvider.ts

@@ -171,6 +171,7 @@ export class PromptProvider {
           }),
           isPlanMode,
         ),
+        taskTracker: config.isTrackerEnabled(),
         operationalGuidelines: this.withSection(
           'operationalGuidelines',
           () => ({

packages/core/src/prompts/snippets.legacy.ts

@@ -18,6 +18,9 @@ import {
   SHELL_TOOL_NAME,
   WRITE_FILE_TOOL_NAME,
   WRITE_TODOS_TOOL_NAME,
+  TRACKER_CREATE_TASK_TOOL_NAME,
+  TRACKER_LIST_TASKS_TOOL_NAME,
+  TRACKER_UPDATE_TASK_TOOL_NAME,
 } from '../tools/tool-names.js';
 
 // --- Options Structs ---
@@ -30,6 +33,7 @@ export interface SystemPromptOptions {
   hookContext?: boolean;
   primaryWorkflows?: PrimaryWorkflowsOptions;
   planningWorkflow?: PlanningWorkflowOptions;
+  taskTracker?: boolean;
   operationalGuidelines?: OperationalGuidelinesOptions;
   sandbox?: SandboxMode;
   interactiveYoloMode?: boolean;
@@ -113,6 +117,8 @@ ${
     : renderPrimaryWorkflows(options.primaryWorkflows)
 }
 
+${options.taskTracker ? renderTaskTracker() : ''}
+
 ${renderOperationalGuidelines(options.operationalGuidelines)}
 
 ${renderInteractiveYoloMode(options.interactiveYoloMode)}
@@ -388,6 +394,23 @@ ${trimmed}
   return `\n---\n\n<loaded_context>\n${sections.join('\n')}\n</loaded_context>`;
 }
 
+export function renderTaskTracker(): string {
+  const trackerCreate = `\`${TRACKER_CREATE_TASK_TOOL_NAME}\``;
+  const trackerList = `\`${TRACKER_LIST_TASKS_TOOL_NAME}\``;
+  const trackerUpdate = `\`${TRACKER_UPDATE_TASK_TOOL_NAME}\``;
+
+  return `
+# TASK MANAGEMENT PROTOCOL
+You are operating with a persistent file-based task tracking system located at \`.tracker/tasks/\`. You must adhere to the following rules:
+
+1.  **NO IN-MEMORY LISTS**: Do not maintain a mental list of tasks or write markdown checkboxes in the chat. Use the provided tools (${trackerCreate}, ${trackerList}, ${trackerUpdate}) for all state management.
+2.  **ATOMICITY**: If a user request involves multiple steps (e.g., "Refactor the backend") or represents a complex task, you must first break this into discrete entries using ${trackerCreate} before writing any code.
+3.  **PLAN MODE INTEGRATION**: If you construct or are following a documented plan (e.g., in Plan Mode), use the tracker to represent the granular execution state of that plan. Maintain a bidirectional understanding between the plan document and the task graph.
+4.  **VERIFICATION**: Before marking a task as complete, verify the work is actually done (e.g., run the test, check the file existence).
+5.  **STATE OVER CHAT**: If the user says "I think we finished that," but the tool says it is 'pending', trust the tool--or verify explicitly before updating.
+6.  **DEPENDENCY MANAGEMENT**: Respect task topology. Never attempt to execute a task if its dependencies are not marked as 'closed'. If you are blocked, focus only on the leaf nodes of the task graph.`.trim();
+}
+
 export function renderPlanningWorkflow(
   options?: PlanningWorkflowOptions,
 ): string {

packages/core/src/prompts/snippets.ts

@@ -27,6 +27,9 @@ import {
   READ_FILE_PARAM_END_LINE,
   SHELL_PARAM_IS_BACKGROUND,
   EDIT_PARAM_OLD_STRING,
+  TRACKER_CREATE_TASK_TOOL_NAME,
+  TRACKER_LIST_TASKS_TOOL_NAME,
+  TRACKER_UPDATE_TASK_TOOL_NAME,
 } from '../tools/tool-names.js';
 import type { HierarchicalMemory } from '../config/memory.js';
 import { DEFAULT_CONTEXT_FILENAME } from '../tools/memoryTool.js';
@@ -41,6 +44,7 @@ export interface SystemPromptOptions {
   hookContext?: boolean;
   primaryWorkflows?: PrimaryWorkflowsOptions;
   planningWorkflow?: PlanningWorkflowOptions;
+  taskTracker?: boolean;
   operationalGuidelines?: OperationalGuidelinesOptions;
   sandbox?: SandboxMode;
   interactiveYoloMode?: boolean;
@@ -120,6 +124,8 @@ ${
     : renderPrimaryWorkflows(options.primaryWorkflows)
 }
 
+${options.taskTracker ? renderTaskTracker() : ''}
+
 ${renderOperationalGuidelines(options.operationalGuidelines)}
 
 ${renderInteractiveYoloMode(options.interactiveYoloMode)}
@@ -464,6 +470,24 @@ ${trimmed}
   return `\n---\n\n<loaded_context>\n${sections.join('\n')}\n</loaded_context>`;
 }
 
+export function renderTaskTracker(): string {
+  const trackerCreate = formatToolName(TRACKER_CREATE_TASK_TOOL_NAME);
+  const trackerList = formatToolName(TRACKER_LIST_TASKS_TOOL_NAME);
+  const trackerUpdate = formatToolName(TRACKER_UPDATE_TASK_TOOL_NAME);
+
+  return `
+# TASK MANAGEMENT PROTOCOL
+You are operating with a persistent file-based task tracking system located at \`.tracker/tasks/\`. You must adhere to the following rules:
+
+1.  **NO IN-MEMORY LISTS**: Do not maintain a mental list of tasks or write markdown checkboxes in the chat. Use the provided tools (${trackerCreate}, ${trackerList}, ${trackerUpdate}) for all state management.
+2.  **IMMEDIATE DECOMPOSITION**: Upon receiving a task, evaluate its functional complexity and scope. If the request involves more than a single atomic modification, or necessitates research before execution, you MUST immediately decompose it into discrete entries using ${trackerCreate}.
+3.  **IGNORE FORMATTING BIAS**: Trigger the protocol based on the **objective complexity** of the goal, regardless of whether the user provided a structured list or a single block of text/paragraph. "Paragraph-style" goals that imply multiple actions are multi-step projects and MUST be tracked.
+4.  **PLAN MODE INTEGRATION**: If you construct or are following a documented plan (e.g., in Plan Mode), use the tracker to represent the granular execution state of that plan. Maintain a bidirectional understanding between the plan document and the task graph.
+5.  **VERIFICATION**: Before marking a task as complete, verify the work is actually done (e.g., run the test, check the file existence).
+6.  **STATE OVER CHAT**: If the user says "I think we finished that," but the tool says it is 'pending', trust the tool--or verify explicitly before updating.
+7.  **DEPENDENCY MANAGEMENT**: Respect task topology. Never attempt to execute a task if its dependencies are not marked as 'closed'. If you are blocked, focus only on the leaf nodes of the task graph.`.trim();
+}
+
 export function renderPlanningWorkflow(
   options?: PlanningWorkflowOptions,
 ): string {

packages/core/src/tools/tool-names.ts

@@ -154,19 +154,19 @@ export const LS_TOOL_NAME_LEGACY = 'list_directory'; // Just to be safe if anyth
 
 export const EDIT_TOOL_NAMES = new Set([EDIT_TOOL_NAME, WRITE_FILE_TOOL_NAME]);
 
-export const TRACKER_CREATE_TASK_TOOL_NAME = 'tracker_create_task';
-export const TRACKER_UPDATE_TASK_TOOL_NAME = 'tracker_update_task';
-export const TRACKER_GET_TASK_TOOL_NAME = 'tracker_get_task';
-export const TRACKER_LIST_TASKS_TOOL_NAME = 'tracker_list_tasks';
-export const TRACKER_ADD_DEPENDENCY_TOOL_NAME = 'tracker_add_dependency';
-export const TRACKER_VISUALIZE_TOOL_NAME = 'tracker_visualize';
-
 // Tool Display Names
 export const WRITE_FILE_DISPLAY_NAME = 'WriteFile';
 export const EDIT_DISPLAY_NAME = 'Edit';
 export const ASK_USER_DISPLAY_NAME = 'Ask User';
 export const READ_FILE_DISPLAY_NAME = 'ReadFile';
 export const GLOB_DISPLAY_NAME = 'FindFiles';
+export const TRACKER_INIT_TOOL_NAME = 'tracker_init';
+export const TRACKER_CREATE_TASK_TOOL_NAME = 'tracker_create_task';
+export const TRACKER_UPDATE_TASK_TOOL_NAME = 'tracker_update_task';
+export const TRACKER_GET_TASK_TOOL_NAME = 'tracker_get_task';
+export const TRACKER_LIST_TASKS_TOOL_NAME = 'tracker_list_tasks';
+export const TRACKER_ADD_DEPENDENCY_TOOL_NAME = 'tracker_add_dependency';
+export const TRACKER_VISUALIZE_TOOL_NAME = 'tracker_visualize';
 
 /**
  * Mapping of legacy tool names to their current names.
@@ -229,6 +229,13 @@ export const ALL_BUILTIN_TOOL_NAMES = [
   GET_INTERNAL_DOCS_TOOL_NAME,
   ENTER_PLAN_MODE_TOOL_NAME,
   EXIT_PLAN_MODE_TOOL_NAME,
+  TRACKER_INIT_TOOL_NAME,
+  TRACKER_CREATE_TASK_TOOL_NAME,
+  TRACKER_UPDATE_TASK_TOOL_NAME,
+  TRACKER_GET_TASK_TOOL_NAME,
+  TRACKER_LIST_TASKS_TOOL_NAME,
+  TRACKER_ADD_DEPENDENCY_TOOL_NAME,
+  TRACKER_VISUALIZE_TOOL_NAME,
 ] as const;
 
 /**