v2.0.3: CLI client + human inbox + cost tracking + Boomerang new_task

Phase 7 of v2.0. BooCoder gains a terminal-driven UX and subagent
isolation primitive.

CLI (src/cli.ts): standalone entry point for terminal use.
- boocode run "task" [--agent x] [--model y] — create + stream output
- boocode ls [--state x] — formatted task table
- boocode attach <id> — WS stream of running task
- boocode send <id> "msg" — follow-up message to task session
Connects to BOOCODER_URL (default http://100.114.205.53:9502).

Human inbox (routes/inbox.ts): GET /api/inbox (failed/blocked tasks),
POST /api/inbox/:id/retry (reset to pending for re-dispatch).

Cost tracking: dispatcher aggregates tokens_used from all messages in
the task's session after completion, stores in tasks.cost_tokens.
GET /api/stats/costs?group_by=project|agent|day for aggregation.

Boomerang subagent isolation (3 new tools):
- new_task: creates child task with parent_task_id linkage, runs in
  fresh isolated session. Orchestrator sees only output_summary.
- list_tasks: query child tasks of current parent
- check_task_status: read task state + output_summary

The orchestrator pattern: an agent with tools: [new_task, list_tasks,
check_task_status] can ONLY dispatch — can't read files or MCP. This
is the Roo Code Boomerang Tasks capability-restriction principle.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

apps/coder/src/services/tools/new_task.ts

@@ -0,0 +1,65 @@
+import { z } from 'zod';
+import type { ToolDef, ToolContext } from './types.js';
+import { getInferenceContext } from './inference_context.js';
+
+const NewTaskInput = z.object({
+  input: z.string().min(1).describe('Task description for the child subtask'),
+  agent: z.string().optional().describe('Optional: dispatch to a specific agent'),
+  model: z.string().optional().describe('Optional: model override for the subtask'),
+});
+
+type NewTaskInputT = z.infer<typeof NewTaskInput>;
+
+export const newTaskTool: ToolDef<NewTaskInputT> = {
+  name: 'new_task',
+  description:
+    'Spawn a subtask that runs in isolation. The subtask gets its own session and ' +
+    'worktree. Use check_task_status to monitor progress. Only the output_summary is ' +
+    'accessible to the parent — full isolation (Boomerang pattern).',
+  inputSchema: NewTaskInput,
+  jsonSchema: {
+    type: 'function',
+    function: {
+      name: 'new_task',
+      description:
+        'Spawn a subtask that runs in isolation. The subtask gets its own session and ' +
+        'worktree. Use check_task_status to monitor progress.',
+      parameters: {
+        type: 'object',
+        properties: {
+          input: { type: 'string', description: 'Task description for the child subtask' },
+          agent: { type: 'string', description: 'Optional: dispatch to a specific agent' },
+          model: { type: 'string', description: 'Optional: model override for the subtask' },
+        },
+        required: ['input'],
+      },
+    },
+  },
+
+  async execute(input: NewTaskInputT, _projectRoot: string, context: ToolContext): Promise<unknown> {
+    const { sql } = context;
+    // Get the current task's project_id from the inference context
+    const ctx = getInferenceContext();
+    const currentTaskId = ctx.taskId;
+
+    // Look up the project_id from the current session
+    const [session] = await sql<{ project_id: string }[]>`
+      SELECT project_id FROM sessions WHERE id = ${ctx.sessionId}
+    `;
+    if (!session) {
+      return { error: 'Cannot determine project_id from current session' };
+    }
+
+    const [task] = await sql<{ id: string; state: string }[]>`
+      INSERT INTO tasks (project_id, parent_task_id, input, agent, model)
+      VALUES (${session.project_id}, ${currentTaskId}, ${input.input}, ${input.agent ?? null}, ${input.model ?? null})
+      RETURNING id, state
+    `;
+
+    return {
+      message: `Subtask created (id: ${task!.id}). It will run in isolation. Use check_task_status to monitor.`,
+      task_id: task!.id,
+      state: task!.state,
+    };
+  },
+};