fix: address todo follow-up review

packages/coding-agent/CHANGELOG.md

@@ -14,6 +14,7 @@
 - Fixed llama.cpp discovery mapping unlimited `max_tokens = -1` / `n_predict = -1` output limits to the generic 32K discovery cap instead of the discovered runtime context window. ([#3781](https://github.com/can1357/oh-my-pi/issues/3781))
 - Fixed the bash interceptor blocking `echo` / `printf` redirects to `/dev/null`, `/dev/tty`, `/dev/stdout`, and `/dev/stderr` device sinks while still directing real file writes to the write tool. ([#3763](https://github.com/can1357/oh-my-pi/issues/3763))
 - Fixed the `edit` tool persisting unbounded full-file `oldText` / `newText` snapshots in tool-result `details`, inflating per-turn session JSONL lines (hundreds of KB per edit on large files). `details.oldText`/`details.newText` are now pruned when their combined length exceeds 32 KB; the visible diff, path, line, and diagnostic metadata are preserved, and ACP `diff` content still flows for smaller edits. ([#3786](https://github.com/can1357/oh-my-pi/issues/3786))
+- Fixed todo HUD and goal-mode continuations preserving accurate persisted todo context across `/btw` branches, hidden goal prompts, and discovery-only todo tool states.
 
 ## [16.2.5] - 2026-06-28
 

packages/coding-agent/test/goals/goal-mode-integration.test.ts

@@ -10,6 +10,7 @@ import { initTheme } from "@oh-my-pi/pi-coding-agent/modes/theme/theme";
 import { AgentSession } from "@oh-my-pi/pi-coding-agent/session/agent-session";
 import { AuthStorage } from "@oh-my-pi/pi-coding-agent/session/auth-storage";
 import { SessionManager } from "@oh-my-pi/pi-coding-agent/session/session-manager";
+import type { DiscoverableTool } from "@oh-my-pi/pi-coding-agent/tool-discovery/tool-index";
 import { createTools, type Tool, type ToolSession } from "@oh-my-pi/pi-coding-agent/tools";
 import type { TodoPhase } from "@oh-my-pi/pi-coding-agent/tools/todo";
 import { TempDir } from "@oh-my-pi/pi-utils";
@@ -338,7 +339,7 @@ describe("InteractiveMode goal mode integration", () => {
 			isToolDiscoveryEnabled: () => harness.session.isToolDiscoveryEnabled(),
 			getSelectedDiscoveredToolNames: () => harness.session.getSelectedDiscoveredToolNames(),
 			activateDiscoveredTools: (toolNames: string[]) => harness.session.activateDiscoveredTools(toolNames),
-			getDiscoverableTools: (filter?: Parameters<AgentSession["getDiscoverableTools"]>[0]) =>
+			getDiscoverableTools: (filter?: { source?: DiscoverableTool["source"] }) =>
 				harness.session.getDiscoverableTools(filter),
 		});
 		for (const tool of await createTools(harness.toolSession, ["search_tool_bm25"])) {

packages/coding-agent/test/interactive-mode-todo-clear.test.ts

@@ -1,6 +1,7 @@
 import { afterEach, beforeAll, beforeEach, describe, expect, it, vi } from "bun:test";
 import * as path from "node:path";
 import { Agent } from "@oh-my-pi/pi-agent-core";
+import type { AssistantMessage } from "@oh-my-pi/pi-ai";
 import { ModelRegistry } from "@oh-my-pi/pi-coding-agent/config/model-registry";
 import { resetSettingsForTest, Settings } from "@oh-my-pi/pi-coding-agent/config/settings";
 import { InteractiveMode } from "@oh-my-pi/pi-coding-agent/modes/interactive-mode";
@@ -193,7 +194,7 @@ describe("InteractiveMode todo HUD persistence", () => {
 		vi.spyOn(mode, "updateEditorBorderColor").mockImplementation(() => {});
 		vi.spyOn(mode, "showStatus").mockImplementation(() => {});
 
-		await mode.handleBtwBranch("why did this fail?", {} as Parameters<InteractiveMode["handleBtwBranch"]>[1]);
+		await mode.handleBtwBranch("why did this fail?", {} as AssistantMessage);
 
 		expect(mode.todoReminderContainer.children).toHaveLength(0);
 	});