feat: add tiered tool-repetition detection (soft warning + hard stop) #772
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,261 @@ | ||
| // npx vitest src/core/assistant-message/__tests__/presentAssistantMessage-tool-repetition.spec.ts | ||
|
|
||
| import { presentAssistantMessage } from "../presentAssistantMessage" | ||
|
|
||
| // Mock dependencies | ||
| vi.mock("../../task/Task") | ||
| vi.mock("../../tools/validateToolUse", () => ({ | ||
| validateToolUse: vi.fn(), | ||
| isValidToolName: vi.fn(() => true), | ||
| })) | ||
|
|
||
| // Mock the read_file tool so we can assert the normal execution path is taken | ||
| // when the repetition detector allows a tool call. The handle spy simulates a | ||
| // successful tool run by pushing a tool_result, mirroring the real tool. | ||
| // `vi.hoisted` is required because `vi.mock` factories are hoisted above | ||
| // top-level variable declarations. | ||
| const { readFileHandle } = vi.hoisted(() => ({ | ||
| readFileHandle: vi.fn(async (_cline: any, block: any, { pushToolResult }: any) => { | ||
| pushToolResult(`[read_file for '${block?.params?.path ?? block?.nativeArgs?.path}'] Result`) | ||
| }), | ||
| })) | ||
|
|
||
| vi.mock("../../tools/ReadFileTool", () => ({ | ||
| readFileTool: { | ||
| handle: readFileHandle, | ||
| getReadFileToolDescription: vi.fn(() => "[read_file]"), | ||
| }, | ||
| })) | ||
|
|
||
| const captureConsecutiveMistakeError = vi.fn() | ||
| const captureException = vi.fn() | ||
| const captureToolUsage = vi.fn() | ||
|
|
||
| vi.mock("@roo-code/telemetry", () => ({ | ||
| TelemetryService: { | ||
| instance: { | ||
| get captureToolUsage() { | ||
| return captureToolUsage | ||
| }, | ||
| get captureConsecutiveMistakeError() { | ||
| return captureConsecutiveMistakeError | ||
| }, | ||
| get captureException() { | ||
| return captureException | ||
| }, | ||
| captureEvent: vi.fn(), | ||
| }, | ||
| }, | ||
| })) | ||
|
|
||
| describe("presentAssistantMessage - Tool Repetition Detection", () => { | ||
| let mockTask: any | ||
|
|
||
| beforeEach(() => { | ||
| vi.clearAllMocks() | ||
|
|
||
| mockTask = { | ||
| taskId: "test-task-id", | ||
| instanceId: "test-instance", | ||
| abort: false, | ||
| presentAssistantMessageLocked: false, | ||
| presentAssistantMessageHasPendingUpdates: false, | ||
| currentStreamingContentIndex: 0, | ||
| assistantMessageContent: [], | ||
| userMessageContent: [], | ||
| didCompleteReadingStream: false, | ||
| didRejectTool: false, | ||
| didAlreadyUseTool: false, | ||
| consecutiveMistakeCount: 0, | ||
| consecutiveMistakeLimit: 5, | ||
| clineMessages: [], | ||
| apiConfiguration: { apiProvider: "anthropic" }, | ||
| api: { | ||
| getModel: () => ({ id: "test-model", info: {} }), | ||
| }, | ||
| recordToolUsage: vi.fn(), | ||
| recordToolError: vi.fn(), | ||
| toolRepetitionDetector: { | ||
| check: vi.fn().mockReturnValue({ action: "allow" }), | ||
| }, | ||
| providerRef: { | ||
| deref: () => ({ | ||
| getState: vi.fn().mockResolvedValue({ | ||
| mode: "code", | ||
| customModes: [], | ||
| }), | ||
| }), | ||
| }, | ||
| say: vi.fn().mockResolvedValue(undefined), | ||
| ask: vi.fn().mockResolvedValue({ response: "yesButtonClicked" }), | ||
| } | ||
|
|
||
| mockTask.pushToolResultToUserContent = vi.fn().mockImplementation((toolResult: any) => { | ||
| const existingResult = mockTask.userMessageContent.find( | ||
| (block: any) => block.type === "tool_result" && block.tool_use_id === toolResult.tool_use_id, | ||
| ) | ||
| if (existingResult) { | ||
| return false | ||
| } | ||
| mockTask.userMessageContent.push(toolResult) | ||
| return true | ||
| }) | ||
| }) | ||
|
|
||
| it("should soft block without involving the user and return an error to the model", async () => { | ||
| const toolCallId = "tool_call_soft_block" | ||
| mockTask.assistantMessageContent = [ | ||
| { | ||
| type: "tool_use", | ||
| id: toolCallId, | ||
| name: "read_file", | ||
| params: { path: "test.txt" }, | ||
| nativeArgs: { path: "test.txt" }, | ||
| partial: false, | ||
| }, | ||
| ] | ||
|
|
||
| mockTask.toolRepetitionDetector.check = vi.fn().mockReturnValue({ | ||
| action: "soft_block", | ||
| message: "The tool 'read_file' was blocked because it was just called with identical parameters.", | ||
| }) | ||
|
|
||
| await presentAssistantMessage(mockTask) | ||
|
|
||
| // The user should NOT have been asked anything for a soft block. | ||
| expect(mockTask.ask).not.toHaveBeenCalled() | ||
|
|
||
| // A tool_result with the soft block message should have been pushed. | ||
| const toolResult = mockTask.userMessageContent.find( | ||
| (item: any) => item.type === "tool_result" && item.tool_use_id === toolCallId, | ||
| ) | ||
| expect(toolResult).toBeDefined() | ||
| expect(toolResult.content).toContain("read_file") | ||
| expect(toolResult.content).toContain("identical parameters") | ||
|
|
||
| // No telemetry escalation for a soft block. | ||
| expect(captureConsecutiveMistakeError).not.toHaveBeenCalled() | ||
| expect(captureException).not.toHaveBeenCalled() | ||
| }) | ||
|
|
||
| it("should hard block, ask the user for guidance, and record telemetry", async () => { | ||
| const toolCallId = "tool_call_hard_block" | ||
| mockTask.assistantMessageContent = [ | ||
| { | ||
| type: "tool_use", | ||
| id: toolCallId, | ||
| name: "read_file", | ||
| params: { path: "test.txt" }, | ||
| nativeArgs: { path: "test.txt" }, | ||
| partial: false, | ||
| }, | ||
| ] | ||
|
|
||
| mockTask.toolRepetitionDetector.check = vi.fn().mockReturnValue({ | ||
| action: "hard_block", | ||
| askUser: { | ||
| messageKey: "mistake_limit_reached", | ||
| messageDetail: "Roo appears to be stuck in a loop calling {toolName} repeatedly.", | ||
| }, | ||
| }) | ||
|
|
||
| mockTask.ask = vi.fn().mockResolvedValue({ response: "yesButtonClicked" }) | ||
|
|
||
| await presentAssistantMessage(mockTask) | ||
|
|
||
| // The user must be asked for guidance with the resolved message key. | ||
| expect(mockTask.ask).toHaveBeenCalledWith( | ||
| "mistake_limit_reached", | ||
| expect.stringContaining("read_file"), | ||
| ) | ||
|
|
||
| // Telemetry escalation should have fired for a hard block. | ||
| expect(captureConsecutiveMistakeError).toHaveBeenCalledWith("test-task-id") | ||
| expect(captureException).toHaveBeenCalled() | ||
|
|
||
| // A tool_result describing the repetition limit should have been pushed. | ||
| const toolResult = mockTask.userMessageContent.find( | ||
| (item: any) => item.type === "tool_result" && item.tool_use_id === toolCallId, | ||
| ) | ||
| expect(toolResult).toBeDefined() | ||
| expect(toolResult.content).toContain("read_file") | ||
| }) | ||
|
|
||
| it("should incorporate user feedback when the user responds to a hard block", async () => { | ||
| const toolCallId = "tool_call_hard_block_feedback" | ||
| mockTask.assistantMessageContent = [ | ||
| { | ||
| type: "tool_use", | ||
| id: toolCallId, | ||
| name: "read_file", | ||
| params: { path: "test.txt" }, | ||
| nativeArgs: { path: "test.txt" }, | ||
| partial: false, | ||
| }, | ||
| ] | ||
|
|
||
| mockTask.toolRepetitionDetector.check = vi.fn().mockReturnValue({ | ||
| action: "hard_block", | ||
| askUser: { | ||
| messageKey: "mistake_limit_reached", | ||
| messageDetail: "Stuck calling {toolName} repeatedly.", | ||
| }, | ||
| }) | ||
|
|
||
| mockTask.ask = vi.fn().mockResolvedValue({ | ||
| response: "messageResponse", | ||
| text: "try a different file", | ||
| images: [], | ||
| }) | ||
|
|
||
| await presentAssistantMessage(mockTask) | ||
|
|
||
| // User feedback should have been surfaced to the chat. | ||
| expect(mockTask.say).toHaveBeenCalledWith("user_feedback", "try a different file", []) | ||
|
|
||
| // And appended to the user message content. | ||
| const feedbackBlock = mockTask.userMessageContent.find( | ||
| (item: any) => item.type === "text" && String(item.text).includes("try a different file"), | ||
| ) | ||
| expect(feedbackBlock).toBeDefined() | ||
| expect(feedbackBlock.text).toContain("Tool repetition limit reached") | ||
| }) | ||
|
|
||
| it("should execute the tool normally when the detector allows it", async () => { | ||
| const toolCallId = "tool_call_allow" | ||
| mockTask.assistantMessageContent = [ | ||
| { | ||
| type: "tool_use", | ||
| id: toolCallId, | ||
| name: "read_file", | ||
| params: { path: "test.txt" }, | ||
| nativeArgs: { path: "test.txt" }, | ||
| partial: false, | ||
| }, | ||
| ] | ||
|
|
||
| mockTask.toolRepetitionDetector.check = vi.fn().mockReturnValue({ action: "allow" }) | ||
|
|
||
| await presentAssistantMessage(mockTask) | ||
|
|
||
| // The detector should have been consulted. | ||
| expect(mockTask.toolRepetitionDetector.check).toHaveBeenCalled() | ||
| // No hard block ask should have occurred. | ||
| expect(mockTask.ask).not.toHaveBeenCalledWith("mistake_limit_reached", expect.anything()) | ||
|
|
||
| // The tool must have actually continued into normal execution: the | ||
| // read_file tool runner should have been dispatched with this block. | ||
| expect(readFileHandle).toHaveBeenCalledTimes(1) | ||
| const [, dispatchedBlock] = readFileHandle.mock.calls[0] | ||
| expect(dispatchedBlock).toMatchObject({ name: "read_file", id: toolCallId }) | ||
|
|
||
| // And the normal execution path should have produced a tool_result | ||
| // (no soft/hard block error message). | ||
| const toolResult = mockTask.userMessageContent.find( | ||
| (item: any) => item.type === "tool_result" && item.tool_use_id === toolCallId, | ||
| ) | ||
| expect(toolResult).toBeDefined() | ||
| expect(toolResult.content).toContain("read_file") | ||
| expect(toolResult.is_error).toBeUndefined() | ||
| }) | ||
|
coderabbitai[bot] marked this conversation as resolved.
|
||
| }) | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This checks
askwasn't called and the right tool_result was pushed, but is there an assertion anywhere that the tool itself (readFileHandle) was NOT invoked? A missingbreakafter the soft-block branch would fall through to dispatch and this test would still pass.