Files
kandev/apps/web/components/task/task-chat-panel.tsx
T
José Almeida bce9fd266a feat: multi-question support for ask_user_question_kandev (#828)
* feat(backend): multi-question support for ask_user_question_kandev

Reshape the MCP tool from a single prompt+options pair to a questions[]
array (1-4 entries, single-choice each, custom text fallback). The
clarification HTTP handler gates the response until every question has an
answer, fans out one chat message per question (sharing pending_id, with
question_index/question_total in metadata), and returns a JSON map keyed
by question id back to the agent. The orchestrator resume prompt now
formats multi-line Q1/A1, Q2/A2 summaries when the bundle has more than
one question. Mock-agent gets a clarification-multi scenario and
ask_user_question_test.go covers the new schema + map response. No
back-compat shim — local-first tool, single-PR breaking change.

* feat(web): refresh clarification overlay with multi-question + modern UI

Each per-question chat message gets its own card with a "Question N of M"
progress chip, a ✓ Answered badge once committed, and shadcn-aligned
option buttons with kbd hints. A new useClarificationGroup hook coordinates
the per-question commits (decision A) and posts every answer in a single
batch once the user finishes the bundle; X rejects the whole group. The
single-question case keeps the 1-9 / Esc keyboard shortcuts. Playwright
e2e adds 6 multi-question scenarios (rendering, full happy path, partial
state, mixed option/custom, skip-rejects-bundle, sibling-state) on top of
the existing single-question regression suite.

* fix(web): clarification carousel UX with stepper navigation + race fix

Replaces the vertically stacked card layout with a horizontal stepper:
numbered circle dots up top show step state (active/answered/pending),
the single active question card sits in the middle, and Back / Next /
Submit buttons sit at the bottom. Picking an option auto-advances to the
next step (single-question case auto-submits, no explicit Submit button).
The stepper steps are clickable so users can revise prior answers.
Keyboard shortcuts: 1-9 for option pick, ArrowLeft/Right for navigation,
Esc for skip.

Also fixes a race that surfaced as a 400 BadRequest on the
/api/v1/clarification/:id/respond endpoint when the user clicked an
option before all N messages of a multi-question bundle had streamed in:
findPendingClarificationGroup now gates on metadata.question_total and
returns an empty array until the full bundle has arrived.

Playwright suite expanded from 6 to 12 multi-question scenarios:
stepper rendering, auto-advance, Back, stepper jump, Submit gating,
happy path, mixed custom-text, revise-via-jump, skip-from-mid-step,
number-key shortcuts, Esc skip, first-step Back disabled.

* fix(clarification): address CodeRabbit/Greptile/Claude review comments

Backend:
- Single source of truth for question validation: rename/export
  NormalizeAndValidateQuestions and call it from both the HTTP and the
  WebSocket-side MCP handlers, so duplicate question ids and bad option
  counts can't slip through either path.
- httpRespond now validates that every answer references a real question
  id from the bundle and that no id is answered twice (regression tests
  added).
- questionsFromMessages reorders by metadata.question_index so the
  reconstructed bundle preserves the original question order.
- adapters.CreateClarificationRequestMessages rolls back already-created
  messages when a later write in the bundle fails.
- extractQuestionAnswers always returns a JSON map (even on rejection,
  with an envelope { rejected, reject_reason }) and simplifyAnswer
  preserves both selected_option and custom_text when both are present.
- MCP tool description: rewrite the response example to match the actual
  output shape; add minItems / maxItems to the questions and options
  arrays in the JSON Schema.
- Test helper: drop the local contains() shim, use strings.Contains.

Frontend:
- useMemo([props]) was a no-op (props is always a fresh ref); split into
  [message, messages] so it actually memoizes.
- Drop the dead "expired" SubmitState branch — the backend never returns
  410 from this endpoint.
- Remove the asymmetric option-list dimming: the agent now receives both
  selected_option and custom_text together, so a half-grey state was
  misleading.
- New tests: pending-clarification.test.ts covers
  findPendingClarificationGroup including the question_total gate, and
  use-clarification-group.test.ts covers the hook's record/submit/skip
  flows and the override-merge path used by the single-question case.

* fix(web): treat clarification 409 as success so duplicate submit doesn't stick the overlay

* fix(clarification): unblock agent on persistence failure + nil-guard rejected branch

- Add Store.CancelRequest(pendingID) and call it from both the WebSocket
  MCP handler and the HTTP create-request handler when message creation
  fails. Without it, the agent's WaitForResponse parked for the full
  2-hour MCP timeout while no clarification cards ever reached the user.
  Both handlers now return an explicit error response on failure.
- applyAnswersToMessages now guards h.repo == nil in the rejected branch
  (parity with sibling expectedQuestionIDs).
- extractQuestionAnswers stubs for unanswered questions in a rejected
  bundle now carry { answered: false, rejected: true } so agents reading
  per-question payloads see the rejection without inspecting the
  envelope.
- Hook test: added 409-as-ok regression coverage for both submit and
  skip paths so the previous fix can't drift.
- Store test: covers CancelRequest unblocking a parked WaitForResponse.

* fix: drop unused e2e helper + use channel sync in CancelRequest test

* refactor(web): drop dead 'message' prop from clarification overlay

Both call sites pass 'messages' (the bundle group), so the legacy
single-message prop was unused. Tightening the prop type also lets the
useMemo depend on a single value.

* fix(web): ignore modifier-keyed events in clarification carousel shortcuts

Bare 1..9 / ArrowLeft / ArrowRight / Esc still drive the carousel, but
Cmd/Ctrl+1..9 (browser tab switch) and Alt+ArrowLeft (browser back) now
pass through as the user expects.
2026-05-05 19:14:57 +01:00

301 lines
11 KiB
TypeScript

"use client";
import { useCallback, useEffect, useMemo, useRef, useState, memo } from "react";
import { PanelRoot, PanelBody } from "./panel-primitives";
import { useSettingsData } from "@/hooks/domains/settings/use-settings-data";
import { type ChatInputContainerHandle } from "@/components/task/chat/chat-input-container";
import {
QueuedMessageIndicator,
type QueuedMessageIndicatorHandle,
} from "@/components/task/chat/queued-message-indicator";
import { MessageList } from "@/components/task/chat/message-list";
import { useIsTaskArchived } from "./task-archived-context";
import { useChatPanelState } from "./chat/use-chat-panel-state";
import { ChatInputArea, useSubmitHandler, useChatPanelHandlers } from "./chat/chat-input-area";
import { ClarificationInputOverlay } from "./chat/clarification-input-overlay";
import { PanelSearchBar } from "@/components/search/panel-search-bar";
import { SessionSearchHits } from "@/components/task/chat/session-search-hits";
import { usePanelSearch } from "@/hooks/use-panel-search";
import { useSessionSearch } from "@/hooks/domains/session/use-session-search";
import { useLazyLoadMessages } from "@/hooks/use-lazy-load-messages";
import { useAppStore } from "@/components/state-provider";
type QueuedOverlayProps = {
isQueued: boolean;
queuedMessage: { content: string } | null | undefined;
isArchived: boolean;
indicatorRef: React.RefObject<QueuedMessageIndicatorHandle | null>;
onCancel: () => void;
onUpdate: (content: string) => Promise<void>;
onEditComplete: () => void;
};
function QueuedMessageOverlay({
isQueued,
queuedMessage,
isArchived,
indicatorRef,
onCancel,
onUpdate,
onEditComplete,
}: QueuedOverlayProps) {
if (!isQueued || !queuedMessage || isArchived) return null;
return (
<div className="flex-shrink-0 bg-card px-3 pt-1.5">
<QueuedMessageIndicator
ref={indicatorRef}
content={queuedMessage.content}
onCancel={onCancel}
onUpdate={onUpdate}
isVisible={true}
onEditComplete={onEditComplete}
/>
</div>
);
}
function useClarificationKey(agentMessageCount: number) {
const lastCountRef = useRef(agentMessageCount);
const [clarificationKey, setClarificationKey] = useState(0);
useEffect(() => {
lastCountRef.current = agentMessageCount;
}, [agentMessageCount]);
const handleClarificationResolved = useCallback(() => setClarificationKey((k) => k + 1), []);
return { clarificationKey, handleClarificationResolved };
}
function SessionSearchOverlay({
search,
agentLabel,
agentName,
}: {
search: ReturnType<typeof useSessionSearch>;
agentLabel: string | null;
agentName: string | null;
}) {
const currentIdx = search.activeHitId
? search.hits.findIndex((h) => h.id === search.activeHitId)
: -1;
const total = search.hits.length;
const handleNext = useCallback(() => {
if (!total) return;
const next = search.hits[(Math.max(currentIdx, -1) + 1) % total];
if (next) search.setActiveHit(next.id);
}, [search, currentIdx, total]);
const handlePrev = useCallback(() => {
if (!total) return;
const prevIdx = (Math.max(currentIdx, 0) - 1 + total) % total;
const prev = search.hits[prevIdx];
if (prev) search.setActiveHit(prev.id);
}, [search, currentIdx, total]);
if (!search.isOpen) return null;
return (
<div className="absolute top-2 right-2 z-20 flex flex-col items-end gap-1">
<PanelSearchBar
className="static"
value={search.query}
onChange={search.setQuery}
onNext={handleNext}
onPrev={handlePrev}
onClose={search.close}
matchInfo={{ current: currentIdx >= 0 ? currentIdx + 1 : 0, total }}
isLoading={search.isSearching}
// Session search already debounces in useDebouncedSearch; skip the
// bar's debounce so we don't stack 150ms + 180ms per keystroke.
debounceMs={0}
/>
<SessionSearchHits
hits={search.hits}
query={search.query}
activeHitId={search.activeHitId}
onSelect={search.setActiveHit}
isSearching={search.isSearching}
agentLabel={agentLabel}
agentName={agentName}
/>
</div>
);
}
/** Returns the AgentProfileOption for the session's profile, or null. Uses
* primitive profile id to avoid getSnapshot-cache errors from returning
* fresh objects on every selector call. */
function useSessionAgentProfile(sessionId: string | null | undefined) {
const profileId = useAppStore((state) =>
sessionId ? (state.taskSessions.items[sessionId]?.agent_profile_id ?? null) : null,
);
return useAppStore((state) =>
profileId
? (state.agentProfiles.items.find((p: { id: string }) => p.id === profileId) ?? null)
: null,
);
}
/** Resolves the agent profile name + registry slug for the given session.
* Label is "Profile Name" from the "Agent • Profile Name" store label; slug
* feeds <AgentLogo> which fetches the logo by agent type. */
function useSessionAgentIdentity(sessionId: string | null | undefined): {
label: string | null;
name: string | null;
} {
const profile = useSessionAgentProfile(sessionId);
return useMemo(() => {
if (!profile) return { label: null, name: null };
const parts = profile.label.split(" \u2022 ");
const label = parts[1] || parts[0] || profile.label;
return { label, name: profile.agent_name ?? null };
}, [profile]);
}
type TaskChatPanelProps = {
onSend?: (message: string) => void;
sessionId?: string | null;
onOpenFile?: (path: string) => void;
showRequestChangesTooltip?: boolean;
onRequestChangesTooltipDismiss?: () => void;
/** Callback to open a file at a specific line (for comment clicks) */
onOpenFileAtLine?: (filePath: string) => void;
/** Hide the sessions dropdown (session tabs in dockview replace it) */
hideSessionsDropdown?: boolean;
};
// eslint-disable-next-line max-lines-per-function -- composes many sub-panels; each concern already factored into its own hook
export const TaskChatPanel = memo(function TaskChatPanel({
onSend,
sessionId = null,
onOpenFile,
showRequestChangesTooltip = false,
onRequestChangesTooltipDismiss,
onOpenFileAtLine,
hideSessionsDropdown,
}: TaskChatPanelProps) {
const isArchived = useIsTaskArchived();
const chatInputRef = useRef<ChatInputContainerHandle>(null);
const queuedMessageRef = useRef<QueuedMessageIndicatorHandle>(null);
useSettingsData(true);
const panelState = useChatPanelState({ sessionId, onOpenFile, onOpenFileAtLine });
const { isSending, handleSubmit } = useSubmitHandler(panelState, onSend);
const {
resolvedSessionId,
session,
task,
taskId,
isWorking,
messagesLoading,
groupedItems,
allMessages,
footerActionMessages,
permissionsByToolCallId,
childrenByParentToolCallId,
agentMessageCount,
cancelQueue,
pendingClarification,
pendingClarificationGroup,
isQueued,
queuedMessage,
updateQueueContent,
} = panelState;
const { handleCancelTurn, handleCancelQueue, handleQueueEditComplete } = useChatPanelHandlers(
resolvedSessionId,
cancelQueue,
chatInputRef,
);
const { clarificationKey, handleClarificationResolved } = useClarificationKey(agentMessageCount);
const panelRef = useRef<HTMLDivElement>(null);
const { loadMore } = useLazyLoadMessages(resolvedSessionId);
const search = useSessionSearch(resolvedSessionId, loadMore);
const { label: agentLabel, name: agentName } = useSessionAgentIdentity(resolvedSessionId);
usePanelSearch({
containerRef: panelRef,
isOpen: search.isOpen,
onOpen: search.open,
onClose: search.close,
});
// The message list has no focus-capturing child (unlike TipTap/xterm in the
// plan/terminal panels), so clicking a message leaves focus on <body>. Make
// the panel root itself focusable and route non-interactive clicks to it so
// Ctrl+F can detect focus within the session panel.
const handlePanelMouseDown = useCallback((e: React.MouseEvent<HTMLDivElement>) => {
const target = e.target as HTMLElement | null;
if (!target) return;
// Exclude `tabindex="-1"` so we don't match PanelRoot itself (which is
// marked focus-receivable but shouldn't short-circuit this handler).
if (
target.closest(
"input, textarea, select, button, a, [contenteditable], [tabindex]:not([tabindex='-1'])",
)
) {
return;
}
panelRef.current?.focus({ preventScroll: true });
}, []);
return (
<PanelRoot
ref={panelRef}
data-testid="session-chat"
data-panel-kind="session"
tabIndex={-1}
onMouseDown={handlePanelMouseDown}
className="outline-none"
>
<PanelBody padding={false} className="relative">
<MessageList
items={groupedItems}
messages={allMessages}
footerActionMessages={footerActionMessages}
permissionsByToolCallId={permissionsByToolCallId}
childrenByParentToolCallId={childrenByParentToolCallId}
taskId={taskId ?? undefined}
sessionId={resolvedSessionId}
messagesLoading={messagesLoading}
isWorking={isWorking}
sessionState={session?.state}
taskState={task?.state}
worktreePath={session?.worktree_path}
onOpenFile={onOpenFile}
/>
<SessionSearchOverlay search={search} agentLabel={agentLabel} agentName={agentName} />
</PanelBody>
{pendingClarification && !isArchived && (
<div className="flex-shrink-0 border-t border-sky-400/30 bg-card px-1">
<ClarificationInputOverlay
messages={pendingClarificationGroup}
onResolved={handleClarificationResolved}
/>
</div>
)}
<QueuedMessageOverlay
isQueued={isQueued}
queuedMessage={queuedMessage}
isArchived={isArchived}
indicatorRef={queuedMessageRef}
onCancel={handleCancelQueue}
onUpdate={updateQueueContent}
onEditComplete={handleQueueEditComplete}
/>
{isArchived ? (
<div className="bg-muted/50 flex-shrink-0 px-4 py-3 text-center text-sm text-muted-foreground border-t">
This task is archived and read-only.
</div>
) : (
<ChatInputArea
chatInputRef={chatInputRef}
clarificationKey={clarificationKey}
onClarificationResolved={handleClarificationResolved}
handleSubmit={handleSubmit}
handleCancelTurn={handleCancelTurn}
showRequestChangesTooltip={showRequestChangesTooltip}
onRequestChangesTooltipDismiss={onRequestChangesTooltipDismiss}
panelState={panelState}
isSending={isSending}
hideSessionsDropdown={hideSessionsDropdown}
/>
)}
</PanelRoot>
);
});