fusion/packages/dashboard/src/mission-interview.ts

/**
 * Mission Interview Session Management
 *
 * Manages AI-guided interview sessions for mission specification.
 * Uses an AI agent to conduct back-and-forth conversations that
 * produce structured mission plans (milestones, slices, features).
 *
 * Architecture mirrors planning.ts but targets the mission hierarchy.
 *
 * Features:
 * - AI agent integration with real-time streaming via SSE
 * - Rate limiting per IP
 * - Session expiration and cleanup
 * - SSE streaming via MissionInterviewStreamManager
 * - Prompt override support for project-level customization
 */

import type { PlanningQuestion, PromptOverrideMap, TaskStore } from "@fusion/core";
import { resolvePrompt } from "@fusion/core";
import { randomUUID } from "node:crypto";
import { EventEmitter } from "node:events";
import type { AiSessionStore, AiSessionRow, AiSessionStatus } from "./ai-session-store.js";
import { SessionEventBuffer, type SessionBufferedEvent } from "./sse-buffer.js";
import {
  createSessionDiagnostics,
  resetDiagnosticsSink,
  nonfatal,
} from "./ai-session-diagnostics.js";
import { GenerationGuard, isAbortError } from "./ai-session-timeout.js";

import { createFnAgent as engineCreateFnAgent } from "@fusion/engine";
import { createPlanningBoardTools } from "./planning-board-tools.js";

// eslint-disable-next-line @typescript-eslint/no-explicit-any
type AgentResult = any;
const MISSION_INTERVIEW_BUILTIN_WEB_TOOLS = ["WebSearch", "WebFetch"] as const;
// eslint-disable-next-line @typescript-eslint/no-explicit-any
const createFnAgent: any = engineCreateFnAgent;

/**
 * Shared diagnostics helper for the mission-interview module.
 * Uses the shared ai-session-diagnostics helper for consistent scoped logging.
 * @see ai-session-diagnostics.ts for the shared contract
 */
const diagnostics = createSessionDiagnostics("mission-interview");

/**
 * Get the current diagnostics logger (for backward compatibility).
 * @internal - exposed for test hook
 */
export function __getMissionInterviewDiagnostics() {
  return diagnostics;
}

/**
 * Inject a diagnostics sink (test-only).
 * Delegates to the shared ai-session-diagnostics sink.
 * When a sink is injected, all mission-interview module diagnostics route through it.
 * This allows tests to assert on diagnostics without global console spies.
 * @internal - exposed for test hook
 */
export function __setMissionInterviewDiagnostics(_logger: unknown): void {
  // For backward compatibility, we keep this function but it now delegates
  // to the shared helper's sink mechanism. The actual sink injection
  // should use setDiagnosticsSink() from ai-session-diagnostics.
  if (_logger === null) {
    resetDiagnosticsSink();
  }
}

function ensureEngineReady(): Promise<void> {
  return Promise.resolve();
}

// ── Constants ───────────────────────────────────────────────────────────────

/** Session TTL in milliseconds (7 days) */
const SESSION_TTL_MS = 7 * 24 * 60 * 60 * 1000;

/** Cleanup interval in milliseconds (5 minutes) */
const CLEANUP_INTERVAL_MS = 5 * 60 * 1000;

/** Max interview sessions per IP per hour */
const MAX_SESSIONS_PER_IP_PER_HOUR = 5;

/** Rate limiting window in milliseconds (1 hour) */
const RATE_LIMIT_WINDOW_MS = 60 * 60 * 1000;

/** Max number of retry attempts when AI returns unparseable output */
const MAX_PARSE_RETRIES = 1;

/**
 * Per-turn generation timeout. Mission interview turns produce larger plans
 * than planning, so this is more generous than planning's 120 s. Bounds the
 * worst case for a silently-stalled model stream or hung tool call.
 */
export const GENERATION_TIMEOUT_MS = 180_000;

const generationGuard = new GenerationGuard();

const MISSION_INTERVIEW_DRAFT_STATUSES = ["generating", "awaiting_input", "error"] as const;

function isMissionInterviewDraftStatus(
  status: AiSessionStatus,
): status is Extract<AiSessionStatus, (typeof MISSION_INTERVIEW_DRAFT_STATUSES)[number]> {
  return (MISSION_INTERVIEW_DRAFT_STATUSES as readonly string[]).includes(status);
}

/** Mission interview system prompt */
export const MISSION_INTERVIEW_SYSTEM_PROMPT = `You are a mission planning assistant for a project management system.

Your job: help users transform high-level goals into structured mission plans with milestones, slices, and features — each with verification criteria.

## Mission Hierarchy
- Mission: The top-level objective (the user will provide this)
- Milestone: A major phase or deliverable within the mission (e.g., "Foundation & Infrastructure", "Core Feature Development", "Polish & Release"). Each milestone has verification criteria that define how to confirm the phase is complete.
- Slice: A focused work unit within a milestone that can be activated and worked on independently (e.g., "Auth system setup", "API endpoints", "UI components"). Each slice has verification criteria.
- Feature: A specific deliverable within a slice, detailed enough to become a task (e.g., "JWT token refresh endpoint", "Password reset email template"). Each feature has acceptance criteria.

## Conversation Flow
1. The user describes their mission goal
2. Ask clarifying questions to understand scope, constraints, technical context, user needs, and priorities
3. Push back on vague objectives — ask for specifics
4. Challenge unrealistic scope — suggest phasing
5. Once you have enough information (typically 4-8 questions), produce the structured plan
6. The plan should be thorough — break every milestone into slices, every slice into features

## Question Types to Use
PREFER structured question types over free-text. This makes the interview faster and more focused.

- "single_select" (DEFAULT): Use for most questions. Provide 3–6 options covering the common choices.
  ALWAYS include an "Other (please describe)" or "Custom" option as the LAST option so users can provide free-form input.
  Examples: tech stack, deployment target, priority level, integration approach, architecture style.
- "multi_select": Use when multiple options can apply simultaneously (e.g., features to include, platforms to support, security requirements).
  Provide 4–6 options. Include an "Other (please describe)" option at the end.
- "confirm": Use for simple yes/no or go/no-go decisions (e.g., "Should we include offline support?", "Is backwards compatibility required?").
- "text": Use ONLY when genuinely needed — asking for a project name, URL, specific API endpoint, or other unique free-form values that cannot be reasonably optioned.

## Question Design Guidelines
- Provide specific, well-crafted option labels and descriptions so users can quickly select without thinking
- Options should be mutually exclusive and collectively exhaustive for single_select
- Use domain-appropriate jargon in option labels (developers understand "GraphQL", "REST", "gRPC")
- Include 3–6 options per question — never fewer than 3, rarely more than 6
- The last option should always be something like { "id": "other", "label": "Other (please describe)" } for single_select, or similar for multi_select
- Start with big-picture scope questions, then narrow into specifics
- Ask about target users, key constraints, technical preferences, timeline
- Each milestone should represent a meaningful phase boundary or checkpoint
- Each slice should be independently shippable work
- Features should be specific and actionable
- ALWAYS include verification/acceptance criteria at every level:
  - Milestone: "verification" field — how to confirm this phase is complete (e.g., "All API endpoints return correct responses, integration tests pass")
  - Milestone: optional "acceptanceCriteria" field for explicit milestone-level completion bars (if omitted, acceptance can be auto-derived from child feature criteria/descriptions)
  - Slice: "verification" field — how to confirm this work unit is done (e.g., "Auth flow works end-to-end from signup through login")
  - Feature: "acceptanceCriteria" field — how to verify this specific deliverable (e.g., "JWT tokens expire after 1 hour and refresh correctly")
- Suggest sensible defaults and push for specificity
- Aim for 2-4 milestones, 1-3 slices per milestone, 2-5 features per slice
- Keep the plan realistic and achievable

## Board tools
- fn_task_list — list active tasks
- fn_task_get — read a task's full details and PROMPT.md
Use these to avoid duplicating an existing in-flight plan and to anchor your questions against current backlog context.

## Response Format
Always respond with valid JSON in one of these formats:

For single_select question (DEFAULT — use this for most questions):
{"type": "question", "data": {"id": "q-tech-stack", "type": "single_select", "question": "What is the primary technology stack?", "description": "Select the main framework or stack for the project", "options": [{"id": "react-ts", "label": "React + TypeScript", "description": "Component-based UI with type safety"}, {"id": "nextjs", "label": "Next.js", "description": "Full-stack React framework with SSR"}, {"id": "vue", "label": "Vue.js", "description": "Progressive JavaScript framework"}, {"id": "backend-only", "label": "Backend / API only", "description": "No frontend, pure server-side project"}, {"id": "other", "label": "Other (please describe)", "description": "Specify your custom stack"}]}}

For multi_select question:
{"type": "question", "data": {"id": "q-platforms", "type": "multi_select", "question": "Which platforms should be supported?", "description": "Select all that apply", "options": [{"id": "web", "label": "Web browser", "description": "Desktop and mobile web"}, {"id": "ios", "label": "iOS", "description": "iPhone and iPad"}, {"id": "android", "label": "Android", "description": "Phones and tablets"}, {"id": "desktop", "label": "Desktop app", "description": "Electron or native desktop"}, {"id": "api", "label": "API / headless", "description": "Programmatic access only"}, {"id": "other", "label": "Other (please describe)", "description": "Specify additional platforms"}]}}

For confirm question:
{"type": "question", "data": {"id": "q-auth", "type": "confirm", "question": "Should the system require user authentication?", "description": "Login, sessions, and access control"}}

For text question (use ONLY for names, URLs, or truly unique free-form input):
{"type": "question", "data": {"id": "q-project-name", "type": "text", "question": "What is the project or product name?", "description": "Used in documentation and configuration"}}

For completion (when you have enough information):
{"type": "complete", "data": {"missionTitle": "Refined mission title", "missionDescription": "Comprehensive mission description based on the conversation", "milestones": [{"title": "Milestone title", "description": "What this phase achieves", "verification": "How to confirm this milestone is complete", "acceptanceCriteria": "Optional explicit milestone criteria (omit to auto-derive from features)", "slices": [{"title": "Slice title", "description": "What this work unit covers", "verification": "How to confirm this slice is done", "features": [{"title": "Feature title", "description": "What to build", "acceptanceCriteria": "How to verify this feature works"}]}]}]}}`;

// ── Types ───────────────────────────────────────────────────────────────────

/** A feature within a slice in the generated plan */
export interface MissionPlanFeature {
  title: string;
  description?: string;
  acceptanceCriteria?: string;
}

/** A slice within a milestone in the generated plan */
export interface MissionPlanSlice {
  title: string;
  description?: string;
  verification?: string;
  features: MissionPlanFeature[];
}

/** A milestone in the generated plan */
export interface MissionPlanMilestone {
  title: string;
  description?: string;
  verification?: string;
  acceptanceCriteria?: string;
  slices: MissionPlanSlice[];
}

/** The complete mission plan summary produced by the interview */
export interface MissionPlanSummary {
  missionTitle?: string;
  missionDescription?: string;
  milestones: MissionPlanMilestone[];
}

/** Response from interview: either a question or a completed plan */
export type MissionInterviewResponse =
  | { type: "question"; data: PlanningQuestion }
  | { type: "complete"; data: MissionPlanSummary };

export interface MissionInterviewDraftSummary {
  id: string;
  title: string;
  status: Extract<AiSessionStatus, (typeof MISSION_INTERVIEW_DRAFT_STATUSES)[number]>;
  projectId: string | null;
  createdAt: string;
  updatedAt: string;
  hasConversation: boolean;
}

/** SSE event types for mission interview streaming */
export type MissionInterviewStreamEvent =
  | { type: "thinking"; data: string }
  | { type: "text"; data: string }
  | { type: "question"; data: PlanningQuestion }
  | { type: "summary"; data: MissionPlanSummary }
  | { type: "error"; data: string }
  | { type: "complete" };

/** Callback function for streaming events */
export type MissionInterviewStreamCallback = (event: MissionInterviewStreamEvent, eventId?: number) => void;

interface MissionInterviewHistoryEntry {
  question: PlanningQuestion;
  response: unknown;
  thinkingOutput?: string;
}

/** In-memory interview session */
interface MissionInterviewSession {
  id: string;
  ip: string;
  projectId: string | null;
  missionId: string;
  missionTitle: string;
  history: MissionInterviewHistoryEntry[];
  currentQuestion?: PlanningQuestion;
  summary?: MissionPlanSummary;
  /** Last terminal error for retry UX */
  error?: string;
  agent?: AgentResult;
  thinkingOutput: string;
  /** Thinking output generated while producing currentQuestion */
  lastGeneratedThinking: string;
  /** Model override for this interview session */
  modelProvider?: string;
  modelId?: string;
  /**
   * Captured at session creation so rehydrated sessions can rebuild their
   * agent without the caller threading context through every API.
   */
  store?: TaskStore;
  rootDir?: string;
  createdAt: Date;
  updatedAt: Date;
}

interface RateLimitEntry {
  count: number;
  firstRequestAt: Date;
}

// ── In-Memory Storage ───────────────────────────────────────────────────────

const sessions = new Map<string, MissionInterviewSession>();
const rateLimits = new Map<string, RateLimitEntry>();

// ── AI Session Persistence ────────────────────────────────────────────────

let _aiSessionStore: AiSessionStore | undefined;
let _aiSessionDeletedListener: ((sessionId: string) => void) | undefined;

function safeParseJson<T>(
  text: string | null,
  fallback: T,
  options?: { throwOnError?: boolean; fieldName?: string },
): T {
  if (!text) {
    return fallback;
  }

  try {
    return JSON.parse(text) as T;
  } catch (error) {
    if (options?.throwOnError) {
      const fieldSuffix = options.fieldName ? ` in ${options.fieldName}` : "";
      throw new Error(`Invalid JSON${fieldSuffix}: ${(error as Error).message}`);
    }
    return fallback;
  }
}

export function setAiSessionStore(store: AiSessionStore): void {
  if (_aiSessionStore && _aiSessionDeletedListener) {
    _aiSessionStore.off("ai_session:deleted", _aiSessionDeletedListener);
  }

  _aiSessionStore = store;
  _aiSessionDeletedListener = (sessionId: string) => {
    cleanupInMemoryMissionSession(sessionId);
  };
  _aiSessionStore.on("ai_session:deleted", _aiSessionDeletedListener);
}

function cleanupInMemoryMissionSession(sessionId: string): boolean {
  const session = sessions.get(sessionId);
  if (!session) {
    return false;
  }

  // Abort any in-flight generation so prompt() rejects promptly.
  generationGuard.stop(sessionId);

  if (session.agent) {
    try { session.agent.session.dispose?.(); } catch { /* ignore */ }
    session.agent = undefined;
  }

  missionInterviewStreamManager.cleanupSession(sessionId);
  sessions.delete(sessionId);
  return true;
}

function persistMissionSession(session: MissionInterviewSession, status: "generating" | "awaiting_input" | "complete" | "error", error?: string): void {
  if (!_aiSessionStore) return;
  const row: AiSessionRow = {
    id: session.id,
    type: "mission_interview",
    status,
    title: session.missionTitle.slice(0, 120),
    inputPayload: JSON.stringify({
      ip: session.ip,
      projectId: session.projectId,
      missionTitle: session.missionTitle,
      missionId: session.missionId,
      modelProvider: session.modelProvider,
      modelId: session.modelId,
    }),
    conversationHistory: JSON.stringify(session.history),
    currentQuestion: session.currentQuestion ? JSON.stringify(session.currentQuestion) : null,
    result: session.summary ? JSON.stringify(session.summary) : null,
    thinkingOutput: session.thinkingOutput,
    error: error ?? null,
    projectId: session.projectId,
    createdAt: session.createdAt.toISOString(),
    updatedAt: new Date().toISOString(),
    lockedByTab: null,
    lockedAt: null,
  };
  _aiSessionStore.upsert(row);
}

function persistMissionThinking(sessionId: string, thinkingOutput: string): void {
  if (!_aiSessionStore) return;
  _aiSessionStore.updateThinking(sessionId, thinkingOutput);
}

function unpersistMissionSession(sessionId: string): void {
  if (!_aiSessionStore) return;
  _aiSessionStore.delete(sessionId);
}

function buildMissionInterviewSessionFromRow(row: AiSessionRow): MissionInterviewSession {
  const payload = safeParseJson<{ ip?: string; projectId?: string | null; missionId?: string; missionTitle?: string; modelProvider?: string; modelId?: string }>(
    row.inputPayload,
    {},
    { throwOnError: true, fieldName: "inputPayload" },
  );

  const createdAt = new Date(row.createdAt);
  const updatedAt = new Date(row.updatedAt);

  if (Number.isNaN(createdAt.getTime()) || Number.isNaN(updatedAt.getTime())) {
    throw new Error("Invalid session timestamps");
  }

  return {
    id: row.id,
    ip: payload.ip ?? "",
    projectId: row.projectId ?? payload.projectId ?? null,
    missionId: payload.missionId ?? "",
    missionTitle: payload.missionTitle ?? row.title,
    history: safeParseJson<MissionInterviewHistoryEntry[]>(
      row.conversationHistory,
      [],
      { throwOnError: true, fieldName: "conversationHistory" },
    ),
    currentQuestion: row.currentQuestion
      ? (safeParseJson<PlanningQuestion | null>(row.currentQuestion, null, {
          throwOnError: true,
          fieldName: "currentQuestion",
        }) ?? undefined)
      : undefined,
    summary: row.result
      ? (safeParseJson<MissionPlanSummary | null>(row.result, null, {
          throwOnError: true,
          fieldName: "result",
        }) ?? undefined)
      : undefined,
    thinkingOutput: row.thinkingOutput,
    lastGeneratedThinking: row.thinkingOutput || "",
    error: row.error ?? undefined,
    modelProvider: payload.modelProvider,
    modelId: payload.modelId,
    createdAt,
    updatedAt,
    agent: undefined,
  };
}

export function rehydrateFromStore(store: AiSessionStore): number {
  let rows: AiSessionRow[] = [];

  try {
    rows = store.listRecoverable().filter((row) => row.type === "mission_interview");
  } catch (error) {
    diagnostics.errorFromException("Failed to list recoverable sessions", error, { operation: "list-recoverable" });
    return 0;
  }

  let rehydrated = 0;
  for (const row of rows) {
    try {
      const session = buildMissionInterviewSessionFromRow(row);
      sessions.set(session.id, session);
      rehydrated += 1;
    } catch (error) {
      diagnostics.errorFromException("Failed to rehydrate session", error, { sessionId: row.id, operation: "rehydrate" });
    }
  }

  return rehydrated;
}

// ── Cleanup Interval ────────────────────────────────────────────────────────

function cleanupExpiredSessions(): void {
  const now = Date.now();
  for (const [id, session] of sessions) {
    if (now - session.updatedAt.getTime() > SESSION_TTL_MS) {
      cleanupInMemoryMissionSession(id);
    }
  }
  for (const [ip, entry] of rateLimits) {
    if (now - entry.firstRequestAt.getTime() > RATE_LIMIT_WINDOW_MS) {
      rateLimits.delete(ip);
    }
  }
}

const cleanupInterval = setInterval(cleanupExpiredSessions, CLEANUP_INTERVAL_MS);
cleanupInterval.unref?.();
process.on("beforeExit", () => clearInterval(cleanupInterval));

// ── Stream Manager ──────────────────────────────────────────────────────────

export class MissionInterviewStreamManager extends EventEmitter {
  private readonly sessions = new Map<string, Set<MissionInterviewStreamCallback>>();
  private readonly buffers = new Map<string, SessionEventBuffer>();

  constructor(private readonly bufferSize = 100) {
    super();
  }

  subscribe(sessionId: string, callback: MissionInterviewStreamCallback): () => void {
    if (!this.sessions.has(sessionId)) {
      this.sessions.set(sessionId, new Set());
    }
    const callbacks = this.sessions.get(sessionId)!;
    callbacks.add(callback);
    return () => {
      callbacks.delete(callback);
      if (callbacks.size === 0) {
        this.sessions.delete(sessionId);
      }
    };
  }

  private getBuffer(sessionId: string): SessionEventBuffer {
    let buffer = this.buffers.get(sessionId);
    if (!buffer) {
      buffer = new SessionEventBuffer(this.bufferSize);
      this.buffers.set(sessionId, buffer);
    }
    return buffer;
  }

  broadcast(sessionId: string, event: MissionInterviewStreamEvent): number {
    const serialized = JSON.stringify((event as { data?: unknown }).data ?? {});
    const eventData = typeof serialized === "string" ? serialized : "{}";
    const eventId = this.getBuffer(sessionId).push(event.type, eventData);

    const callbacks = this.sessions.get(sessionId);
    if (!callbacks) return eventId;

    for (const callback of callbacks) {
      nonfatal(
        () => callback(event, eventId),
        diagnostics,
        "Error broadcasting to client",
        { sessionId, operation: "broadcast" }
      );
    }

    return eventId;
  }

  getBufferedEvents(sessionId: string, sinceId: number): SessionBufferedEvent[] {
    const buffer = this.buffers.get(sessionId);
    if (!buffer) return [];
    return buffer.getEventsSince(sinceId);
  }

  hasSubscribers(sessionId: string): boolean {
    const callbacks = this.sessions.get(sessionId);
    return callbacks !== undefined && callbacks.size > 0;
  }

  cleanupSession(sessionId: string): void {
    this.sessions.delete(sessionId);
    this.buffers.delete(sessionId);
  }

  reset(): void {
    this.sessions.clear();
    this.buffers.clear();
    this.removeAllListeners();
  }
}

export const missionInterviewStreamManager = new MissionInterviewStreamManager();

// ── Rate Limiting ───────────────────────────────────────────────────────────

export function checkRateLimit(ip: string): boolean {
  const now = Date.now();
  const entry = rateLimits.get(ip);

  if (!entry) {
    rateLimits.set(ip, { count: 1, firstRequestAt: new Date() });
    return true;
  }

  if (now - entry.firstRequestAt.getTime() > RATE_LIMIT_WINDOW_MS) {
    rateLimits.set(ip, { count: 1, firstRequestAt: new Date() });
    return true;
  }

  if (entry.count >= MAX_SESSIONS_PER_IP_PER_HOUR) {
    return false;
  }

  entry.count++;
  return true;
}

export function getRateLimitResetTime(ip: string): Date | null {
  const entry = rateLimits.get(ip);
  if (!entry) return null;
  return new Date(entry.firstRequestAt.getTime() + RATE_LIMIT_WINDOW_MS);
}

// ── JSON Parsing Utilities ─────────────────────────────────────────────────

/**
 * Extract the best JSON candidate from AI response text.
 * Handles markdown-wrapped JSON, embedded prose, and multiple objects.
 */
export function extractJsonCandidate(text: string): string | null {
  if (!text || !text.trim()) return null;

  // 1. Try markdown code blocks first
  const codeBlockMatch = text.match(/```(?:json)?\s*([\s\S]*?)\s*```/);
  if (codeBlockMatch?.[1]) {
    const candidate = codeBlockMatch[1].trim();
    if (candidate.startsWith("{")) return candidate;
  }

  // 2. Find all top-level brace-delimited objects using balanced brace counting
  const candidates: Array<{ text: string }> = [];
  for (let i = 0; i < text.length; i++) {
    if (text[i] === "{") {
      let depth = 0;
      let inString = false;
      let escape = false;
      for (let j = i; j < text.length; j++) {
        const ch = text[j];
        if (escape) { escape = false; continue; }
        if (ch === "\\") { escape = true; continue; }
        if (ch === '"') { inString = !inString; continue; }
        if (inString) continue;
        if (ch === "{") depth++;
        if (ch === "}") depth--;
        if (depth === 0) {
          const candidate = text.slice(i, j + 1).trim();
          try {
            JSON.parse(candidate);
            candidates.push({ text: candidate });
          } catch { /* not valid JSON */ }
          break;
        }
      }
    }
  }

  if (candidates.length > 0) {
    candidates.sort((a, b) => b.text.length - a.text.length);
    return candidates[0].text;
  }

  // 3. Last resort: try the full trimmed text
  const trimmed = text.trim();
  if (trimmed.startsWith("{")) return trimmed;

  return null;
}

/**
 * Attempt to repair common JSON issues.
 */
export function repairJson(text: string): string {
  let repaired = text;
  repaired = repaired.replace(/,\s*([}\]])/g, "$1");

  let openBraces = 0;
  let openBrackets = 0;
  let inString = false;
  let escape = false;
  for (const ch of repaired) {
    if (escape) { escape = false; continue; }
    if (ch === "\\") { escape = true; continue; }
    if (ch === '"') { inString = !inString; continue; }
    if (inString) continue;
    if (ch === "{") openBraces++;
    if (ch === "}") openBraces--;
    if (ch === "[") openBrackets++;
    if (ch === "]") openBrackets--;
  }

  if (inString) repaired += '"';

  // Re-count after potential string fix
  openBraces = 0;
  openBrackets = 0;
  inString = false;
  escape = false;
  for (const ch of repaired) {
    if (escape) { escape = false; continue; }
    if (ch === "\\") { escape = true; continue; }
    if (ch === '"') { inString = !inString; continue; }
    if (inString) continue;
    if (ch === "{") openBraces++;
    if (ch === "}") openBraces--;
    if (ch === "[") openBrackets++;
    if (ch === "]") openBrackets--;
  }

  repaired += "]".repeat(Math.max(0, openBrackets));
  repaired += "}".repeat(Math.max(0, openBraces));

  return repaired;
}

/**
 * Parse AI agent response into a MissionInterviewResponse.
 * Handles markdown wrapping, embedded prose, truncated JSON.
 */
export function parseMissionAgentResponse(text: string): MissionInterviewResponse {
  const candidate = extractJsonCandidate(text);

  if (!candidate) {
    diagnostics.error("No JSON candidate found in agent response", { inputSnippet: text.slice(0, 500), operation: "parse-json" });
    throw new Error("AI returned no valid JSON. Please try again.");
  }

  let parsed: unknown;
  try {
    parsed = JSON.parse(candidate);
  } catch {
    try {
      const repaired = repairJson(candidate);
      parsed = JSON.parse(repaired);
    } catch (repairErr) {
      diagnostics.error("Failed to parse agent response (repair also failed)", { inputSnippet: candidate.slice(0, 500), operation: "parse-json-repair" });
      throw new Error(
        `Failed to parse AI response: ${repairErr instanceof Error ? repairErr.message : "Unknown error"}. Please try again.`
      );
    }
  }

  if (
    typeof parsed === "object" &&
    parsed !== null &&
    "type" in parsed &&
    "data" in parsed
  ) {
    const typed = parsed as { type: string; data: unknown };
    if (typed.type === "question" && typed.data !== null && typed.data !== undefined) {
      return parsed as MissionInterviewResponse;
    }
    if (typed.type === "complete" && typed.data !== null && typeof typed.data === "object") {
      const data = typed.data as Record<string, unknown>;
      if (Array.isArray(data.milestones)) {
        return parsed as MissionInterviewResponse;
      }
    }
  }

  diagnostics.error("Invalid response structure from AI", { parsedSnippet: JSON.stringify(parsed).slice(0, 500), operation: "parse-validate" });
  throw new Error("AI returned an invalid response structure. Please try again.");
}

// ── Response Formatting ────────────────────────────────────────────────────

/**
 * Format user response as a message for the AI agent.
 */
function formatResponseForAgent(
  question: PlanningQuestion,
  responses: Record<string, unknown>
): string {
  const responseValue = responses[question.id];
  const comment = typeof responses._comment === "string" ? responses._comment.trim() : "";

  let formatted: string;

  switch (question.type) {
    case "text":
      formatted = `Question: ${question.question}\n\nAnswer: ${responseValue}`;
      break;
    case "single_select":
      if (typeof responseValue === "string") {
        const option = question.options?.find((o) => o.id === responseValue);
        formatted = `Question: ${question.question}\n\nSelected: ${option?.label || responseValue}`;
        break;
      }
      formatted = `Question: ${question.question}\n\nAnswer: ${responseValue}`;
      break;
    case "multi_select":
      if (Array.isArray(responseValue)) {
        const selected = responseValue.map((id) => {
          const option = question.options?.find((o) => o.id === id);
          return option?.label || id;
        });
        formatted = `Question: ${question.question}\n\nSelected: ${selected.join(", ")}`;
        break;
      }
      formatted = `Question: ${question.question}\n\nAnswer: ${responseValue}`;
      break;
    case "confirm":
      formatted = `Question: ${question.question}\n\nAnswer: ${responseValue === true ? "Yes" : "No"}`;
      break;
    default:
      formatted = `Question: ${question.question}\n\nAnswer: ${JSON.stringify(responseValue)}`;
      break;
  }

  return comment.length > 0 ? `${formatted}\n\nAdditional context: ${comment}` : formatted;
}

function coerceResponseRecord(question: PlanningQuestion, response: unknown): Record<string, unknown> {
  if (response && typeof response === "object" && !Array.isArray(response)) {
    return response as Record<string, unknown>;
  }

  return {
    [question.id]: response,
  };
}

function disposeMissionAgentForRetry(session: MissionInterviewSession): void {
  if (!session.agent) {
    return;
  }

  nonfatal(
    () => session.agent.session.dispose?.(),
    diagnostics,
    "Error disposing agent for retry",
    { sessionId: session.id, operation: "dispose-retry" }
  );

  session.agent = undefined;
}

// ── AI Agent Integration ───────────────────────────────────────────────────

/**
 * Initialize the AI agent for a session and start the first turn.
 */
async function initializeAgent(
  session: MissionInterviewSession,
  rootDir: string,
  store: TaskStore,
  promptOverrides?: PromptOverrideMap,
): Promise<void> {
  try {
    session.agent = await createMissionInterviewAgent(session, rootDir, store, promptOverrides);
    session.updatedAt = new Date();

    // Send initial message to get first question
    await continueAgentConversation(
      session,
      `I want to plan a mission: "${session.missionTitle}". Interview me to understand what I need, then produce a structured plan.`,
    );
  } catch (err) {
    const errorMessage = err instanceof Error ? err.message : "Failed to initialize AI agent";
    diagnostics.errorFromException("Agent initialization error for session", err, { sessionId: session.id, operation: "initialize-agent" });
    session.error = errorMessage;
    session.updatedAt = new Date();
    persistMissionSession(session, "error", errorMessage);
    missionInterviewStreamManager.broadcast(session.id, {
      type: "error",
      data: errorMessage,
    });
  }
}

async function createMissionInterviewAgent(
  session: MissionInterviewSession,
  rootDir: string,
  store: TaskStore,
  promptOverrides?: PromptOverrideMap,
): Promise<AgentResult> {
  await ensureEngineReady();

  const effectivePrompt = resolvePrompt("mission-interview-system", promptOverrides);

  return createFnAgent({
    cwd: rootDir,
    systemPrompt: effectivePrompt,
    tools: "readonly",
    builtinToolsAllowlist: [...MISSION_INTERVIEW_BUILTIN_WEB_TOOLS],
    customTools: [...createPlanningBoardTools(store)],
    ...(session.modelProvider && session.modelId
      ? {
          defaultProvider: session.modelProvider,
          defaultModelId: session.modelId,
        }
      : {}),
    onThinking: (delta: string) => {
      session.thinkingOutput += delta;
      persistMissionThinking(session.id, session.thinkingOutput);
      missionInterviewStreamManager.broadcast(session.id, {
        type: "thinking",
        data: delta,
      });
    },
    onText: (delta: string) => {
      session.thinkingOutput += delta;
      missionInterviewStreamManager.broadcast(session.id, {
        type: "text",
        data: delta,
      });
    },
  });
}

function formatMissionInterviewHistory(
  history: Array<{ question: PlanningQuestion; response: unknown }>,
): string {
  if (history.length === 0) {
    return "";
  }

  return history
    .map(({ question, response }) => {
      const responseRecord =
        response && typeof response === "object" && !Array.isArray(response)
          ? (response as Record<string, unknown>)
          : undefined;
      const responseValue = responseRecord ? responseRecord[question.id] : response;
      const comment = typeof responseRecord?._comment === "string" ? responseRecord._comment.trim() : "";

      const lines = [
        `Q: ${question.question}`,
        `A: ${typeof responseValue === "string" ? responseValue : JSON.stringify(responseValue ?? null)}`,
      ];

      if (comment.length > 0) {
        lines.push(`Comment: ${comment}`);
      }

      return lines.join("\n");
    })
    .join("\n\n");
}

async function ensureMissionInterviewAgent(
  session: MissionInterviewSession,
  rootDir: string | undefined,
  store: TaskStore | undefined,
  historyForReplay: Array<{ question: PlanningQuestion; response: unknown }>,
  promptOverrides?: PromptOverrideMap,
): Promise<void> {
  if (session.agent) {
    return;
  }

  const effectiveRootDir = rootDir ?? session.rootDir;
  const effectiveStore = store ?? session.store;

  if (!effectiveRootDir) {
    throw new InvalidSessionStateError(
      "AI agent not available for this session and cannot be resumed without project context",
    );
  }
  if (!effectiveStore) {
    throw new InvalidSessionStateError(
      "AI agent not available for this session and cannot be resumed without task store context",
    );
  }

  session.agent = await createMissionInterviewAgent(session, effectiveRootDir, effectiveStore, promptOverrides);

  if (historyForReplay.length === 0) {
    return;
  }

  const historySummary = formatMissionInterviewHistory(historyForReplay);
  if (!historySummary) {
    return;
  }

  await generationGuard.run(
    session.id,
    GENERATION_TIMEOUT_MS,
    {
      onTimeout: () => setMissionSessionError(
        session,
        "AI generation timed out while restoring context. You can retry or start a new session.",
      ),
      onUserStop: () => setMissionSessionError(
        session,
        "Generation stopped by user. You can retry or start a new session.",
      ),
    },
    () => session.agent!.session.prompt(
      [
        "Previous conversation summary:",
        historySummary,
        "Use this context when handling the next user response.",
      ].join("\n\n"),
    ),
  );
}

function setMissionSessionError(session: MissionInterviewSession, message: string): void {
  session.error = message;
  session.updatedAt = new Date();
  persistMissionSession(session, "error", message);
  missionInterviewStreamManager.broadcast(session.id, {
    type: "error",
    data: message,
  });
}

/**
 * Manually abort an in-flight mission interview generation. Returns true if
 * a generation was active and got aborted.
 */
export function stopMissionInterviewGeneration(sessionId: string): boolean {
  return generationGuard.stop(sessionId);
}

/**
 * Continue the AI conversation with a user message.
 * Includes bounded recovery: one retry on parse failure.
 *
 * The entire body — including the parse-retry inner `prompt()` calls — runs
 * inside `generationGuard.run`, so a stalled stream or hung tool call on
 * either the first or retry attempt is bounded by `GENERATION_TIMEOUT_MS`.
 */
async function continueAgentConversation(session: MissionInterviewSession, message: string): Promise<void> {
  if (!session.agent) {
    throw new InvalidSessionStateError("AI agent not initialized");
  }

  try {
    await generationGuard.run(
      session.id,
      GENERATION_TIMEOUT_MS,
      {
        onTimeout: () => setMissionSessionError(
          session,
          "AI generation timed out. You can retry or start a new session.",
        ),
        onUserStop: () => setMissionSessionError(
          session,
          "Generation stopped by user. You can retry or start a new session.",
        ),
      },
      async () => {
        const agent = session.agent!;
        session.thinkingOutput = "";

        await agent.session.prompt(message);

        // Get the response text from the agent's state
        interface AgentMessage {
          role: string;
          content?: string | Array<{ type: string; text: string }>;
        }
        const lastMessage = (agent.session.state.messages as AgentMessage[])
          .filter((m: AgentMessage) => m.role === "assistant")
          .pop();

        let responseText = session.thinkingOutput;
        if (lastMessage?.content) {
          if (typeof lastMessage.content === "string") {
            responseText = lastMessage.content;
          } else if (Array.isArray(lastMessage.content)) {
            responseText = lastMessage.content
              .filter((c: { type: string; text: string }): c is { type: "text"; text: string } => c.type === "text")
              .map((c: { type: string; text: string }) => c.text)
              .join("");
          }
        }

        // Parse with retry
        let parsed: MissionInterviewResponse | undefined;
        let lastError: Error | undefined;

        for (let attempt = 0; attempt <= MAX_PARSE_RETRIES; attempt++) {
          try {
            parsed = parseMissionAgentResponse(responseText);
            break;
          } catch (err) {
            lastError = err instanceof Error ? err : new Error(String(err));

            if (attempt < MAX_PARSE_RETRIES) {
              diagnostics.warn(
                "Parse attempt failed, requesting reformat",
                { sessionId: session.id, attempt: attempt + 1, operation: "parse-retry" }
              );
              try {
                session.thinkingOutput = "";
                await agent.session.prompt(
                  "Your previous response could not be parsed as JSON. " +
                  'Please respond with ONLY a valid JSON object: either {"type":"question","data":{...}} ' +
                  'or {"type":"complete","data":{"missionTitle":"...","missionDescription":"...","milestones":[...]}}. ' +
                  "No markdown, no explanation, just the JSON."
                );

                const retryMessage = (agent.session.state.messages as AgentMessage[])
                  .filter((m: AgentMessage) => m.role === "assistant")
                  .pop();

                let retryText = session.thinkingOutput;
                if (retryMessage?.content) {
                  if (typeof retryMessage.content === "string") {
                    retryText = retryMessage.content;
                  } else if (Array.isArray(retryMessage.content)) {
                    retryText = retryMessage.content
                      .filter((c: { type: string; text: string }): c is { type: "text"; text: string } => c.type === "text")
                      .map((c: { type: string; text: string }) => c.text)
                      .join("");
                  }
                }
                responseText = retryText;
              } catch (retryErr) {
                diagnostics.errorFromException("Retry prompt failed for session", retryErr, { sessionId: session.id, operation: "retry-prompt" });
                break;
              }
            }
          }
        }

        if (!parsed) {
          const errorMsg = `${lastError?.message || "Failed to parse AI response"} You can try responding again or start a new session.`;
          diagnostics.error(
            "All parse attempts exhausted for session",
            { sessionId: session.id, message: errorMsg, operation: "parse-exhausted" }
          );
          setMissionSessionError(session, errorMsg);
          return;
        }

        if (parsed.type === "question") {
          session.currentQuestion = parsed.data;
          session.error = undefined;
          session.lastGeneratedThinking = session.thinkingOutput;
          session.updatedAt = new Date();
          persistMissionSession(session, "awaiting_input");
          missionInterviewStreamManager.broadcast(session.id, {
            type: "question",
            data: parsed.data,
          });
        } else if (parsed.type === "complete") {
          session.summary = parsed.data;
          session.currentQuestion = undefined;
          session.error = undefined;
          session.updatedAt = new Date();
          persistMissionSession(session, "complete");
          missionInterviewStreamManager.broadcast(session.id, {
            type: "summary",
            data: parsed.data,
          });
          missionInterviewStreamManager.broadcast(session.id, { type: "complete" });
        }
      },
    );
  } catch (err) {
    // Timeout / user-stop already published an error state via the guard
    // handlers. Don't double-broadcast a generic AbortError.
    if (isAbortError(err)) {
      return;
    }
    const errorMessage = err instanceof Error ? err.message : "AI processing failed";
    diagnostics.errorFromException("Agent conversation error for session", err, { sessionId: session.id, operation: "conversation" });
    setMissionSessionError(session, errorMessage);
  }
}

// ── Session Management ──────────────────────────────────────────────────────

/**
 * Create a new mission interview session with AI agent streaming.
 * Returns sessionId immediately; client connects to SSE to receive events.
 */
export async function createMissionInterviewSession(
  ip: string,
  missionTitle: string,
  rootDir: string,
  store: TaskStore,
  promptOverrides?: PromptOverrideMap,
  modelProvider?: string,
  modelId?: string,
  projectId?: string | null,
): Promise<string> {
  if (!checkRateLimit(ip)) {
    const resetTime = getRateLimitResetTime(ip);
    throw new RateLimitError(
      `Rate limit exceeded. Maximum ${MAX_SESSIONS_PER_IP_PER_HOUR} sessions per hour. ` +
        `Reset at ${resetTime?.toISOString() || "unknown"}`
    );
  }

  const sessionId = randomUUID();

  const session: MissionInterviewSession = {
    id: sessionId,
    ip,
    projectId: projectId ?? null,
    missionId: "",
    missionTitle,
    history: [],
    thinkingOutput: "",
    lastGeneratedThinking: "",
    modelProvider,
    modelId,
    store,
    rootDir,
    createdAt: new Date(),
    updatedAt: new Date(),
  };

  sessions.set(sessionId, session);
  persistMissionSession(session, "generating");

  // Initialize AI agent in background
  initializeAgent(session, rootDir, store, promptOverrides).catch((err) => {
    diagnostics.errorFromException("Failed to initialize agent for session", err, { sessionId, operation: "initialize-agent" });
    persistMissionSession(session, "error", err.message || "Failed to initialize AI agent");
    missionInterviewStreamManager.broadcast(sessionId, {
      type: "error",
      data: err.message || "Failed to initialize AI agent",
    });
  });

  return sessionId;
}

/**
 * Submit a response to the current question.
 * Supports AI agent mode with streaming.
 */
export async function submitMissionInterviewResponse(
  sessionId: string,
  responses: Record<string, unknown>,
  rootDir?: string,
  store?: TaskStore,
  promptOverrides?: PromptOverrideMap,
): Promise<MissionInterviewResponse> {
  const session = getMissionInterviewSession(sessionId);
  if (!session) {
    throw new SessionNotFoundError(`Mission interview session ${sessionId} not found or expired`);
  }

  if (store && !session.store) session.store = store;
  if (rootDir && !session.rootDir) session.rootDir = rootDir;

  if (!session.currentQuestion) {
    throw new InvalidSessionStateError("No active question in session");
  }

  // Record the response
  session.history.push({
    question: session.currentQuestion,
    response: responses,
    thinkingOutput: session.lastGeneratedThinking || "",
  });
  session.error = undefined;
  persistMissionSession(session, "generating");

  if (!session.agent) {
    const replayHistory = session.history.slice(0, -1);
    await ensureMissionInterviewAgent(session, rootDir, store, replayHistory, promptOverrides);
  }

  const message = formatResponseForAgent(session.currentQuestion, responses);
  await continueAgentConversation(session, message);

  if (session.summary) {
    return { type: "complete", data: session.summary };
  }
  if (session.currentQuestion) {
    return { type: "question", data: session.currentQuestion };
  }
  // Fallback — should not happen with a working agent
  return {
    type: "question",
    data: {
      id: "q-fallback",
      type: "text",
      question: "Could you tell me more about what you want to build?",
      description: "The AI is processing your response. Please provide more details.",
    },
  };
}

export async function retryMissionInterviewSession(
  sessionId: string,
  rootDir: string,
  store?: TaskStore,
  promptOverrides?: PromptOverrideMap,
): Promise<void> {
  const session = getMissionInterviewSession(sessionId);
  if (!session) {
    throw new SessionNotFoundError(`Mission interview session ${sessionId} not found or expired`);
  }

  if (store && !session.store) session.store = store;
  if (rootDir && !session.rootDir) session.rootDir = rootDir;

  const persisted = _aiSessionStore?.get(sessionId);
  if (persisted && persisted.type !== "mission_interview") {
    throw new SessionNotFoundError(`Mission interview session ${sessionId} not found or expired`);
  }

  const inErrorState = persisted ? persisted.status === "error" : Boolean(session.error);
  if (!inErrorState) {
    throw new InvalidSessionStateError(`Mission interview session ${sessionId} is not in an error state`);
  }

  disposeMissionAgentForRetry(session);

  session.error = undefined;
  session.summary = undefined;
  session.updatedAt = new Date();
  persistMissionSession(session, "generating");

  if (session.history.length === 0) {
    await ensureMissionInterviewAgent(session, rootDir, store, [], promptOverrides);
    await continueAgentConversation(
      session,
      `I want to plan a mission: "${session.missionTitle}". Interview me to understand what I need, then produce a structured plan.`,
    );
    return;
  }

  const replayHistory = session.history.slice(0, -1);
  const lastEntry = session.history[session.history.length - 1];

  await ensureMissionInterviewAgent(session, rootDir, store, replayHistory, promptOverrides);
  const replayMessage = formatResponseForAgent(
    lastEntry.question,
    coerceResponseRecord(lastEntry.question, lastEntry.response),
  );
  await continueAgentConversation(session, replayMessage);
}

export async function cancelMissionInterviewSession(sessionId: string): Promise<void> {
  const removed = cleanupInMemoryMissionSession(sessionId);
  if (!removed) {
    throw new SessionNotFoundError(`Mission interview session ${sessionId} not found or expired`);
  }

  unpersistMissionSession(sessionId);
}

export function listMissionInterviewDrafts(projectId?: string): MissionInterviewDraftSummary[] {
  if (!_aiSessionStore) {
    return [];
  }

  return _aiSessionStore
    .listActive(projectId)
    .filter((session) => {
      if (session.type !== "mission_interview") {
        return false;
      }
      if (!isMissionInterviewDraftStatus(session.status)) {
        return false;
      }
      if (projectId) {
        return session.projectId === projectId;
      }
      return session.projectId == null;
    })
    .map((session) => {
      const row = _aiSessionStore?.get(session.id);
      const conversation = row ? safeParseJson<unknown[]>(row.conversationHistory, []) : [];
      return {
        id: session.id,
        title: session.title,
        status: session.status as MissionInterviewDraftSummary["status"],
        projectId: session.projectId,
        createdAt: row?.createdAt ?? session.updatedAt,
        updatedAt: session.updatedAt,
        hasConversation: conversation.length > 0,
      };
    });
}

export async function discardMissionInterviewSession(sessionId: string): Promise<{ removed: boolean }> {
  try {
    await cancelMissionInterviewSession(sessionId);
    return { removed: true };
  } catch (error) {
    if (!(error instanceof SessionNotFoundError)) {
      throw error;
    }
  }

  const removed = _aiSessionStore?.deleteByIdAndType(sessionId, "mission_interview") ?? false;
  return { removed };
}

export function getMissionInterviewSession(sessionId: string): MissionInterviewSession | undefined {
  const inMemory = sessions.get(sessionId);
  if (inMemory) {
    return inMemory;
  }

  if (!_aiSessionStore) {
    return undefined;
  }

  const row = _aiSessionStore.get(sessionId);
  if (!row || row.type !== "mission_interview") {
    return undefined;
  }

  try {
    const restored = buildMissionInterviewSessionFromRow(row);
    sessions.set(restored.id, restored);
    return restored;
  } catch (error) {
    diagnostics.errorFromException("Failed to restore session from SQLite", error, { sessionId, operation: "restore" });
    return undefined;
  }
}

export function getMissionInterviewSummary(sessionId: string): MissionPlanSummary | undefined {
  return getMissionInterviewSession(sessionId)?.summary;
}

export function cleanupMissionInterviewSession(sessionId: string): void {
  cleanupInMemoryMissionSession(sessionId);
  unpersistMissionSession(sessionId);
}

/**
 * Reset all mission interview state. Used for testing only.
 */
/**
 * Test-only: register a stub session so SSE/buffer tests can drive the stream
 * manager directly without spinning up the real AI agent (which is
 * deliberately blocked in the vitest harness).
 */
export function __registerMissionInterviewSessionForTest(sessionId: string, missionTitle = "Test Mission"): void {
  sessions.set(sessionId, {
    id: sessionId,
    ip: "127.0.0.1",
    projectId: null,
    missionId: "",
    missionTitle,
    history: [],
    thinkingOutput: "",
    lastGeneratedThinking: "",
    createdAt: new Date(),
    updatedAt: new Date(),
  });
}

export function __resetMissionInterviewState(): void {
  for (const [id] of sessions) {
    cleanupInMemoryMissionSession(id);
  }
  sessions.clear();
  rateLimits.clear();
  missionInterviewStreamManager.reset();
  generationGuard.reset();

  if (_aiSessionStore && _aiSessionDeletedListener) {
    _aiSessionStore.off("ai_session:deleted", _aiSessionDeletedListener);
  }
  _aiSessionDeletedListener = undefined;
  _aiSessionStore = undefined;

  // Reset diagnostics sink to default
  resetDiagnosticsSink();
}

// ── Custom Errors ───────────────────────────────────────────────────────────

export class RateLimitError extends Error {
  constructor(message: string) {
    super(message);
    this.name = "RateLimitError";
  }
}

export class SessionNotFoundError extends Error {
  constructor(message: string) {
    super(message);
    this.name = "SessionNotFoundError";
  }
}

export class InvalidSessionStateError extends Error {
  constructor(message: string) {
    super(message);
    this.name = "InvalidSessionStateError";
  }
}