@gajae-code/coding-agent 0.7.7 → 0.7.9

package/src/coordinator-mcp/server.ts

@@ -119,7 +119,7 @@ interface TurnRecord {
 		target: string | null;
 		tmux_keys_sent?: boolean;
 		prompt_acknowledged?: boolean;
-		state?: "queued" | "tmux_keys_sent" | "acknowledged" | "unavailable";
+		state?: "queued" | "tmux_keys_sent" | "acknowledged" | "unavailable" | "unacknowledged";
 		attempts: Array<{
 			delivered: boolean;
 			created_at: string;
@@ -214,6 +214,9 @@ interface CoordinatorEventInput {
 }
 
 const MISSING_FINAL_RESPONSE_ADVISORY = "completion_missing_final_response";
+const PROMPT_ACK_TIMEOUT_REASON = "runtime_prompt_ack_timeout";
+const DEFAULT_RUNTIME_PROMPT_ACK_TIMEOUT_MS = 10_000;
+const MAX_RUNTIME_PROMPT_ACK_TIMEOUT_MS = 5 * 60 * 1000;
 const ACTIVE_TURN_STATUSES = new Set<TurnStatus>(["delivering", "active", "waiting_for_answer", "completing"]);
 const TERMINAL_TURN_STATUSES = new Set<TurnStatus>(["completed", "failed", "cancelled", "superseded"]);
 const TURN_ID_PATTERN = /^turn-[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/i;
@@ -317,8 +320,14 @@ function toolSchema(name: CoordinatorToolName): {
 				properties: {
 					session_id: sessionId,
 					turn_id: { type: "string" },
-					timeout_ms: { type: "number" },
-					poll_interval_ms: { type: "number" },
+					timeout_ms: {
+						type: "number",
+						description: "Bounded await timeout in milliseconds, capped at 30 minutes.",
+					},
+					poll_interval_ms: {
+						type: "number",
+						description: "Bounded polling interval in milliseconds, capped at 10 seconds.",
+					},
 					lines: { type: "number" },
 				},
 				required: ["turn_id"],
@@ -406,7 +415,10 @@ function toolSchema(name: CoordinatorToolName): {
 					after_seq: { type: "number" },
 					session_id: sessionId,
 					event_types: { type: "array", items: { type: "string" } },
-					timeout_ms: { type: "number" },
+					timeout_ms: {
+						type: "number",
+						description: "Bounded event long-poll timeout in milliseconds, capped at 30 seconds.",
+					},
 					limit: { type: "number" },
 				},
 			},
@@ -447,7 +459,8 @@ function toolSchema(name: CoordinatorToolName): {
 					await_completion: { type: "boolean", description: "If true, poll the turn until terminal or timeout." },
 					timeout_ms: {
 						type: "number",
-						description: "Bounded await timeout; same cap semantics as gjc_coordinator_await_turn.",
+						description:
+							"Bounded await timeout in milliseconds, capped at 30 minutes like gjc_coordinator_await_turn.",
 					},
 					poll_interval_ms: { type: "number", description: "Bounded await polling interval." },
 					lines: { type: "number", description: "Bounded advisory tail lines returned with await/read payloads." },
@@ -1039,6 +1052,135 @@ async function markTurnTerminalFromSessionState(
 	return resolved;
 }
 
+function runtimeStateAcknowledgesTurn(turn: TurnRecord, sessionState: CoordinatorSessionState | null): boolean {
+	return (
+		sessionState?.source === "agent_session_event" &&
+		sessionState.current_turn_id === turn.turn_id &&
+		(sessionState.state === "running" ||
+			sessionState.state === "needs_user_input" ||
+			sessionState.state === "completed" ||
+			sessionState.state === "errored")
+	);
+}
+
+async function markTurnAcknowledgedFromRuntimeState(
+	namespaceDir: string,
+	turn: TurnRecord,
+	sessionState: CoordinatorSessionState,
+): Promise<TurnRecord> {
+	if (turn.delivery.prompt_acknowledged === true && turn.delivery.state === "acknowledged") return turn;
+	const timestamp = new Date().toISOString();
+	const acknowledged: TurnRecord = {
+		...turn,
+		delivery: {
+			...turn.delivery,
+			delivered: true,
+			prompt_acknowledged: true,
+			state: "acknowledged",
+			attempts: [
+				...turn.delivery.attempts,
+				{
+					delivered: true,
+					created_at: sessionState.updated_at,
+					reason: "runtime_prompt_acknowledged",
+					channel: "runtime_ack",
+					tmux_keys_sent: turn.delivery.tmux_keys_sent,
+				},
+			],
+		},
+		updated_at: timestamp,
+	};
+	await writeTurnRecord(namespaceDir, acknowledged);
+	await writeActiveTurn(namespaceDir, acknowledged);
+	return acknowledged;
+}
+
+function turnAwaitingRuntimeAckExpired(turn: TurnRecord, nowMs: number, ackTimeoutMs: number): boolean {
+	if (!ACTIVE_TURN_STATUSES.has(turn.status)) return false;
+	if (turn.delivery.tmux_keys_sent !== true) return false;
+	if (turn.delivery.prompt_acknowledged === true) return false;
+	if (turn.delivery.state !== "tmux_keys_sent") return false;
+	const deliveredAt =
+		turn.delivery.attempts.findLast(attempt => attempt.channel === "tmux_keys")?.created_at ?? turn.updated_at;
+	const deliveredMs = Date.parse(deliveredAt);
+	return Number.isFinite(deliveredMs) && nowMs - deliveredMs >= ackTimeoutMs;
+}
+
+async function markTurnFailedForUnacknowledgedDelivery(
+	namespaceDir: string,
+	turn: TurnRecord,
+	ackTimeoutMs: number,
+): Promise<TurnRecord> {
+	const timestamp = new Date().toISOString();
+	const message = `Tmux key delivery succeeded, but the GJC runtime did not acknowledge the prompt or emit turn_start within ${ackTimeoutMs}ms. The turn never started; stop waiting and inspect/retry the coordinator session.`;
+	const failed: TurnRecord = {
+		...turn,
+		status: "failed",
+		delivery: {
+			...turn.delivery,
+			delivered: false,
+			queued: false,
+			prompt_acknowledged: false,
+			state: "unacknowledged",
+			attempts: [
+				...turn.delivery.attempts,
+				{
+					delivered: false,
+					created_at: timestamp,
+					reason: PROMPT_ACK_TIMEOUT_REASON,
+					channel: "runtime_ack",
+					tmux_keys_sent: true,
+				},
+			],
+		},
+		final_response: {
+			text: message,
+			format: "markdown",
+			source: "coordinator_delivery_ack_timeout",
+			artifact_path: null,
+			truncated: false,
+		},
+		error: { code: PROMPT_ACK_TIMEOUT_REASON, message, recoverable: true },
+		evidence: [
+			...turn.evidence,
+			{
+				type: PROMPT_ACK_TIMEOUT_REASON,
+				message,
+				tmux_keys_sent: true,
+				prompt_acknowledged: false,
+				created_at: timestamp,
+			},
+		],
+		liveness: { checked_at: timestamp, live: turn.liveness.live, reason: PROMPT_ACK_TIMEOUT_REASON },
+		updated_at: timestamp,
+		completed_at: timestamp,
+	};
+	await writeTurnRecord(namespaceDir, failed);
+	await clearActiveTurn(namespaceDir, failed);
+	await writeSessionState(namespaceDir, failed.session_id, "stale", {
+		lastTurnId: failed.turn_id,
+		live: failed.liveness.live,
+		reason: PROMPT_ACK_TIMEOUT_REASON,
+	});
+	return failed;
+}
+
+async function reconcileRuntimeAcknowledgement(
+	namespaceDir: string,
+	turn: TurnRecord,
+	sessionState: CoordinatorSessionState | null,
+	ackTimeoutMs: number,
+	options: { failOnTimeout: boolean } = { failOnTimeout: true },
+): Promise<TurnRecord> {
+	if (sessionState && runtimeStateAcknowledgesTurn(turn, sessionState)) {
+		return await markTurnAcknowledgedFromRuntimeState(namespaceDir, turn, sessionState);
+	}
+	if (options.failOnTimeout && turnAwaitingRuntimeAckExpired(turn, Date.now(), ackTimeoutMs)) {
+		return await markTurnFailedForUnacknowledgedDelivery(namespaceDir, turn, ackTimeoutMs);
+	}
+	return turn;
+}
+
 function shellQuote(value: string): string {
 	return `'${value.replaceAll("'", "'\\''")}'`;
 }
@@ -1086,16 +1228,33 @@ function asTerminalTurnStatus(status: unknown): TurnStatus | null {
 	return null;
 }
 
-function boundedTimeoutMs(value: unknown): number {
+export const COORDINATOR_AWAIT_TURN_TIMEOUT_MAX_MS = 30 * 60 * 1000;
+export const COORDINATOR_RUNTIME_PROMPT_ACK_TIMEOUT_MAX_MS = MAX_RUNTIME_PROMPT_ACK_TIMEOUT_MS;
+export const COORDINATOR_EVENT_WATCH_TIMEOUT_MAX_MS = 30_000;
+export const COORDINATOR_POLL_INTERVAL_MAX_MS = 10_000;
+
+function parsePositiveIntegerMs(value: unknown, fallback: number): number {
 	const parsed = typeof value === "number" ? value : Number.parseInt(String(value ?? ""), 10);
-	if (!Number.isFinite(parsed) || parsed <= 0) return 1000;
-	return Math.min(parsed, 30_000);
+	return Number.isFinite(parsed) && parsed > 0 ? parsed : fallback;
 }
 
-function boundedPollIntervalMs(value: unknown): number {
-	const parsed = typeof value === "number" ? value : Number.parseInt(String(value ?? ""), 10);
-	if (!Number.isFinite(parsed) || parsed <= 0) return 100;
-	return Math.min(Math.max(parsed, 10), 1000);
+export function boundedAwaitTurnTimeoutMs(value: unknown): number {
+	return Math.min(parsePositiveIntegerMs(value, 1000), COORDINATOR_AWAIT_TURN_TIMEOUT_MAX_MS);
+}
+
+export function boundedRuntimePromptAckTimeoutMs(value: unknown): number {
+	return Math.min(
+		parsePositiveIntegerMs(value, DEFAULT_RUNTIME_PROMPT_ACK_TIMEOUT_MS),
+		COORDINATOR_RUNTIME_PROMPT_ACK_TIMEOUT_MAX_MS,
+	);
+}
+
+export function boundedEventWatchTimeoutMs(value: unknown): number {
+	return Math.min(parsePositiveIntegerMs(value, 1000), COORDINATOR_EVENT_WATCH_TIMEOUT_MAX_MS);
+}
+
+export function boundedPollIntervalMs(value: unknown): number {
+	return Math.min(Math.max(parsePositiveIntegerMs(value, 100), 10), COORDINATOR_POLL_INTERVAL_MAX_MS);
 }
 async function runCommand(command: string[]): Promise<{ exitCode: number; stdout: string; stderr: string }> {
 	const proc = Bun.spawn(command, { stdout: "pipe", stderr: "pipe" });
@@ -1114,8 +1273,10 @@ async function sendTmuxPromptKeys(
 	prompt: string,
 	runner: CommandRunner = runCommand,
 ): Promise<boolean> {
-	const sent = await runner(["tmux", "send-keys", "-t", target, prompt, "C-m", "C-m"]);
-	return sent.exitCode === 0;
+	const typed = await runner(["tmux", "send-keys", "-t", target, "-l", prompt]);
+	if (typed.exitCode !== 0) return false;
+	const submitted = await runner(["tmux", "send-keys", "-t", target, "-l", "\x1b[13;5u"]);
+	return submitted.exitCode === 0;
 }
 
 function boundedLineCount(value: unknown): number {
@@ -1387,7 +1548,9 @@ export async function readCoordinatorArtifact(
 }
 
 export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions = {}) {
-	const config = buildCoordinatorMcpConfig(options.env ?? process.env);
+	const env = options.env ?? process.env;
+	const config = buildCoordinatorMcpConfig(env);
+	const promptAckTimeoutMs = boundedRuntimePromptAckTimeoutMs(env.GJC_COORDINATOR_MCP_PROMPT_ACK_TIMEOUT_MS);
 	const services = options.services ?? {};
 	const namespaceDir = coordinatorNamespacePath(config);
 	const commandRunner = services.commandRunner ?? runCommand;
@@ -1485,14 +1648,17 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 		await writeTurnRecord(namespaceDir, activeTurn);
 		await writeActiveTurn(namespaceDir, activeTurn);
 		const sessionState = await readSessionState(namespaceDir, activeTurn.session_id);
-		const runtimeStateAlreadySettled =
-			sessionState?.current_turn_id === activeTurn.turn_id &&
-			(sessionState.state === "completed" || sessionState.state === "errored");
-		if (!runtimeStateAlreadySettled) {
-			await writeSessionState(namespaceDir, activeTurn.session_id, tmuxKeysSent ? "running" : "stale", {
+		const runtimeStateAlreadyAcknowledged =
+			sessionState !== null && runtimeStateAcknowledgesTurn(activeTurn, sessionState);
+		const resolvedTurn =
+			runtimeStateAlreadyAcknowledged && sessionState
+				? await markTurnAcknowledgedFromRuntimeState(namespaceDir, activeTurn, sessionState)
+				: activeTurn;
+		if (!runtimeStateAlreadyAcknowledged && !tmuxKeysSent) {
+			await writeSessionState(namespaceDir, activeTurn.session_id, "stale", {
 				currentTurnId: activeTurn.turn_id,
 				live,
-				reason: tmuxKeysSent ? null : "tmux_delivery_unavailable",
+				reason: "tmux_delivery_unavailable",
 			});
 		}
 		await appendCoordinatorEvent(namespaceDir, {
@@ -1505,7 +1671,7 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 			payloadRef: path.relative(namespaceDir, turnFile(namespaceDir, activeTurn.turn_id)),
 			metadata: { target: typeof target === "string" ? target : null, live },
 		});
-		return activeTurn;
+		return resolvedTurn;
 	}
 
 	async function promoteNextQueuedTurn(sessionId: string): Promise<TurnRecord | null> {
@@ -1533,33 +1699,67 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 		let resolvedTurn = turn;
 		let advisoryStatus: Record<string, unknown> = { live: false };
 		let sessionState = await readSessionState(namespaceDir, turn.session_id);
+		resolvedTurn = await reconcileRuntimeAcknowledgement(
+			namespaceDir,
+			resolvedTurn,
+			sessionState,
+			promptAckTimeoutMs,
+			{ failOnTimeout: false },
+		);
+		if (resolvedTurn !== turn) sessionState = await readSessionState(namespaceDir, resolvedTurn.session_id);
 		if (
 			sessionState &&
-			ACTIVE_TURN_STATUSES.has(turn.status) &&
-			sessionState.current_turn_id === turn.turn_id &&
+			ACTIVE_TURN_STATUSES.has(resolvedTurn.status) &&
+			(sessionState.current_turn_id === resolvedTurn.turn_id ||
+				(sessionState.state === "errored" &&
+					sessionState.source === "agent_session_event" &&
+					sessionState.current_turn_id == null)) &&
 			(sessionState.state === "completed" || sessionState.state === "errored")
 		) {
-			resolvedTurn = await markTurnTerminalFromSessionState(namespaceDir, turn, sessionState);
+			resolvedTurn = await markTurnTerminalFromSessionState(namespaceDir, resolvedTurn, sessionState);
 			sessionState = await readSessionState(namespaceDir, resolvedTurn.session_id);
 		} else if (
 			sessionState &&
-			ACTIVE_TURN_STATUSES.has(turn.status) &&
-			sessionState.current_turn_id === turn.turn_id &&
+			ACTIVE_TURN_STATUSES.has(resolvedTurn.status) &&
+			sessionState.current_turn_id === resolvedTurn.turn_id &&
 			sessionState.state === "stale" &&
 			sessionState.reason === "tmux_delivery_unavailable" &&
-			turn.delivery.state === "unavailable" &&
+			resolvedTurn.delivery.state === "unavailable" &&
 			session &&
 			hasTmuxIdentity(session)
 		) {
-			resolvedTurn = await markTurnFailedForUnavailableSession(namespaceDir, turn, "tmux_delivery_unavailable");
+			resolvedTurn = await markTurnFailedForUnavailableSession(
+				namespaceDir,
+				resolvedTurn,
+				"tmux_delivery_unavailable",
+			);
 			sessionState = await readSessionState(namespaceDir, resolvedTurn.session_id);
-		} else if (!session && ACTIVE_TURN_STATUSES.has(turn.status)) {
-			resolvedTurn = await markTurnFailedForUnavailableSession(namespaceDir, turn, "session_record_missing");
+		} else if (!session && ACTIVE_TURN_STATUSES.has(resolvedTurn.status)) {
+			resolvedTurn = await markTurnFailedForUnavailableSession(namespaceDir, resolvedTurn, "session_record_missing");
 			sessionState = await readSessionState(namespaceDir, resolvedTurn.session_id);
 		} else if (session) {
 			advisoryStatus = await inspectTmuxSession(session, boundedLineCount(lines), commandRunner);
-			if (ACTIVE_TURN_STATUSES.has(turn.status) && hasTmuxIdentity(session) && advisoryStatus.live === false) {
-				resolvedTurn = await markTurnFailedForUnavailableSession(namespaceDir, turn, "tmux_session_missing");
+			if (
+				ACTIVE_TURN_STATUSES.has(resolvedTurn.status) &&
+				hasTmuxIdentity(session) &&
+				advisoryStatus.live === false
+			) {
+				resolvedTurn = await markTurnFailedForUnavailableSession(
+					namespaceDir,
+					resolvedTurn,
+					"tmux_session_missing",
+				);
+				sessionState = await readSessionState(namespaceDir, resolvedTurn.session_id);
+			}
+		}
+		if (ACTIVE_TURN_STATUSES.has(resolvedTurn.status)) {
+			resolvedTurn = await reconcileRuntimeAcknowledgement(
+				namespaceDir,
+				resolvedTurn,
+				sessionState,
+				promptAckTimeoutMs,
+			);
+			if (!ACTIVE_TURN_STATUSES.has(resolvedTurn.status)) {
 				sessionState = await readSessionState(namespaceDir, resolvedTurn.session_id);
 			}
 		}
@@ -1579,6 +1779,46 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 		};
 	}
 
+	async function reconcileActiveTurnAcknowledgements(): Promise<void> {
+		const turns = (await listJsonFiles(turnsDir(namespaceDir)))
+			.map(turn => asRecord(turn) as TurnRecord | null)
+			.filter((turn): turn is TurnRecord => turn !== null && ACTIVE_TURN_STATUSES.has(turn.status));
+		for (const turn of turns) {
+			let sessionState = await readSessionState(namespaceDir, turn.session_id);
+			const resolvedTurn = await reconcileRuntimeAcknowledgement(
+				namespaceDir,
+				turn,
+				sessionState,
+				promptAckTimeoutMs,
+				{ failOnTimeout: false },
+			);
+			if (!ACTIVE_TURN_STATUSES.has(resolvedTurn.status)) continue;
+			if (resolvedTurn !== turn) sessionState = await readSessionState(namespaceDir, resolvedTurn.session_id);
+			const session = asRecord(await readJsonFile(sessionFile(resolvedTurn.session_id)));
+			if (
+				sessionState &&
+				sessionState.current_turn_id === resolvedTurn.turn_id &&
+				sessionState.state === "stale" &&
+				sessionState.reason === "tmux_delivery_unavailable" &&
+				resolvedTurn.delivery.state === "unavailable" &&
+				session &&
+				hasTmuxIdentity(session)
+			) {
+				await markTurnFailedForUnavailableSession(namespaceDir, resolvedTurn, "tmux_delivery_unavailable");
+				continue;
+			}
+			if (!session) {
+				await markTurnFailedForUnavailableSession(namespaceDir, resolvedTurn, "session_record_missing");
+				continue;
+			}
+			if (hasTmuxIdentity(session) && (await hasTmuxSession(session, commandRunner)) === false) {
+				await markTurnFailedForUnavailableSession(namespaceDir, resolvedTurn, "tmux_session_missing");
+				continue;
+			}
+			await reconcileRuntimeAcknowledgement(namespaceDir, resolvedTurn, sessionState, promptAckTimeoutMs);
+		}
+	}
+
 	async function callTool(name: string, args: Record<string, unknown> = {}): Promise<Record<string, unknown>> {
 		try {
 			if (name === "gjc_coordinator_list_sessions") return { ok: true, sessions: await listSessions() };
@@ -1619,6 +1859,7 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 				};
 			}
 			if (name === "gjc_coordinator_read_status") {
+				await reconcileActiveTurnAcknowledgements();
 				const sessionId = args.session_id;
 				if (sessionId) {
 					const session = asRecord(await readJsonFile(sessionFile(sessionId)));
@@ -1651,6 +1892,7 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 			if (name === "gjc_coordinator_read_artifact")
 				return await readCoordinatorArtifact(config, { path: args.path });
 			if (name === "gjc_coordinator_read_coordination_status") {
+				await reconcileActiveTurnAcknowledgements();
 				const sessions = jsonRecords(await listSessions());
 				const sessionStates = jsonRecords(await listJsonFiles(path.join(namespaceDir, "session-states")));
 				const turns = jsonRecords(await listJsonFiles(turnsDir(namespaceDir)));
@@ -1684,13 +1926,15 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 				};
 			}
 			if (name === "gjc_coordinator_watch_events") {
+				await reconcileActiveTurnAcknowledgements();
 				const limit = boundedEventLimit(args.limit);
-				const timeoutMs = boundedTimeoutMs(args.timeout_ms);
+				const timeoutMs = boundedEventWatchTimeoutMs(args.timeout_ms);
 				let events = await readCoordinatorEvents(namespaceDir);
 				let matched = filterCoordinatorEvents(events, args, limit);
 				let timedOut = false;
 				if (matched.length === 0 && timeoutMs > 0) {
 					await waitForCoordinatorEvents(namespaceDir, timeoutMs);
+					await reconcileActiveTurnAcknowledgements();
 					events = await readCoordinatorEvents(namespaceDir);
 					matched = filterCoordinatorEvents(events, args, limit);
 					timedOut = matched.length === 0;
@@ -1819,7 +2063,7 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 				};
 				if (promptAliasIgnored) base.prompt_alias_ignored = true;
 				if (args.await_completion === true && !shouldQueue) {
-					const timeoutMs = boundedTimeoutMs(args.timeout_ms);
+					const timeoutMs = boundedAwaitTurnTimeoutMs(args.timeout_ms);
 					const pollIntervalMs = boundedPollIntervalMs(args.poll_interval_ms);
 					const deadline = Date.now() + timeoutMs;
 					let payload = await readTurnPayload(turn.turn_id, sessionId, args.lines);
@@ -1979,7 +2223,7 @@ export function createCoordinatorMcpServer(options: CoordinatorMcpServerOptions
 				return await readTurnPayload(args.turn_id, args.session_id, args.lines);
 			}
 			if (name === "gjc_coordinator_await_turn") {
-				const timeoutMs = boundedTimeoutMs(args.timeout_ms);
+				const timeoutMs = boundedAwaitTurnTimeoutMs(args.timeout_ms);
 				const pollIntervalMs = boundedPollIntervalMs(args.poll_interval_ms);
 				const deadline = Date.now() + timeoutMs;
 				let payload = await readTurnPayload(args.turn_id, args.session_id, args.lines);

package/src/deep-interview/render-middleware.ts

@@ -49,6 +49,7 @@ interface ThresholdModel {
 type DeepInterviewModel = RoundQuestionModel | TopologyQuestionModel | ProgressModel | ThresholdModel;
 
 function normalizeText(text: string): string {
+	if (typeof text !== "string") return "";
 	return text.trim().replaceAll("\r\n", "\n");
 }
 

package/src/defaults/gjc/skills/deep-interview/SKILL.md

@@ -51,6 +51,10 @@ Inspired by the [Ouroboros project](https://github.com/Q00/ouroboros) which demo
 - Keep prompt payloads budgeted: summarize or trim oversized initial context/history before composing question, scoring, spec, or handoff prompts
 - If the user's initial context is oversized, create a concise prompt-safe summary first and wait for that summary before ambiguity scoring, question generation, or downstream execution handoff
 - Do not proceed to execution until ambiguity ≤ the resolved threshold for this run and the user explicitly approves a scoped execution path
+- Treat user wording such as `implementation`, "implementation plan", Korean `구현`, or "구현 계획" as describing the eventual target, not permission to implement now.
+- While still in deep-interview, do not implement, edit/write code, launch implementation workers, or start task/skill/ultragoal implementation; continue interviewing for scope, risks, acceptance criteria, and unknowns.
+- When the user wants interview output for eventual implementation, say: "I can interview for an implementation plan, but I won't implement during deep-interview." Then continue clarifying scope, risks, acceptance criteria, and unknowns.
+- Implementation requires an explicit phase transition/approval after the interview: deep-interview must first produce its spec/handoff, the workflow phase must explicitly transition out of deep-interview, and execution approval must be captured by a downstream execution path.
 - Allow early exit with a clear warning if ambiguity is still high
 - Persist interview state for resume across session interruptions
 - A multi-persona lateral-review panel convenes at ambiguity-milestone transitions (and before synthesizing any agent-supplied answer) to expose blind spots from independent perspectives
@@ -74,7 +78,7 @@ Inspired by the [Ouroboros project](https://github.com/Q00/ouroboros) which demo
 
 ## Native Plugin Invocation Guard (Issue #3030)
 
-If this raw bundled skill is loaded by GJC's native skill loader through `/skill:deep-interview`, do not treat that path as permission to skip rendered GJC setup. The user-facing invocation is `/skill:deep-interview`; do not recommend or advertise CLI bridge commands as the deep-interview entrypoint. Regardless of invocation path, Phase 0 below remains blocking and must resolve `gjc.deepInterview.ambiguityThreshold` from settings before any announcement, state write, question, or ambiguity score.
+If this raw bundled skill is loaded by GJC's native skill loader through `/skill:deep-interview`, do not treat that path as permission to skip rendered GJC setup. The user-facing invocation is `/skill:deep-interview`; do not recommend or advertise CLI bridge commands as the deep-interview entrypoint. Regardless of invocation path, Phase 0 below remains blocking and must resolve `gjc.deepInterview.ambiguityThreshold` from pre-resolved native state or settings before any announcement, state write, question, or ambiguity score.
 
 ## Corrupt current-session state recovery
 
@@ -84,20 +88,24 @@ When deep-interview detects its own current-session state is corrupt, tampered,
 
 Complete this phase before Phase 1, before brownfield exploration, before GJC state persistence, before Round 0, and before any ambiguity scoring. Do not continue if the resolved threshold and source are unknown.
 
-1. **Read threshold settings in precedence order**:
+1. **Prefer pre-resolved native state**:
+   - First inspect active deep-interview state with `gjc state deep-interview read --json`.
+   - If state contains a finite numeric `threshold` and a non-empty `threshold_source`, use those values, set `<resolvedThreshold>`, `<resolvedThresholdPercent>`, and `<resolvedThresholdSource>`, and skip optional settings-file reads. This is the normal `/skill:deep-interview` path because the native hook already resolved settings quietly before loading the skill.
+2. **Only if native state lacks a resolved threshold, read threshold settings in precedence order**:
    - User settings: `[$GJC_CONFIG_DIR|~/.gjc]/settings.json`
    - Project settings: `./.gjc/settings.json` (overrides user settings)
-2. **Resolve threshold and source**:
-   - Read `gjc.deepInterview.ambiguityThreshold` from both files when present.
+   - Read `gjc.deepInterview.ambiguityThreshold` only from files that are known to exist; optional settings-file absence is expected and must not be surfaced as failed `Read` calls.
+   - Do not probe arbitrary ancestor candidates such as `../../.gjc/settings.json`; use the current project `.gjc/settings.json` and user settings only.
+3. **Resolve threshold and source**:
    - Use the project value when valid; otherwise use the user value when valid; otherwise use the default `0.05`.
    - Set these run variables exactly: `<resolvedThreshold>`, `<resolvedThresholdPercent>`, and `<resolvedThresholdSource>` (for example `./.gjc/settings.json`, `[$GJC_CONFIG_DIR|~/.gjc]/settings.json`, or `default`).
-3. **Emit the required first line to the user before any other interview announcement**:
+4. **Emit the required first line to the user before any other interview announcement**:
 
 ```
 Deep Interview threshold: <resolvedThresholdPercent> (source: <resolvedThresholdSource>)
 ```
 
-4. **Carry threshold source forward mechanically**:
+5. **Carry threshold source forward mechanically**:
    - Substitute `<resolvedThreshold>`, `<resolvedThresholdPercent>`, and `<resolvedThresholdSource>` throughout the remaining instructions before continuing.
    - Include `threshold_source` in the first `gjc state write` payload and preserve it on later state updates; do not edit `.gjc/_session-{sessionid}/state` files directly unless an explicit force override is active.
    - Include both threshold and source in the final spec metadata.

package/src/defaults/gjc/skills/ralplan/SKILL.md

@@ -19,7 +19,7 @@ Ralplan is the consensus planning workflow. It triggers iterative planning with
 
 ## Flags
 
-- `--interactive`: Enables user prompts at key decision points (draft review in step 2 and final approval in step 6). Without this flag the workflow runs fully automated — Planner → Architect → Critic loop — marks the final plan `pending approval`, outputs it, and stops without asking for confirmation or executing changes.
+- `--interactive`: Enables extra mid-loop user prompts (draft review in step 2 and one-at-a-time reconciliation in step 6c). Regardless of this flag, the workflow always finishes the post-interview gate with an `ask`-tool prompt offering Refine further / Approve ultragoal / Approve team / Stop here, and never auto-executes — execution always requires explicit approval through that prompt.
 - `--deliberate`: Forces deliberate mode for high-risk work. Adds pre-mortem (3 scenarios) and expanded test planning (unit/integration/e2e/observability). Without this flag, deliberate mode can still auto-enable when the request explicitly signals high risk (auth/security, migrations, destructive changes, production incidents, compliance/PII, public API breakage).
 - `--architect openai-code`: Use OpenAI code for the Architect pass when OpenAI code CLI is available. Otherwise, briefly note the fallback and keep the default GJC Architect review.
 - `--critic openai-code`: Use OpenAI code for the Critic pass when OpenAI code CLI is available. Otherwise, briefly note the fallback and keep the default GJC Critic review.
@@ -83,13 +83,20 @@ The consensus workflow:
 6. **Post-ralplan interview** (intent reconciliation gate): After Critic returns `APPROVE` and before the plan is finalized, reconcile the consensus plan against the user's actual intent. The goal is to make sure ralplan did not silently bake in assumptions that conflict with what the user wants.
    a. **Collect open items** from the run: every assumption the Planner/Architect/Critic resolved by assumption rather than by stated fact, every ambiguity flagged during review, and every decision the loop made without explicit user input. Source these from the persisted `planner`/`architect`/`critic`/`revision` stage artifacts, not from memory.
    b. **Cross-check prior context for conflicts**: glob `.gjc/_session-{sessionid}/specs/deep-interview-*.md` and other prior specs/plans/context relevant by topic. For each, list points where the consensus plan contradicts, weakens, or expands beyond a previously crystallized decision, constraint, or non-goal. Cite the conflicting artifact and line/section.
-   c. **Reconcile with the user**:
-      - *(--interactive only)* Use the `ask` tool to confirm the open assumptions and conflicts **one at a time**, weakest/highest-impact first, polishing intent. If any confirmation reveals that the plan diverges from user intent, route the consolidated correction back into the re-review loop (step 5b Planner revision) and re-run Architect + Critic before returning here. Cap at the same 5-iteration ceiling.
-      - *(automated mode)* Do not ask. Embed every unconfirmed assumption and every detected prior-context conflict into the final plan under an **## Intent Reconciliation** section as explicit open confirmations the user must review at the `pending approval` gate, so nothing is silently assumed.
+   c. **Reconcile with the user via the `ask` tool (always, regardless of `--interactive`)**: Never stop idle with plain-text prose after the consensus loop. Every reconciliation question MUST go through the `ask` tool with contextual options plus free-text.
+      - If open items exist, confirm the open assumptions and conflicts **one at a time** with the `ask` tool, weakest/highest-impact first, polishing intent. If any confirmation reveals that the plan diverges from user intent, route the consolidated correction back into the re-review loop (step 5b Planner revision) and re-run Architect + Critic before returning here. Cap at the same 5-iteration ceiling.
+      - If the plan is crystal clear (no open assumptions or prior-context conflicts), skip straight to the step 8 final-options `ask` instead of inventing filler questions.
+      - For every confirmed open item, embed the resolved outcome into the final plan under an **## Intent Reconciliation** section so the `pending approval` artifact records each decision; record any item the user explicitly defers as an open confirmation under that same section.
    d. Persist the reconciliation with `gjc ralplan --write --stage post-interview --stage_n <N> --artifact "..." --json`, then return the receipt/path plus a compact status (reconciled-clean / reconciled-with-revision / open-confirmations-pending) instead of pasting the full body.
-7. On reconciliation completion, mark the plan `pending approval` unless explicit execution approval has already been captured, persist the ADR/final plan via `gjc ralplan --write --stage final --stage_n <N> --artifact "..."`, and do not directly edit `.gjc/_session-{sessionid}/plans`. *(--interactive only)* If `--interactive` is set, use the `ask` tool to present the plan with approval options (Approve execution via ultragoal (Recommended) / Approve execution via team (only when tmux-based interactive worker parallelization is required) / Compact then return for execution approval / Request changes / Reject). Final plan must include ADR (Decision, Drivers, Alternatives considered, Why chosen, Consequences, Follow-ups) and, when present, the **## Intent Reconciliation** section. Otherwise, output the final plan and stop before any mutation or delegation.
-8. *(--interactive only)* User chooses: Approve ultragoal execution (recommended), Approve team execution (tmux parallelization only), Request changes, or Reject
-9. *(--interactive only)* On approval: invoke `/skill:ultragoal` for execution by default; invoke `/skill:team` only when the user explicitly needs tmux-based interactive worker parallelization -- never implement directly
+7. On reconciliation completion, mark the plan `pending approval` unless explicit execution approval has already been captured, persist the ADR/final plan via `gjc ralplan --write --stage final --stage_n <N> --artifact "..."`, and do not directly edit `.gjc/_session-{sessionid}/plans`. Final plan must include ADR (Decision, Drivers, Alternatives considered, Why chosen, Consequences, Follow-ups) and, when present, the **## Intent Reconciliation** section.
+8. **Always** present the finalized plan via the `ask` tool (regardless of `--interactive`) with `workflowGate: { stage: "ralplan", kind: "approval" }` on the final question so RPC/headless clients receive a `ralplan`/`approval` workflow gate, not a deep-interview question gate. Use these options:
+   - **Refine further** — re-run the consensus loop / request changes, then return here
+   - **Approve execution via ultragoal (Recommended)** — goal-tracked autonomous execution
+   - **Approve execution via team** — only when tmux-based interactive worker parallelization is required
+   - **Stop here** — keep the plan as `pending approval` and make no further changes
+
+   Always include a free-text option. Do not stop with plain text and no `ask`; the post-interview gate's terminal action is this `ask`.
+9. On approval: invoke `/skill:ultragoal` for execution by default; invoke `/skill:team` only when the user explicitly needs tmux-based interactive worker parallelization. On **Refine further**, return to the step 5 re-review loop. On **Stop here**, leave the `pending approval` artifact and stop. Never implement directly.
 
    Before invoking `/skill:team` or `/skill:ultragoal`, mark ralplan ready for handoff so the skill tool's chain guard permits the transition:
 

package/src/gjc-runtime/deep-interview-runtime.ts

@@ -100,7 +100,7 @@ async function resolveSpecContent(rawSpec: string, cwd: string): Promise<string>
 		if (stat.isFile()) return await fs.readFile(candidate, "utf-8");
 	} catch (error) {
 		const err = error as NodeJS.ErrnoException;
-		if (err.code !== "ENOENT" && err.code !== "ENOTDIR") {
+		if (err.code !== "ENOENT" && err.code !== "ENOTDIR" && err.code !== "ENAMETOOLONG") {
 			throw new DeepInterviewCommandError(2, `failed to read --spec ${candidate}: ${err.message}`);
 		}
 	}