fix(kiro): improve direct session cache reuse

Reshape Kiro direct requests so resumed client sessions reuse Kiro's
cache-affinity fields instead of starting unrelated CodeWhisperer
conversations.

- keep conversationState.conversationId stable when the client sends an
  explicit session id (x-session-id, session_id, conversation_id, Claude
  Code session metadata)
- add a stable conversationState.agentContinuationId per Kiro session
- send conversationState.agentTaskType: "vibe" and agentMode: "vibe",
  matching the normal Kiro CLI/KAS chat path
- move Kiro thinking instructions into Kiro-compatible systemPrompt /
  additionalModelRequestFields instead of generic top-level thinking
- keep volatile timestamp context out of the top-level systemPrompt; it
  remains only in user content fallback
- suppress additionalModelRequestFields for legacy 4.5-era Claude/Kiro
  models that reject it, while defaulting future Claude/Kiro model ids
  to supported
- preserve Kiro meteringEvent credit usage internally for accounting
  without leaking provider-specific fields into OpenAI-compatible usage
- prevent unrelated headerless Kiro requests from sharing one
  connection-wide continuation
- cap/evict continuation sessions so long-running processes do not grow
  the continuation map unbounded
- treat generated headerless Kiro sessions as one-shot so they do not
  evict real explicit-session continuations
- keep credit-only Kiro metering valid for internal persistence when
  token metrics are unavailable

open-sse/config/kiroConstants.js

@@ -131,6 +131,50 @@ export function resolveKiroThinkingBudget(body, headers, model) {
   return null;
 }
 
+export function extractKiroEffortLevel(body) {
+  const effort =
+    body?.output_config?.effort ??
+    body?.reasoning_effort ??
+    (typeof body?.reasoning === "object" ? body.reasoning?.effort : null);
+  if (typeof effort !== "string") return null;
+  const normalized = effort.toLowerCase();
+  if (normalized === "none" || normalized === "off" || normalized === "disabled") return null;
+  if (normalized === "xhigh" || normalized === "max") return "high";
+  if (["low", "medium", "high"].includes(normalized)) return normalized;
+  return null;
+}
+
+export function buildKiroAdditionalModelRequestFields(body) {
+  const effort = extractKiroEffortLevel(body);
+  if (!effort) return undefined;
+  // Mirrors Kiro CLI/KAS buildEffortRequestFields("output_config").
+  return {
+    thinking: { type: "adaptive", display: "summarized" },
+    output_config: { effort },
+  };
+}
+
+export function supportsKiroAdditionalModelRequestFields(model) {
+  if (typeof model !== "string") return false;
+  const normalized = model.toLowerCase().replace(/-/g, ".");
+  if (!normalized.includes("claude")) return false;
+  const match = normalized.match(/(?:^|[/.])claude(?:[/.][a-z]+)*[/.](\d+)(?:[/.](\d+))?(?:[/.]|$)/);
+  if (!match) return false;
+  const [, majorText, minorText] = match;
+  const major = Number(majorText);
+  const minor = minorText === undefined ? null : Number(minorText);
+  const dateSuffixMinor = minor !== null && minor >= 1000;
+  // Kiro rejected additionalModelRequestFields on legacy 4.5 models in live smoke.
+  // Default future Claude/Kiro models to supported so new model releases do not
+  // need a code allowlist update.
+  return !(major < 4 || (major === 4 && (minor === null || minor <= 5 || dateSuffixMinor)));
+}
+
+export function buildKiroAdditionalModelRequestFieldsForModel(body, model) {
+  if (!supportsKiroAdditionalModelRequestFields(model)) return undefined;
+  return buildKiroAdditionalModelRequestFields(body);
+}
+
 /**
  * Detect whether an inbound request is asking for reasoning / thinking output.
  * Thin wrapper over resolveKiroThinkingBudget (single source of truth).

open-sse/translator/index.js

@@ -103,8 +103,16 @@ export function translateRequest(sourceFormat, targetFormat, model, body, stream
     }
   }
 
-  // Normalize thinking to the target provider-native format (config-driven, capability-aware)
+  // Normalize thinking to the target provider-native format (config-driven, capability-aware).
-  applyThinking(targetFormat, model, result, provider, thinkingIntent);
+  // Kiro's GenerateAssistantResponse request does not accept the generic top-level
+  // `thinking` field; its translators map thinking intent to KAS-compatible
+  // systemPrompt/additionalModelRequestFields instead.
+  const kiroThinkingMappedByTranslator =
+    targetFormat === FORMATS.KIRO &&
+    (sourceFormat === FORMATS.OPENAI || sourceFormat === FORMATS.CLAUDE);
+  if (!kiroThinkingMappedByTranslator) {
+    applyThinking(targetFormat, model, result, provider, thinkingIntent);
+  }
 
   // Always normalize to clean OpenAI format when target is OpenAI
   // This handles hybrid requests (e.g., OpenAI messages + Claude tools)

open-sse/translator/request/claude-to-kiro.js

@@ -24,13 +24,15 @@
  */
 import { register } from "../index.js";
 import { FORMATS } from "../formats.js";
-import { v4 as uuidv4 } from "uuid";
+import { applyKiroSessionReplay } from "../../utils/kiroSessionReplay.js";
+import { resolveContinuationId, resolveSessionIdentity } from "../../utils/sessionManager.js";
 import {
   resolveKiroModel,
   resolveKiroThinkingBudget,
   buildThinkingSystemPrefix,
   KIRO_AGENTIC_SYSTEM_PROMPT,
   resolveDefaultProfileArn,
+  buildKiroAdditionalModelRequestFieldsForModel,
 } from "../../config/kiroConstants.js";
 import { DEFAULT_IMAGE_MIME } from "../schema/index.js";
 import { ROLE, CLAUDE_BLOCK } from "../schema/index.js";
@@ -363,6 +365,18 @@ function reconcileOrphanedToolResults(history, currentMessage) {
   }
 }
 
+function extractClaudeSystemText(system) {
+  if (!system) return "";
+  if (typeof system === "string") return system;
+  if (Array.isArray(system)) {
+    return system.map((s) => {
+      if (typeof s === "string") return s;
+      return s?.text || "";
+    }).filter(Boolean).join("\n");
+  }
+  return "";
+}
+
 /**
  * Build a Kiro payload directly from a Claude Messages API request body.
  */
@@ -402,62 +416,75 @@ export function claudeToKiroRequest(model, body, stream, credentials) {
     ? (credentials?.providerSpecificData?.profileArn || "")
     : (credentials?.providerSpecificData?.profileArn || resolveDefaultProfileArn(authMethod));
 
-  let finalContent = currentMessage?.userInputMessage?.content || "";
+  // Kiro CLI/KAS sends system prompt as top-level `systemPrompt`. Keep a
-
+  // content fallback too because the CodeWhisperer surface does not always
-  // System prompt: pass via native systemInstruction field (Kiro/Q API supports it)
+  // enforce top-level systemPrompt for direct calls.
-  // and also prepend as <instructions> in user content as fallback for upstreams
-  // that don't support the native field.
-  let systemInstruction = undefined;
-  if (body.system) {
-    let systemText = "";
-    if (typeof body.system === "string") {
-      systemText = body.system;
-    } else if (Array.isArray(body.system)) {
-      systemText = body.system.map((s) => s.text || "").join("\n");
-    }
-    if (systemText) {
-      systemInstruction = systemText;
-      finalContent = `<instructions>\n${systemText}\n</instructions>\n\n${finalContent}`;
-    }
-  }
-
-  // Prefix order: thinking_mode tag, timestamp marker, then agentic prompt.
   const timestamp = new Date().toISOString();
-  const prefixParts = [];
+  const systemPromptParts = [];
-  if (thinkingBudget !== null) prefixParts.push(buildThinkingSystemPrefix(thinkingBudget));
+  if (thinkingBudget !== null) systemPromptParts.push(buildThinkingSystemPrefix(thinkingBudget));
-  prefixParts.push(`[Context: Current time is ${timestamp}]`);
+  if (agentic) systemPromptParts.push(KIRO_AGENTIC_SYSTEM_PROMPT);
-  if (agentic) prefixParts.push(KIRO_AGENTIC_SYSTEM_PROMPT);
+  const systemInstruction = extractClaudeSystemText(body.system);
-  finalContent = `${prefixParts.join("\n\n")}\n\n${finalContent}`;
+  if (systemInstruction) systemPromptParts.push(systemInstruction);
+  const systemPrompt = systemPromptParts.filter(Boolean).join("\n\n");
+  const currentTimeContext = `[Context: Current time is ${timestamp}]`;
+  const contentPrefix = [systemPrompt, currentTimeContext].filter(Boolean).join("\n\n");
 
+  const sessionIdentity = resolveSessionIdentity({
+    headers: credentials?.rawHeaders,
+    body,
+    connectionId: credentials?.connectionId,
+    scope: "kiro",
+  });
+  const conversationId = sessionIdentity.sessionId;
+  const continuationId = resolveContinuationId({
+    sessionId: conversationId,
+    connectionId: credentials?.connectionId,
+    scope: "kiro",
+    ephemeral: sessionIdentity.ephemeral,
+  });
+  const replay = applyKiroSessionReplay({
+    conversationId,
+    connectionId: credentials?.connectionId,
+    modelId: upstreamModel,
+    systemPrompt,
+    contentPrefix,
+    currentContentPrefix: currentTimeContext,
+    history,
+    currentMessage,
+  });
+  const replayCurrent = replay.currentMessage?.userInputMessage || {};
   const userInputMessage = {
-    content: finalContent,
+    content: replayCurrent.content || "",
     modelId: upstreamModel,
     origin: "AI_EDITOR",
-    ...(currentMessage?.userInputMessage?.userInputMessageContext && {
+    ...(replayCurrent.userInputMessageContext && {
-      userInputMessageContext:
+      userInputMessageContext: replayCurrent.userInputMessageContext,
-        currentMessage.userInputMessage.userInputMessageContext,
     }),
-    ...(currentMessage?.userInputMessage?.images && {
+    ...(replayCurrent.images && {
-      images: currentMessage.userInputMessage.images,
+      images: replayCurrent.images,
     }),
   };
 
-  if (systemInstruction) {
-    userInputMessage.systemInstruction = systemInstruction;
-  }
-
   const payload = {
     conversationState: {
       chatTriggerType: "MANUAL",
-      conversationId: uuidv4(),
+      conversationId,
+      agentContinuationId: continuationId,
+      agentTaskType: "vibe",
       currentMessage: {
         userInputMessage,
       },
-      history,
+      history: replay.history,
     },
+    agentMode: "vibe",
   };
 
   if (profileArn) payload.profileArn = profileArn;
+  if (systemPrompt) payload.systemPrompt = systemPrompt;
+  const additionalModelRequestFields = buildKiroAdditionalModelRequestFieldsForModel(body, upstreamModel);
+  if (additionalModelRequestFields) {
+    payload.additionalModelRequestFields = additionalModelRequestFields;
+  }
 
   if (maxTokens || temperature !== undefined || topP !== undefined) {
     payload.inferenceConfig = {};

open-sse/translator/request/openai-to-kiro.js

@@ -5,13 +5,15 @@
 import { register } from "../index.js";
 import { FORMATS } from "../formats.js";
 import { v4 as uuidv4 } from "uuid";
-import { resolveSessionId } from "../../utils/sessionManager.js";
+import { applyKiroSessionReplay } from "../../utils/kiroSessionReplay.js";
+import { resolveContinuationId, resolveSessionIdentity } from "../../utils/sessionManager.js";
 import {
   resolveKiroModel,
   resolveKiroThinkingBudget,
   buildThinkingSystemPrefix,
   KIRO_AGENTIC_SYSTEM_PROMPT,
-  resolveDefaultProfileArn
+  resolveDefaultProfileArn,
+  buildKiroAdditionalModelRequestFieldsForModel
 } from "../../config/kiroConstants.js";
 import { parseDataUri } from "../concerns/image.js";
 import { DEFAULT_IMAGE_MIME } from "../schema/index.js";
@@ -546,47 +548,74 @@ export function openaiToKiroRequest(model, body, stream, credentials) {
     ? (credentials?.providerSpecificData?.profileArn || "")
     : (credentials?.providerSpecificData?.profileArn || resolveDefaultProfileArn(authMethod));
 
-  let finalContent = currentMessage?.userInputMessage?.content || "";
-
   const timestamp = new Date().toISOString();
 
-  // Build the system-prompt prefix that goes ABOVE the user message body.
+  // Kiro CLI/KAS sends these as top-level systemPrompt. Keep a content fallback
-  // Order: thinking_mode tag first (so Kiro sees it before any user text),
+  // too because the CodeWhisperer surface does not always enforce top-level
-  // then context/timestamp marker, then optional agentic chunked-write prompt.
+  // systemPrompt for direct calls.
-  const prefixParts = [];
+  const systemPromptParts = [];
   if (thinkingBudget !== null) {
-    prefixParts.push(buildThinkingSystemPrefix(thinkingBudget));
+    systemPromptParts.push(buildThinkingSystemPrefix(thinkingBudget));
   }
-  prefixParts.push(`[Context: Current time is ${timestamp}]`);
   if (agentic) {
-    prefixParts.push(KIRO_AGENTIC_SYSTEM_PROMPT);
+    systemPromptParts.push(KIRO_AGENTIC_SYSTEM_PROMPT);
   }
-  finalContent = `${prefixParts.join("\n\n")}\n\n${finalContent}`;
+  const systemPrompt = systemPromptParts.filter(Boolean).join("\n\n");
+  const currentTimeContext = `[Context: Current time is ${timestamp}]`;
+  const contentPrefix = [systemPrompt, currentTimeContext].filter(Boolean).join("\n\n");
+
+  const sessionIdentity = resolveSessionIdentity({ headers: credentials?.rawHeaders, body, connectionId: credentials?.connectionId, scope: "kiro" });
+  const conversationId = sessionIdentity.sessionId;
+  const continuationId = resolveContinuationId({
+    sessionId: conversationId,
+    connectionId: credentials?.connectionId,
+    scope: "kiro",
+    ephemeral: sessionIdentity.ephemeral,
+  });
+  const replay = applyKiroSessionReplay({
+    conversationId,
+    connectionId: credentials?.connectionId,
+    modelId: upstreamModel,
+    systemPrompt,
+    contentPrefix,
+    currentContentPrefix: currentTimeContext,
+    history,
+    currentMessage,
+  });
+  const replayCurrent = replay.currentMessage?.userInputMessage || {};
 
   const payload = {
     conversationState: {
       chatTriggerType: "MANUAL",
-      conversationId: resolveSessionId({ headers: credentials?.rawHeaders, body, connectionId: credentials?.connectionId, scope: "kiro" }),
+      conversationId,
+      agentContinuationId: continuationId,
+      agentTaskType: "vibe",
       currentMessage: {
         userInputMessage: {
-          content: finalContent,
+          content: replayCurrent.content || "",
           modelId: upstreamModel,
           origin: "AI_EDITOR",
-          ...(currentMessage?.userInputMessage?.images?.length > 0 && {
+          ...(replayCurrent.images?.length > 0 && {
-            images: currentMessage.userInputMessage.images
+            images: replayCurrent.images
           }),
-          ...(currentMessage?.userInputMessage?.userInputMessageContext && {
+          ...(replayCurrent.userInputMessageContext && {
-            userInputMessageContext: currentMessage.userInputMessage.userInputMessageContext
+            userInputMessageContext: replayCurrent.userInputMessageContext
           })
         }
       },
-      history: history
+      history: replay.history
-    }
+    },
+    agentMode: "vibe",
   };
 
   if (profileArn) {
     payload.profileArn = profileArn;
   }
+  if (systemPrompt) payload.systemPrompt = systemPrompt;
+  const additionalModelRequestFields = buildKiroAdditionalModelRequestFieldsForModel(body, upstreamModel);
+  if (additionalModelRequestFields) {
+    payload.additionalModelRequestFields = additionalModelRequestFields;
+  }
 
   if (maxTokens || temperature !== undefined || topP !== undefined) {
     payload.inferenceConfig = {};

open-sse/utils/kiroSessionReplay.js

@@ -0,0 +1,125 @@
+import { MEMORY_CONFIG } from "../config/runtimeConfig.js";
+
+const sessionStartStore = new Map();
+const MAX_SESSION_STARTS = 5000;
+
+function clone(value) {
+  return value == null ? value : JSON.parse(JSON.stringify(value));
+}
+
+function sessionKey(connectionId, conversationId) {
+  return `${connectionId || ""}:${conversationId || ""}`;
+}
+
+function ensureUserMessageModelId(message, modelId) {
+  if (message?.userInputMessage && !message.userInputMessage.modelId && modelId) {
+    message.userInputMessage.modelId = modelId;
+  }
+  return message;
+}
+
+function ensureHistoryModelIds(history, modelId) {
+  for (const item of history || []) {
+    ensureUserMessageModelId(item, modelId);
+  }
+  return history;
+}
+
+function prefixUserMessage(message, contentPrefix, modelId) {
+  const out = clone(message) || { userInputMessage: { content: "" } };
+  if (!out.userInputMessage) out.userInputMessage = { content: "" };
+  ensureUserMessageModelId(out, modelId);
+  if (contentPrefix) {
+    const content = out.userInputMessage.content || "";
+    out.userInputMessage.content = content
+      ? `${contentPrefix}\n\n${content}`
+      : contentPrefix;
+  }
+  return out;
+}
+
+function findFirstUserIndex(history) {
+  return history.findIndex((item) => item?.userInputMessage);
+}
+
+function rememberSessionStart(key, entry) {
+  if (sessionStartStore.size >= MAX_SESSION_STARTS) {
+    sessionStartStore.delete(sessionStartStore.keys().next().value);
+  }
+  sessionStartStore.set(key, { ...entry, lastUsed: Date.now() });
+}
+
+/**
+ * Preserve Kiro cacheability by freezing the first user message (`msg0`) for a
+ * session, replaying that exact message as the first history user on later
+ * turns, and injecting volatile current-time context only into the current turn.
+ */
+export function applyKiroSessionReplay({
+  conversationId,
+  connectionId,
+  modelId,
+  systemPrompt = "",
+  contentPrefix = "",
+  currentContentPrefix = "",
+  history = [],
+  currentMessage,
+} = {}) {
+  const key = sessionKey(connectionId, conversationId);
+  const existing = conversationId ? sessionStartStore.get(key) : null;
+  const baseHistory = clone(history) || [];
+  const baseCurrent = clone(currentMessage) || { userInputMessage: { content: "" } };
+
+  if (existing && existing.modelId === modelId && existing.systemPrompt === systemPrompt) {
+    existing.lastUsed = Date.now();
+    const firstUserIndex = findFirstUserIndex(baseHistory);
+    const sessionStart = ensureUserMessageModelId(clone(existing.sessionStart), modelId);
+    if (firstUserIndex >= 0) {
+      baseHistory[firstUserIndex] = sessionStart;
+    } else {
+      baseHistory.unshift(sessionStart);
+    }
+    return {
+      history: ensureHistoryModelIds(baseHistory, modelId),
+      currentMessage: prefixUserMessage(baseCurrent, currentContentPrefix, modelId),
+      replayed: true,
+    };
+  }
+
+  const firstUserIndex = findFirstUserIndex(baseHistory);
+  let sessionStart;
+  let nextCurrent = ensureUserMessageModelId(baseCurrent, modelId);
+  if (firstUserIndex >= 0) {
+    sessionStart = prefixUserMessage(baseHistory[firstUserIndex], contentPrefix, modelId);
+    baseHistory[firstUserIndex] = clone(sessionStart);
+    nextCurrent = prefixUserMessage(baseCurrent, currentContentPrefix, modelId);
+  } else {
+    sessionStart = prefixUserMessage(baseCurrent, contentPrefix, modelId);
+    nextCurrent = clone(sessionStart);
+  }
+
+  if (conversationId) {
+    rememberSessionStart(key, {
+      sessionStart: clone(sessionStart),
+      modelId,
+      systemPrompt,
+    });
+  }
+
+  return {
+    history: ensureHistoryModelIds(baseHistory, modelId),
+    currentMessage: nextCurrent,
+    replayed: false,
+  };
+}
+
+export function clearKiroSessionReplayStore() {
+  sessionStartStore.clear();
+}
+
+const cleanup = setInterval(() => {
+  const now = Date.now();
+  for (const [key, entry] of sessionStartStore) {
+    if (now - entry.lastUsed > MEMORY_CONFIG.sessionTtlMs) sessionStartStore.delete(key);
+  }
+}, MEMORY_CONFIG.sessionCleanupIntervalMs);
+if (cleanup.unref) cleanup.unref();

open-sse/utils/sessionManager.js

@@ -13,6 +13,7 @@ import { MEMORY_CONFIG } from "../config/runtimeConfig.js";
 
 // Runtime storage: Key = connectionId, Value = { sessionId, lastUsed }
 const runtimeSessionStore = new Map();
+const continuationStore = new Map();
 
 // Periodically evict entries that haven't been used within TTL
 const cleanupInterval = setInterval(() => {
@@ -80,6 +81,7 @@ export function generateBinaryStyleId() {
 export function clearSessionStore() {
     runtimeSessionStore.clear();
     assistantSessionStore.clear();
+    continuationStore.clear();
 }
 
 // Conversation-stable session store: Key = hash(scope+assistant text), Value = { sessionId, lastUsed }
@@ -87,9 +89,10 @@ const assistantSessionStore = new Map();
 const ASSISTANT_MIN_LEN = 50;
 const ASSISTANT_CAP_LEN = 50;
 const MAX_ASSISTANT_SESSIONS = 5000;
+const MAX_CONTINUATION_SESSIONS = 5000;
 
 // Client headers/body fields that carry an upstream session id (priority order)
-const SESSION_HEADER_KEYS = ["x-session-id", "session-id", "session_id", "x-amp-thread-id", "x-client-request-id"];
+const SESSION_HEADER_KEYS = ["x-session-id", "session-id", "session_id", "x-amp-thread-id"];
 const CLAUDE_CODE_SESSION_RE = /_session_([a-f0-9-]+)$/;
 
 function sha16(text) {
@@ -131,7 +134,7 @@ function extractAntigravitySession(body) {
     return m ? normalizeSessionId(m[1]) : null;
 }
 
-function extractClientSessionId(headers, body) {
+function extractClientSessionId(headers, body, scope = "") {
     const claude = extractClaudeCodeSession(body?.metadata?.user_id);
     if (claude) return `claude:${claude}`;
     const antigravity = extractAntigravitySession(body);
@@ -140,18 +143,25 @@ function extractClientSessionId(headers, body) {
         const v = headerValue(headers, key);
         if (v) return v;
     }
+    const requestId = scope === "kiro" ? null : headerValue(headers, "x-client-request-id");
+    if (requestId) return requestId;
     const fromBody =
         normalizeSessionId(body?.prompt_cache_key) ||
         normalizeSessionId(body?.session_id) ||
         normalizeSessionId(body?.conversation_id) ||
-        normalizeSessionId(body?.metadata?.user_id);
+        (scope === "kiro" ? null : normalizeSessionId(body?.metadata?.user_id));
     return fromBody || null;
 }
 
+function requestMessages(body) {
+    if (Array.isArray(body?.messages)) return body.messages;
+    if (Array.isArray(body?.input)) return body.input;
+    return [];
+}
+
 // Accumulate assistant text from OpenAI/Responses-style input/messages (cap-limited)
 function accumulateAssistantText(body) {
-    const items = Array.isArray(body?.input) ? body.input
+    const items = requestMessages(body);
-        : Array.isArray(body?.messages) ? body.messages : null;
     if (!items) return "";
     let text = "";
     for (const item of items) {
@@ -193,16 +203,39 @@ function assistantTextSessionId(scope, body) {
  * @param {string} [opts.connectionId] - Connection identifier (fallback scope)
  * @param {string} [opts.workspaceId] - Provider workspace id (account-wide fallback)
  * @param {string} [opts.scope] - Provider scope to isolate cache keys across providers
- * @returns {string} A stable session id
+ * @returns {{sessionId: string, ephemeral: boolean}} A session id plus whether it is one-shot
  */
-export function resolveSessionId({ headers, body, connectionId, workspaceId, scope = "" } = {}) {
+export function resolveSessionIdentity({ headers, body, connectionId, workspaceId, scope = "" } = {}) {
-    const client = extractClientSessionId(headers, body);
+    const client = extractClientSessionId(headers, body, scope);
-    if (client) return client;
+    if (client) return { sessionId: client, ephemeral: false };
-    const fromAssistant = assistantTextSessionId(`${scope}:${connectionId || ""}`, body);
+    const fromAssistant = scope === "kiro" ? null : assistantTextSessionId(`${scope}:${connectionId || ""}`, body);
-    if (fromAssistant) return fromAssistant;
+    if (fromAssistant) return { sessionId: fromAssistant, ephemeral: false };
     const ws = normalizeSessionId(workspaceId);
-    if (ws) return ws;
+    if (ws) return { sessionId: ws, ephemeral: false };
-    return deriveSessionId(connectionId);
+    if (scope === "kiro") return { sessionId: generateBinaryStyleId(), ephemeral: true };
+    return { sessionId: deriveSessionId(connectionId), ephemeral: false };
+}
+
+export function resolveSessionId(opts = {}) {
+    return resolveSessionIdentity(opts).sessionId;
+}
+
+export function resolveContinuationId({ sessionId, connectionId, scope = "", ephemeral = false } = {}) {
+    if (ephemeral) return crypto.randomUUID();
+    const key = `${scope}:${connectionId || ""}:${sessionId || ""}`;
+    const existing = continuationStore.get(key);
+    if (existing) {
+        existing.lastUsed = Date.now();
+        continuationStore.delete(key);
+        continuationStore.set(key, existing);
+        return existing.continuationId;
+    }
+    const continuationId = crypto.randomUUID();
+    if (continuationStore.size >= MAX_CONTINUATION_SESSIONS) {
+        continuationStore.delete(continuationStore.keys().next().value);
+    }
+    continuationStore.set(key, { continuationId, lastUsed: Date.now() });
+    return continuationId;
 }
 
 // Capture session id from request body + credentials (envelope still intact here)
@@ -227,5 +260,8 @@ const assistantCleanup = setInterval(() => {
     for (const [key, entry] of assistantSessionStore) {
         if (now - entry.lastUsed > MEMORY_CONFIG.sessionTtlMs) assistantSessionStore.delete(key);
     }
+    for (const [key, entry] of continuationStore) {
+        if (now - entry.lastUsed > MEMORY_CONFIG.sessionTtlMs) continuationStore.delete(key);
+    }
 }, MEMORY_CONFIG.sessionCleanupIntervalMs);
 if (assistantCleanup.unref) assistantCleanup.unref();

tests/translator/claude-kiro-direct.test.js

@@ -6,8 +6,8 @@ import "./registerAll.js";
 import { translateRequest, translateResponse } from "../../open-sse/translator/index.js";
 import { FORMATS } from "../../open-sse/translator/formats.js";
 
-const C2K = (body) =>
+const C2K = (body, credentials = null, model = "claude-sonnet-4.5") =>
-  translateRequest(FORMATS.CLAUDE, FORMATS.KIRO, "claude-sonnet-4.5", body, true, null, "kiro");
+  translateRequest(FORMATS.CLAUDE, FORMATS.KIRO, model, body, true, credentials, "kiro");
 
 describe("Claude → Kiro (direct route)", () => {
   it("produces a Kiro conversationState payload", () => {
@@ -16,6 +16,27 @@ describe("Claude → Kiro (direct route)", () => {
     expect(out.conversationState.currentMessage.userInputMessage.content).toContain("hello");
   });
 
+  it("keeps conversationId stable from client session headers and replays frozen msg0", () => {
+    const credentials = {
+      rawHeaders: { "x-session-id": "hermes-session-123-claude-replay" },
+      connectionId: "kiro-account-1",
+    };
+    const first = C2K({ messages: [{ role: "user", content: "first" }] }, credentials);
+    const second = C2K({ messages: [{ role: "user", content: "second" }] }, credentials);
+
+    expect(first.conversationState.conversationId).toBe("hermes-session-123-claude-replay");
+    expect(second.conversationState.conversationId).toBe("hermes-session-123-claude-replay");
+    expect(first.conversationState.agentContinuationId).toBeTruthy();
+    expect(second.conversationState.agentContinuationId).toBe(first.conversationState.agentContinuationId);
+    expect(first.conversationState.agentTaskType).toBe("vibe");
+    expect(second.conversationState.history[0].userInputMessage.content).toBe(
+      first.conversationState.currentMessage.userInputMessage.content
+    );
+    expect(second.conversationState.history[0].userInputMessage.modelId).toBe("claude-sonnet-4.5");
+    expect(second.conversationState.currentMessage.userInputMessage.content).toContain("Current time");
+    expect(second.conversationState.currentMessage.userInputMessage.content).toContain("second");
+  });
+
   it("guard 1: with no tools, a dangling tool_result is flattened to text (no structured ref)", () => {
     // Client omitted `tools` but kept a tool_result after compaction.
     const out = C2K({
@@ -60,20 +81,60 @@ describe("Claude → Kiro (direct route)", () => {
       null,
       "kiro"
     );
-    expect(out.conversationState.currentMessage.userInputMessage.content).toContain(
+    expect(out.systemPrompt).toContain(
       "<thinking_mode>enabled</thinking_mode>"
     );
+    expect(out.agentMode).toBe("vibe");
   });
 
-  it("maps output_config.effort high to Kiro max_thinking_length 24576", () => {
+  it("does not send additionalModelRequestFields for Kiro models without effort support", () => {
     const out = C2K({
       output_config: { effort: "high" },
       messages: [{ role: "user", content: "think with adaptive effort" }],
     });
 
-    expect(out.conversationState.currentMessage.userInputMessage.content).toContain(
+    expect(out.additionalModelRequestFields).toBeUndefined();
-      "<max_thinking_length>24576</max_thinking_length>"
+    expect(out.thinking).toBeUndefined();
-    );
+    expect(out.systemPrompt).toContain("<max_thinking_length>24576</max_thinking_length>");
+  });
+
+  it("maps output_config.effort high to Kiro CLI-style additionalModelRequestFields for effort models", () => {
+    const out = C2K({
+      output_config: { effort: "high" },
+      messages: [{ role: "user", content: "think with adaptive effort" }],
+    }, null, "claude-sonnet-5");
+
+    expect(out.additionalModelRequestFields).toEqual({
+      thinking: { type: "adaptive", display: "summarized" },
+      output_config: { effort: "high" },
+    });
+    expect(out.thinking).toBeUndefined();
+    expect(out.systemPrompt).toContain("<max_thinking_length>24576</max_thinking_length>");
+  });
+
+  it("sends Claude system as top-level systemPrompt and keeps a user-content fallback", () => {
+    const out = C2K({
+      system: "system-only instruction",
+      messages: [{ role: "user", content: "hello" }],
+    });
+
+    expect(out.systemPrompt).toContain("system-only instruction");
+    expect(out.conversationState.currentMessage.userInputMessage.content).toContain("system-only instruction");
+  });
+
+  it("keeps top-level systemPrompt stable across turns", () => {
+    const first = C2K({
+      system: "stable instruction",
+      messages: [{ role: "user", content: "first" }],
+    });
+    const second = C2K({
+      system: "stable instruction",
+      messages: [{ role: "user", content: "second" }],
+    });
+
+    expect(first.systemPrompt).toBe(second.systemPrompt);
+    expect(first.systemPrompt).not.toContain("Current time");
+    expect(first.conversationState.currentMessage.userInputMessage.content).toContain("Current time");
   });
 });
 

tests/unit/kiro-thinking-strip.test.js

@@ -1,5 +1,6 @@
 import { describe, it, expect } from "vitest";
 import { KiroExecutor } from "../../open-sse/executors/kiro.js";
+import "../translator/registerAll.js";
 
 function createMockFrame(eventType, payloadObj) {
   const payloadStr = JSON.stringify(payloadObj);
@@ -47,6 +48,13 @@ async function readAllSSE(stream) {
   return result;
 }
 
+async function readNextWithTimeout(reader) {
+  return Promise.race([
+    reader.read(),
+    new Promise((_, reject) => setTimeout(() => reject(new Error("timed out waiting for SSE chunk")), 100)),
+  ]);
+}
+
 describe("KiroExecutor thinking tag stripping", () => {
   it("strips <thinking> tags from assistantResponseEvent", async () => {
     const executor = new KiroExecutor();
@@ -121,4 +129,53 @@ describe("KiroExecutor thinking tag stripping", () => {
     const contentChunks = objects.filter(obj => obj.choices[0].delta.content !== undefined);
     expect(contentChunks.length).toBe(0);
   });
+
+  it("emits a terminal chunk at messageStop before the upstream stream closes", async () => {
+    const executor = new KiroExecutor();
+
+    const f1 = createMockFrame("assistantResponseEvent", { content: "OK" });
+    const f2 = createMockFrame("messageStopEvent", {});
+
+    const readableStream = new ReadableStream({
+      start(controller) {
+        controller.enqueue(f1);
+        controller.enqueue(f2);
+      }
+    });
+
+    const transformedResponse = executor.transformEventStreamToSSE({ body: readableStream }, "claude-test");
+    const reader = transformedResponse.body.getReader();
+    const decoder = new TextDecoder();
+    let output = "";
+    for (let i = 0; i < 4 && !output.includes("\"finish_reason\":\"stop\""); i++) {
+      const { value } = await readNextWithTimeout(reader);
+      output += decoder.decode(value, { stream: true });
+    }
+    await reader.cancel();
+
+    expect(output).toContain("\"finish_reason\":\"stop\"");
+  });
+
+  it("uses tool_calls finish reason for tool streams without messageStop", async () => {
+    const executor = new KiroExecutor();
+
+    const f1 = createMockFrame("toolUseEvent", { toolUseId: "tool-1", name: "read_file", input: { path: "a.txt" } });
+
+    const readableStream = new ReadableStream({
+      start(controller) {
+        controller.enqueue(f1);
+        controller.close();
+      }
+    });
+
+    const transformedResponse = executor.transformEventStreamToSSE({ body: readableStream }, "claude-test");
+    const output = await readAllSSE(transformedResponse.body);
+    const objects = output
+      .split("\n")
+      .filter(line => line.startsWith("data: ") && !line.includes("[DONE]"))
+      .map(line => JSON.parse(line.slice(6)));
+
+    const finalChunk = objects.at(-1);
+    expect(finalChunk.choices[0].finish_reason).toBe("tool_calls");
+  });
 });

tests/unit/openai-to-kiro.test.js

@@ -11,6 +11,7 @@ import { openaiToKiroRequest } from "../../open-sse/translator/request/openai-to
 
 const contentOf = (result) =>
   result.conversationState.currentMessage.userInputMessage.content;
+const systemPromptOf = (result) => result.systemPrompt || "";
 
 describe("openaiToKiroRequest", () => {
   describe("basic message conversion", () => {
@@ -293,7 +294,11 @@ describe("openaiToKiroRequest", () => {
 
       const result = openaiToKiroRequest("claude-sonnet-4.6", body, true, {});
 
-      expect(contentOf(result)).toContain("<max_thinking_length>1024</max_thinking_length>");
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>1024</max_thinking_length>");
+      expect(result.additionalModelRequestFields).toEqual({
+        thinking: { type: "adaptive", display: "summarized" },
+        output_config: { effort: "low" },
+      });
     });
 
     it("maps reasoning_effort high to max_thinking_length 24576", () => {
@@ -304,7 +309,97 @@ describe("openaiToKiroRequest", () => {
 
       const result = openaiToKiroRequest("claude-sonnet-4.6", body, true, {});
 
-      expect(contentOf(result)).toContain("<max_thinking_length>24576</max_thinking_length>");
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>24576</max_thinking_length>");
+      expect(result.additionalModelRequestFields).toEqual({
+        thinking: { type: "adaptive", display: "summarized" },
+        output_config: { effort: "high" },
+      });
+    });
+
+    it("does not send additionalModelRequestFields for legacy Kiro model ids", () => {
+      const body = {
+        reasoning_effort: "high",
+        messages: [{ role: "user", content: "Legacy model id should not get adaptive fields" }]
+      };
+
+      const result = openaiToKiroRequest("claude-sonnet-4.5", body, true, {});
+
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>24576</max_thinking_length>");
+      expect(result.additionalModelRequestFields).toBeUndefined();
+    });
+
+    it("does not send additionalModelRequestFields for date-suffixed Claude 4 model ids", () => {
+      const body = {
+        reasoning_effort: "high",
+        messages: [{ role: "user", content: "Date-suffixed Claude 4 should stay legacy" }]
+      };
+
+      const result = openaiToKiroRequest("claude-sonnet-4-20250514", body, true, {});
+
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>24576</max_thinking_length>");
+      expect(result.additionalModelRequestFields).toBeUndefined();
+    });
+
+    it("does not send additionalModelRequestFields for pre-4 legacy Kiro model ids", () => {
+      const body = {
+        reasoning_effort: "high",
+        messages: [{ role: "user", content: "Older model id should not get adaptive fields" }]
+      };
+
+      const result = openaiToKiroRequest("claude-sonnet-3.7", body, true, {});
+
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>24576</max_thinking_length>");
+      expect(result.additionalModelRequestFields).toBeUndefined();
+    });
+
+    it("does not send additionalModelRequestFields for prefixed pre-4 legacy Kiro model ids", () => {
+      const body = {
+        reasoning_effort: "high",
+        messages: [{ role: "user", content: "Prefixed older model id should not get adaptive fields" }]
+      };
+
+      const result = openaiToKiroRequest("kiro/claude-3-7-sonnet-20250219", body, true, {});
+
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>24576</max_thinking_length>");
+      expect(result.additionalModelRequestFields).toBeUndefined();
+    });
+
+    it("does not send Claude-specific additionalModelRequestFields for prefixed non-Claude aliases", () => {
+      const body = {
+        reasoning_effort: "high",
+        messages: [{ role: "user", content: "Prefixed non-Claude alias should not get adaptive fields" }]
+      };
+
+      const result = openaiToKiroRequest("kiro/gpt-4o", body, true, {});
+
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>24576</max_thinking_length>");
+      expect(result.additionalModelRequestFields).toBeUndefined();
+    });
+
+    it("does not send Claude-specific additionalModelRequestFields for non-Claude aliases", () => {
+      const body = {
+        reasoning_effort: "high",
+        messages: [{ role: "user", content: "Non-Claude aliases should not get Claude adaptive fields" }]
+      };
+
+      const result = openaiToKiroRequest("gpt-4o", body, true, {});
+
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>24576</max_thinking_length>");
+      expect(result.additionalModelRequestFields).toBeUndefined();
+    });
+
+    it("defaults future Kiro model ids to additionalModelRequestFields support", () => {
+      const body = {
+        reasoning_effort: "high",
+        messages: [{ role: "user", content: "Future model id should get adaptive fields" }]
+      };
+
+      const result = openaiToKiroRequest("claude-sonnet-4.60", body, true, {});
+
+      expect(result.additionalModelRequestFields).toEqual({
+        thinking: { type: "adaptive", display: "summarized" },
+        output_config: { effort: "high" },
+      });
     });
 
     it("clamps reasoning_effort max to Kiro max_thinking_length 32000", () => {
@@ -315,7 +410,8 @@ describe("openaiToKiroRequest", () => {
 
       const result = openaiToKiroRequest("claude-sonnet-4.6", body, true, {});
 
-      expect(contentOf(result)).toContain("<max_thinking_length>32000</max_thinking_length>");
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>32000</max_thinking_length>");
+      expect(result.additionalModelRequestFields?.output_config?.effort).toBe("high");
     });
 
     it("clamps OpenAI Responses reasoning.effort xhigh to max_thinking_length 32000", () => {
@@ -326,7 +422,8 @@ describe("openaiToKiroRequest", () => {
 
       const result = openaiToKiroRequest("claude-sonnet-4.6", body, true, {});
 
-      expect(contentOf(result)).toContain("<max_thinking_length>32000</max_thinking_length>");
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>32000</max_thinking_length>");
+      expect(result.additionalModelRequestFields?.output_config?.effort).toBe("high");
     });
 
     it("uses Claude thinking.budget_tokens as max_thinking_length", () => {
@@ -337,7 +434,7 @@ describe("openaiToKiroRequest", () => {
 
       const result = openaiToKiroRequest("claude-sonnet-4.6", body, true, {});
 
-      expect(contentOf(result)).toContain("<max_thinking_length>4096</max_thinking_length>");
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>4096</max_thinking_length>");
     });
 
     it("uses the default budget for synthetic -thinking models with no explicit config", () => {
@@ -347,7 +444,54 @@ describe("openaiToKiroRequest", () => {
 
       const result = openaiToKiroRequest("claude-sonnet-4.6-thinking", body, true, {});
 
-      expect(contentOf(result)).toContain("<max_thinking_length>16000</max_thinking_length>");
+      expect(systemPromptOf(result)).toContain("<max_thinking_length>16000</max_thinking_length>");
+    });
+
+    it("keeps top-level systemPrompt stable across turns", () => {
+      const first = openaiToKiroRequest(
+        "claude-sonnet-4.6-thinking",
+        { messages: [{ role: "user", content: "first" }] },
+        true,
+        {}
+      );
+      const second = openaiToKiroRequest(
+        "claude-sonnet-4.6-thinking",
+        { messages: [{ role: "user", content: "second" }] },
+        true,
+        {}
+      );
+
+      expect(first.systemPrompt).toBe(second.systemPrompt);
+      expect(first.systemPrompt).not.toContain("Current time");
+      expect(first.conversationState.currentMessage.userInputMessage.content).toContain("Current time");
+    });
+
+    it("replays frozen msg0 for explicit Kiro sessions while keeping current time fresh", () => {
+      const credentials = {
+        connectionId: "kiro-account-openai-replay",
+        rawHeaders: { "x-session-id": "hermes-session-openai-replay" },
+      };
+      const first = openaiToKiroRequest(
+        "claude-sonnet-4.6",
+        { messages: [{ role: "user", content: "first turn" }] },
+        true,
+        credentials
+      );
+      const second = openaiToKiroRequest(
+        "claude-sonnet-4.6",
+        { messages: [{ role: "user", content: "second turn" }] },
+        true,
+        credentials
+      );
+
+      expect(second.conversationState.conversationId).toBe("hermes-session-openai-replay");
+      expect(second.conversationState.agentContinuationId).toBe(first.conversationState.agentContinuationId);
+      expect(second.conversationState.history[0].userInputMessage.content).toBe(
+        first.conversationState.currentMessage.userInputMessage.content
+      );
+      expect(second.conversationState.history[0].userInputMessage.modelId).toBe("claude-sonnet-4.6");
+      expect(second.conversationState.currentMessage.userInputMessage.content).toContain("Current time");
+      expect(second.conversationState.currentMessage.userInputMessage.content).toContain("second turn");
     });
 
     it("does not inject thinking prefix for reasoning_effort none", () => {
@@ -358,8 +502,9 @@ describe("openaiToKiroRequest", () => {
 
       const result = openaiToKiroRequest("claude-sonnet-4.6", body, true, {});
 
-      expect(contentOf(result)).not.toContain("<thinking_mode>enabled</thinking_mode>");
+      expect(systemPromptOf(result)).not.toContain("<thinking_mode>enabled</thinking_mode>");
-      expect(contentOf(result)).not.toContain("<max_thinking_length>");
+      expect(systemPromptOf(result)).not.toContain("<max_thinking_length>");
+      expect(result.additionalModelRequestFields).toBeUndefined();
     });
   });
 });

tests/unit/session-manager.test.js

@@ -1,13 +1,15 @@
 // A2: locks resolveSessionId priority/stickiness (codex/kiro/antigravity centralization).
 import { describe, it, expect, beforeEach } from "vitest";
-import { resolveSessionId, deriveSessionId, clearSessionStore } from "../../open-sse/utils/sessionManager.js";
+import { resolveContinuationId, resolveSessionId, resolveSessionIdentity, deriveSessionId, clearSessionStore } from "../../open-sse/utils/sessionManager.js";
 
 // Assistant text must reach ASSISTANT_MIN_LEN (80) to use assistant anchor; else first user message.
 const longAssistant = "x".repeat(80);
 const bodyWithAssistant = { messages: [{ role: "assistant", content: longAssistant }] };
 const bodyWithUserOnly = { messages: [{ role: "user", content: "hello from first user message anchor" }] };
 
-beforeEach(() => clearSessionStore());
+beforeEach(() => {
+  clearSessionStore();
+});
 
 describe("resolveSessionId", () => {
   it("stickiness: same body+connectionId+scope -> same id", () => {
@@ -55,8 +57,170 @@ describe("resolveSessionId", () => {
     expect(got).toBe("client-sess-123");
   });
 
+  it("does not treat request-scoped x-client-request-id as a session override", () => {
+    const first = resolveSessionId({
+      headers: { "x-client-request-id": "req-1" },
+      body: bodyWithUserOnly,
+      connectionId: "conn1",
+      scope: "kiro",
+    });
+    const second = resolveSessionId({
+      headers: { "x-client-request-id": "req-2" },
+      body: bodyWithUserOnly,
+      connectionId: "conn1",
+      scope: "kiro",
+    });
+
+    expect(first).not.toBe("req-1");
+    expect(second).not.toBe("req-2");
+    expect(first).not.toBe(second);
+  });
+
+  it("does not treat request-scoped previous_response_id as a Kiro session override", () => {
+    const first = resolveSessionId({
+      body: { ...bodyWithUserOnly, previous_response_id: "resp-1" },
+      connectionId: "conn1",
+      scope: "kiro",
+    });
+    const second = resolveSessionId({
+      body: { ...bodyWithUserOnly, previous_response_id: "resp-2" },
+      connectionId: "conn1",
+      scope: "kiro",
+    });
+
+    expect(first).not.toBe("resp-1");
+    expect(second).not.toBe("resp-2");
+    expect(first).not.toBe(second);
+  });
+
+  it("does not treat raw metadata.user_id as a Kiro conversation session", () => {
+    const first = resolveSessionId({
+      body: {
+        metadata: { user_id: "user-123" },
+        messages: [{ role: "user", content: "new chat about invoices" }],
+      },
+      connectionId: "conn1",
+      scope: "kiro",
+    });
+    const second = resolveSessionId({
+      body: {
+        metadata: { user_id: "user-123" },
+        messages: [{ role: "user", content: "unrelated new chat about refunds" }],
+      },
+      connectionId: "conn1",
+      scope: "kiro",
+    });
+
+    expect(first).not.toBe("user-123");
+    expect(second).not.toBe("user-123");
+    expect(first).not.toBe(second);
+  });
+
+  it("keeps Claude Code session_id metadata as a Kiro conversation session", () => {
+    const body = {
+      metadata: { user_id: JSON.stringify({ session_id: "claude-code-session-123" }) },
+      messages: [{ role: "user", content: "same Claude Code session" }],
+    };
+
+    expect(resolveSessionId({ body, connectionId: "conn1", scope: "kiro" })).toBe("claude:claude-code-session-123");
+  });
+
+  it("keeps raw metadata.user_id as a non-Kiro session fallback", () => {
+    const got = resolveSessionId({
+      body: {
+        metadata: { user_id: "user-123" },
+        messages: [{ role: "user", content: "non-Kiro provider" }],
+      },
+      connectionId: "conn1",
+      scope: "codex",
+    });
+
+    expect(got).toBe("user-123");
+  });
+
+  it("keeps x-client-request-id as a session override outside Kiro scope", () => {
+    const got = resolveSessionId({
+      headers: { "x-client-request-id": "req-1" },
+      body: bodyWithAssistant,
+      connectionId: "conn1",
+      scope: "codex",
+    });
+
+    expect(got).toBe("req-1");
+  });
+
+
   it("workspaceId path: empty body + workspaceId set -> normalized workspaceId", () => {
     const got = resolveSessionId({ body: {}, connectionId: "conn1", workspaceId: "ws-abc" });
     expect(got).toBe("ws-abc");
   });
+
+  it("uses fresh Kiro sessions for unrelated headerless requests on the same connection", () => {
+    const a = resolveSessionId({ body: bodyWithUserOnly, connectionId: "conn1", scope: "kiro" });
+    const b = resolveSessionId({ body: bodyWithUserOnly, connectionId: "conn1", scope: "kiro" });
+    expect(a).not.toBe(b);
+  });
+
+  it("marks generated headerless Kiro sessions as ephemeral", () => {
+    const generated = resolveSessionIdentity({ body: bodyWithUserOnly, connectionId: "conn1", scope: "kiro" });
+    const explicit = resolveSessionIdentity({
+      headers: { "x-session-id": "client-sess-123" },
+      body: bodyWithUserOnly,
+      connectionId: "conn1",
+      scope: "kiro",
+    });
+
+    expect(generated.ephemeral).toBe(true);
+    expect(explicit).toEqual({ sessionId: "client-sess-123", ephemeral: false });
+  });
+
+  it("does not switch Kiro headerless requests to assistant-text session ids mid-conversation", () => {
+    const withAssistant = { messages: [{ role: "user", content: "same user" }, { role: "assistant", content: "y".repeat(80) }] };
+    const a = resolveSessionId({ body: withAssistant, connectionId: "conn1", scope: "kiro" });
+    const b = resolveSessionId({ body: withAssistant, connectionId: "conn1", scope: "kiro" });
+    expect(a).not.toBe(b);
+  });
+});
+
+describe("resolveContinuationId", () => {
+  it("keeps continuation id stable for the same Kiro session", () => {
+    const opts = { sessionId: "kiro-session-1", connectionId: "conn1", scope: "kiro" };
+    expect(resolveContinuationId(opts)).toBe(resolveContinuationId(opts));
+  });
+
+  it("uses a different continuation id for a different Kiro session", () => {
+    const a = resolveContinuationId({ sessionId: "kiro-session-1", connectionId: "conn1", scope: "kiro" });
+    const b = resolveContinuationId({ sessionId: "kiro-session-2", connectionId: "conn1", scope: "kiro" });
+    expect(a).not.toBe(b);
+  });
+
+  it("does not evict a recently used continuation id when the store exceeds its cap", () => {
+    const first = resolveContinuationId({ sessionId: "kiro-session-0", connectionId: "conn1", scope: "kiro" });
+    for (let i = 1; i < 5000; i++) {
+      resolveContinuationId({ sessionId: `kiro-session-${i}`, connectionId: "conn1", scope: "kiro" });
+    }
+    expect(resolveContinuationId({ sessionId: "kiro-session-0", connectionId: "conn1", scope: "kiro" })).toBe(first);
+    resolveContinuationId({ sessionId: "kiro-session-5000", connectionId: "conn1", scope: "kiro" });
+
+    expect(resolveContinuationId({ sessionId: "kiro-session-0", connectionId: "conn1", scope: "kiro" })).toBe(first);
+  });
+
+  it("evicts old continuation ids when the store exceeds its cap", () => {
+    const first = resolveContinuationId({ sessionId: "kiro-session-0", connectionId: "conn1", scope: "kiro" });
+    for (let i = 1; i <= 5000; i++) {
+      resolveContinuationId({ sessionId: `kiro-session-${i}`, connectionId: "conn1", scope: "kiro" });
+    }
+
+    const afterEviction = resolveContinuationId({ sessionId: "kiro-session-0", connectionId: "conn1", scope: "kiro" });
+    expect(afterEviction).not.toBe(first);
+  });
+
+  it("does not let ephemeral Kiro continuations evict explicit session continuations", () => {
+    const stable = resolveContinuationId({ sessionId: "explicit-session", connectionId: "conn1", scope: "kiro" });
+    for (let i = 0; i <= 5000; i++) {
+      resolveContinuationId({ sessionId: `ephemeral-session-${i}`, connectionId: "conn1", scope: "kiro", ephemeral: true });
+    }
+
+    expect(resolveContinuationId({ sessionId: "explicit-session", connectionId: "conn1", scope: "kiro" })).toBe(stable);
+  });
 });