Files
9router/open-sse/translator/request/openai-to-kiro.js
T
Edison42 9c58ba645e 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
2026-07-16 15:15:05 +07:00

637 lines
23 KiB
JavaScript

/**
* OpenAI to Kiro Request Translator
* Converts OpenAI Chat Completions format to Kiro/AWS CodeWhisperer format
*/
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 { parseDataUri } from "../concerns/image.js";
import { DEFAULT_IMAGE_MIME } from "../schema/index.js";
import { ROLE, OPENAI_BLOCK, CLAUDE_BLOCK } from "../schema/index.js";
/** Render a single tool call as a readable text line. */
function toolCallToText(name, input) {
let argStr;
try {
argStr = typeof input === "string" ? input : JSON.stringify(input ?? {});
} catch {
argStr = "{}";
}
return `[Tool call: ${name || "unknown"}(${argStr})]`;
}
/** Render a tool result (string or content-block array) as a text line. */
function toolResultToText(content) {
const text = Array.isArray(content)
? content.map(c => (typeof c === "string" ? c : c.text || "")).join("\n")
: (typeof content === "string" ? content : "");
return `[Tool result: ${text}]`;
}
/**
* Flatten all tool calls/results in a conversation into plain text.
*
* Kiro's schema validator requires a non-empty
* currentMessage.userInputMessageContext.tools array whenever the history
* references any tool use; otherwise it returns "Improperly formed request"
* (HTTP 400). A client can hit this by omitting the `tools` array on a
* follow-up request — typically after client-side compaction (e.g. OpenCode).
*
* Rather than fabricate stub tool specs — which would advertise tool-calling
* capability the client never requested and may not handle, risking a phantom
* tool call on an otherwise plain turn — we collapse the tool interaction into
* text. The request stays honest, and since no structured tool content
* remains, the validator's "tools required" rule never fires.
*
* Only invoked when the client did NOT send tools; when tools are present the
* structured form is preserved.
*/
function flattenToolInteractions(messages) {
const out = [];
for (const msg of messages) {
// OpenAI tool-result message → user text line
if (msg.role === ROLE.TOOL) {
out.push({ role: ROLE.USER, content: toolResultToText(msg.content) });
continue;
}
if (msg.role === ROLE.ASSISTANT) {
const parts = [];
if (Array.isArray(msg.content)) {
for (const c of msg.content) {
if (c.type === CLAUDE_BLOCK.TOOL_USE) {
parts.push(toolCallToText(c.name, c.input));
} else if (c.type === OPENAI_BLOCK.TEXT || c.text) {
parts.push(c.text || "");
}
}
} else if (typeof msg.content === "string") {
parts.push(msg.content);
}
for (const tc of msg.tool_calls || []) {
parts.push(toolCallToText(tc.function?.name, tc.function?.arguments));
}
out.push({ role: ROLE.ASSISTANT, content: parts.filter(Boolean).join("\n") });
continue;
}
// User messages: replace tool_result blocks with text, keep text + images.
if (msg.role === ROLE.USER && Array.isArray(msg.content)) {
const newContent = msg.content.map(c =>
c.type === CLAUDE_BLOCK.TOOL_RESULT
? { type: OPENAI_BLOCK.TEXT, text: toolResultToText(c.content) }
: c
);
out.push({ ...msg, content: newContent });
continue;
}
out.push(msg);
}
return out;
}
/**
* Reconcile orphaned toolResults — those whose toolUseId has no matching
* toolUse in any assistant message. This happens when client-side compaction
* truncates the conversation and removes the assistant message containing the
* tool_use, but keeps the user message with the corresponding tool_result.
*
* A dangling structured reference makes Kiro return 400, so it must be removed.
* But the client deliberately kept the result content through compaction, so
* rather than discard it we fold it back into the user message as text — the
* same shape flattenToolInteractions() produces. The 400 trigger (the
* structured reference) is gone; the content survives.
*
* `messages` is every carrier that can hold toolResults — both history items
* and the popped-out currentMessage (orphans can land on either).
*/
function reconcileOrphanedToolResults(history, currentMessage) {
// Phase 1: collect all valid toolUseIds from assistant messages in history.
// (currentMessage is always a user turn, so it carries no toolUses.)
const validIds = new Set();
for (const h of history) {
const arm = h.assistantResponseMessage;
if (!arm) continue;
for (const tu of arm.toolUses || []) {
if (tu.toolUseId) validIds.add(tu.toolUseId);
}
}
// Phase 2: across history + currentMessage, keep results with a matching
// toolUse and salvage the rest as text.
const carriers = currentMessage ? [...history, currentMessage] : history;
for (const item of carriers) {
const uim = item.userInputMessage;
const ctx = uim?.userInputMessageContext;
if (!ctx?.toolResults?.length) continue;
const kept = [];
const salvaged = [];
for (const tr of ctx.toolResults) {
if (validIds.has(tr.toolUseId)) {
kept.push(tr);
} else {
salvaged.push(toolResultToText(tr.content));
}
}
if (salvaged.length === 0) continue; // no orphans — leave untouched
// Fold orphaned result content into the user text so it is not lost
const extra = salvaged.join("\n");
uim.content = uim.content ? `${uim.content}\n\n${extra}` : extra;
ctx.toolResults = kept;
if (kept.length === 0 && !ctx.tools?.length) {
delete uim.userInputMessageContext;
}
}
}
/**
* Safely parse JSON string, returning fallback on failure.
*/
function safeJSONParse(str, fallback) {
if (typeof str !== "string") return str ?? fallback;
try { return JSON.parse(str); } catch { return fallback; }
}
/**
* Convert OpenAI messages to Kiro format
* Rules: system/tool/user -> user role, merge consecutive same roles.
*
* Returns { history, currentMessage }.
*/
function convertMessages(messages, tools, model) {
let history = [];
let currentMessage = null;
const clientProvidedTools = tools && tools.length > 0;
// When the client did not send tools, flatten any tool calls/results in the
// history into plain text (see flattenToolInteractions). This keeps the
// request honest and sidesteps Kiro's "tools required" 400, since no
// structured tool content survives to trigger it.
if (!clientProvidedTools) {
messages = flattenToolInteractions(messages);
}
let pendingUserContent = [];
let pendingAssistantContent = [];
let pendingToolResults = [];
let pendingImages = [];
let currentRole = null;
let toolsInjectedToFirstUserMsg = false;
const flushPending = () => {
if (currentRole === "user") {
const content = pendingUserContent.join("\n\n").trim() || "continue";
const userMsg = {
userInputMessage: {
content: content,
modelId: ""
}
};
// Attach images if present (Kiro API supports images field)
if (pendingImages.length > 0) {
userMsg.userInputMessage.images = pendingImages;
}
if (pendingToolResults.length > 0) {
userMsg.userInputMessage.userInputMessageContext = {
toolResults: pendingToolResults
};
}
// Add tools to the user message that has no preceding assistant messages,
// OR the first user message (whichever comes first after any opening
// assistant messages). We track whether any user message has already
// received tools via a flag on the history array.
if (clientProvidedTools && !toolsInjectedToFirstUserMsg) {
if (!userMsg.userInputMessage.userInputMessageContext) {
userMsg.userInputMessage.userInputMessageContext = {};
}
userMsg.userInputMessage.userInputMessageContext.tools = tools.map(t => {
const name = t.function?.name || t.name;
let description = t.function?.description || t.description || "";
if (!description.trim()) {
description = `Tool: ${name}`;
}
const schema = t.function?.parameters || t.parameters || t.input_schema || {};
// Normalize schema: Kiro requires required[] and proper type/properties
const normalizedSchema = Object.keys(schema).length === 0
? { type: "object", properties: {}, required: [] }
: { ...schema, required: schema.required ?? [] };
return {
toolSpecification: {
name,
description,
inputSchema: { json: normalizedSchema }
}
};
});
toolsInjectedToFirstUserMsg = true;
}
history.push(userMsg);
currentMessage = userMsg;
pendingUserContent = [];
pendingToolResults = [];
pendingImages = [];
} else if (currentRole === "assistant") {
const content = pendingAssistantContent.join("\n\n").trim() || "...";
const assistantMsg = {
assistantResponseMessage: {
content: content
}
};
history.push(assistantMsg);
pendingAssistantContent = [];
}
};
for (let i = 0; i < messages.length; i++) {
const msg = messages[i];
let role = msg.role;
// Normalize: system/tool -> user
const wasSystem = role === ROLE.SYSTEM;
if (role === ROLE.SYSTEM || role === ROLE.TOOL) {
role = ROLE.USER;
}
// If role changes, flush pending
if (role !== currentRole && currentRole !== null) {
flushPending();
}
currentRole = role;
if (role === ROLE.USER) {
// Extract content
let content = "";
if (typeof msg.content === "string") {
content = msg.content;
} else if (Array.isArray(msg.content)) {
const textParts = [];
for (const c of msg.content) {
if (c.type === OPENAI_BLOCK.TEXT || c.text) {
textParts.push(c.text || "");
} else if (c.type === OPENAI_BLOCK.IMAGE_URL) {
// OpenAI format: image_url.url with data URI
const url = c.image_url?.url || "";
const parsed = parseDataUri(url);
if (parsed) {
const format = parsed.mimeType.split("/")[1] || parsed.mimeType;
pendingImages.push({ format, source: { bytes: parsed.base64 } });
} else if (url.startsWith("http://") || url.startsWith("https://")) {
// Kiro only supports base64 — fallback to URL text
textParts.push(`[Image: ${url}]`);
}
} else if (c.type === CLAUDE_BLOCK.IMAGE) {
// Claude format: source.type = "base64", source.media_type, source.data
if (c.source?.type === "base64" && c.source?.data) {
const mediaType = c.source.media_type || DEFAULT_IMAGE_MIME;
const format = mediaType.split("/")[1] || mediaType;
pendingImages.push({ format, source: { bytes: c.source.data } });
}
}
}
content = textParts.join("\n");
// Check for tool_result blocks
const toolResultBlocks = msg.content.filter(c => c.type === CLAUDE_BLOCK.TOOL_RESULT);
if (toolResultBlocks.length > 0) {
toolResultBlocks.forEach(block => {
const text = Array.isArray(block.content)
? block.content.map(c => c.text || "").join("\n")
: (typeof block.content === "string" ? block.content : "");
pendingToolResults.push({
toolUseId: block.tool_use_id,
status: "success",
content: [{ text: text }]
});
});
}
}
// Handle tool role (from normalized)
if (msg.role === ROLE.TOOL) {
const toolContent = typeof msg.content === "string" ? msg.content : "";
pendingToolResults.push({
toolUseId: msg.tool_call_id,
status: "success",
content: [{ text: toolContent }]
});
} else if (content) {
// <instructions> tags: Claude models treat these as authoritative directives.
pendingUserContent.push(
wasSystem ? `<instructions>\n${content}\n</instructions>` : content
);
}
} else if (role === ROLE.ASSISTANT) {
// Extract text content and tool uses
let textContent = "";
let toolUses = [];
if (Array.isArray(msg.content)) {
const textBlocks = msg.content.filter(c => c.type === OPENAI_BLOCK.TEXT);
textContent = textBlocks.map(b => b.text).join("\n").trim();
const toolUseBlocks = msg.content.filter(c => c.type === CLAUDE_BLOCK.TOOL_USE);
toolUses = toolUseBlocks;
} else if (typeof msg.content === "string") {
textContent = msg.content.trim();
}
if (msg.tool_calls && msg.tool_calls.length > 0) {
toolUses = msg.tool_calls;
}
if (textContent) {
pendingAssistantContent.push(textContent);
}
// Store tool uses in last assistant message
if (toolUses.length > 0) {
// Flush to create assistant message with toolUses
flushPending();
const lastMsg = history[history.length - 1];
if (lastMsg?.assistantResponseMessage) {
lastMsg.assistantResponseMessage.toolUses = toolUses.map(tc => {
if (tc.function) {
return {
toolUseId: tc.id || uuidv4(),
name: tc.function.name,
input: safeJSONParse(tc.function.arguments, {})
};
} else {
return {
toolUseId: tc.id || uuidv4(),
name: tc.name,
input: tc.input || {}
};
}
});
}
currentRole = null;
}
}
}
// Flush remaining
if (currentRole !== null) {
flushPending();
}
// Pop last userInputMessage as currentMessage (search from end, skip trailing assistant messages)
for (let i = history.length - 1; i >= 0; i--) {
if (history[i].userInputMessage) {
currentMessage = history.splice(i, 1)[0];
break;
}
}
// Grab tools from first history item BEFORE cleanup removes them
const firstHistoryTools = history[0]?.userInputMessage?.userInputMessageContext?.tools;
// Clean up history for Kiro API compatibility
history.forEach(item => {
if (item.userInputMessage?.userInputMessageContext?.tools) {
delete item.userInputMessage.userInputMessageContext.tools;
}
if (item.userInputMessage?.userInputMessageContext &&
Object.keys(item.userInputMessage.userInputMessageContext).length === 0) {
delete item.userInputMessage.userInputMessageContext;
}
if (item.userInputMessage && !item.userInputMessage.modelId) {
item.userInputMessage.modelId = model;
}
});
// Merge consecutive user messages (Kiro requires alternating user/assistant)
// When merging, also combine userInputMessageContext fields so toolResults
// and images from the second message are not silently dropped.
const mergedHistory = [];
for (let i = 0; i < history.length; i++) {
const current = history[i];
if (current.userInputMessage &&
mergedHistory.length > 0 &&
mergedHistory[mergedHistory.length - 1].userInputMessage) {
const prev = mergedHistory[mergedHistory.length - 1];
prev.userInputMessage.content += "\n\n" + current.userInputMessage.content;
// Merge context: combine toolResults, images, etc.
const prevCtx = prev.userInputMessage.userInputMessageContext;
const curCtx = current.userInputMessage.userInputMessageContext;
if (curCtx) {
if (!prevCtx) {
prev.userInputMessage.userInputMessageContext = curCtx;
} else {
if (curCtx.toolResults?.length > 0) {
prevCtx.toolResults = [...(prevCtx.toolResults || []), ...curCtx.toolResults];
}
if (curCtx.tools?.length > 0) {
prevCtx.tools = [...(prevCtx.tools || []), ...curCtx.tools];
}
}
}
} else {
mergedHistory.push(current);
}
}
// When currentMessage is null (no user messages at all — edge case where
// input is only assistant messages), create a minimal currentMessage so
// tools and content can be injected.
if (!currentMessage) {
currentMessage = {
userInputMessage: {
content: "",
modelId: model,
}
};
}
// Reconcile orphaned toolResults across history AND currentMessage — when
// client-side compaction removes assistant messages containing tool_use but
// keeps the tool_result, the dangling reference triggers a Kiro 400. Fold the
// content back into the user text instead of discarding it. Run after
// currentMessage is finalized (an orphan can be merged into it) and before
// tool injection (which may re-add userInputMessageContext).
//
// Only needed on the tools-present path: when the client sent no tools,
// flattenToolInteractions already collapsed every toolResult to text, so
// there is nothing structured left to orphan.
if (clientProvidedTools) {
reconcileOrphanedToolResults(mergedHistory, currentMessage);
}
// Inject tools into currentMessage AFTER cleanup. Tools only exist here when
// the client explicitly sent them (otherwise flattenToolInteractions already
// collapsed all tool content to text upstream, so there is nothing to carry).
const resolvedTools = firstHistoryTools;
if (resolvedTools?.length > 0 &&
!currentMessage.userInputMessage.userInputMessageContext?.tools) {
if (!currentMessage.userInputMessage.userInputMessageContext) {
currentMessage.userInputMessage.userInputMessageContext = {};
}
currentMessage.userInputMessage.userInputMessageContext.tools = resolvedTools;
}
return { history: mergedHistory, currentMessage };
}
/**
* Build Kiro payload from OpenAI format
*
* Two 9router-specific behaviours implemented here:
*
* 1. `-agentic` model suffix. Synthetic variant — same upstream model, but we
* inject a chunked-write system prompt to keep large file writes under
* Kiro's 2-3 minute server timeout. The suffix is stripped before being
* sent upstream.
*
* 2. Thinking / reasoning. Kiro does not accept `thinking.type` or
* `reasoning_effort` natively. The only way to enable reasoning is to
* inject `<thinking_mode>enabled</thinking_mode>` into the user content
* sent upstream. Detection covers Anthropic-Beta header, Claude API
* `thinking`, OpenAI `reasoning_effort`, AMP/Cursor magic tags, and model
* name hints.
*/
export function openaiToKiroRequest(model, body, stream, credentials) {
const messages = body.messages || [];
const tools = body.tools || [];
const maxTokens = 32000;
const temperature = body.temperature;
const topP = body.top_p;
const { upstream: upstreamModel, agentic } = resolveKiroModel(model);
const thinkingBudget = resolveKiroThinkingBudget(body, credentials?.rawHeaders, model);
const { history, currentMessage } = convertMessages(messages, tools, upstreamModel);
// API-key (headless) auth uses a raw CodeWhisperer credential whose profile is
// account-specific. Injecting the shared builder-id/social *default* placeholder
// ARN makes CodeWhisperer reject the request with 403 "bearer token invalid"
// (the ARN doesn't belong to the key's account). So for api_key, only send a
// profileArn that was actually resolved for this connection — never the default.
// OAuth/social keep the default fallback (their tokens accept it).
// api_key / idc / external_idp carry an account-specific (or token-bound)
// profile. The shared builder-id/social default ARN belongs to a different
// account and triggers 403 "bearer token invalid", so never fall back to it —
// send the resolved ARN, or an empty string so CodeWhisperer uses the token's
// own default profile. Only OAuth/social keep the shared placeholder.
const authMethod = credentials?.providerSpecificData?.authMethod;
const accountBoundAuth =
authMethod === "api_key" || authMethod === "idc" || authMethod === "external_idp";
const profileArn = accountBoundAuth
? (credentials?.providerSpecificData?.profileArn || "")
: (credentials?.providerSpecificData?.profileArn || resolveDefaultProfileArn(authMethod));
const timestamp = new Date().toISOString();
// Kiro CLI/KAS sends these as top-level systemPrompt. Keep a content fallback
// too because the CodeWhisperer surface does not always enforce top-level
// systemPrompt for direct calls.
const systemPromptParts = [];
if (thinkingBudget !== null) {
systemPromptParts.push(buildThinkingSystemPrefix(thinkingBudget));
}
if (agentic) {
systemPromptParts.push(KIRO_AGENTIC_SYSTEM_PROMPT);
}
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,
agentContinuationId: continuationId,
agentTaskType: "vibe",
currentMessage: {
userInputMessage: {
content: replayCurrent.content || "",
modelId: upstreamModel,
origin: "AI_EDITOR",
...(replayCurrent.images?.length > 0 && {
images: replayCurrent.images
}),
...(replayCurrent.userInputMessageContext && {
userInputMessageContext: replayCurrent.userInputMessageContext
})
}
},
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 = {};
if (maxTokens) payload.inferenceConfig.maxTokens = maxTokens;
if (temperature !== undefined) payload.inferenceConfig.temperature = temperature;
if (topP !== undefined) payload.inferenceConfig.topP = topP;
}
// Tag payload so the executor can route the upstream model id correctly.
Object.defineProperty(payload, "_kiroUpstreamModel", {
value: upstreamModel,
enumerable: false
});
return payload;
}
register(FORMATS.OPENAI, FORMATS.KIRO, openaiToKiroRequest, null);