Fix : Add reasoning_content placeholder for DeepSeek thinking models

open-sse/utils/reasoningContentInjector.js

@@ -0,0 +1,37 @@
+// Some thinking-mode providers (DeepSeek, Kimi, ...) require reasoning_content
+// to be echoed back on assistant messages. Clients in OpenAI format don't send it,
+// so we inject a non-empty placeholder to satisfy upstream validation.
+
+const PLACEHOLDER = " ";
+
+// Provider-level rules: keyed by executor.provider
+const PROVIDER_RULES = {
+  deepseek: { scope: "all" }
+};
+
+// Model-level rules: matched by predicate against model id
+const MODEL_RULES = [
+  { match: m => m?.startsWith?.("kimi-"), scope: "toolCalls" },
+  { match: m => m?.startsWith?.("deepseek-"), scope: "all" }
+];
+
+function shouldInject(message, scope) {
+  if (message?.role !== "assistant" || "reasoning_content" in message) return false;
+  if (scope === "toolCalls") return Array.isArray(message.tool_calls);
+  return true;
+}
+
+function applyRule(body, rule) {
+  if (!rule || !body?.messages) return body;
+  const messages = body.messages.map(m =>
+    shouldInject(m, rule.scope) ? { ...m, reasoning_content: PLACEHOLDER } : m
+  );
+  return { ...body, messages };
+}
+
+export function injectReasoningContent({ provider, model, body }) {
+  const providerRule = PROVIDER_RULES[provider];
+  const modelRule = MODEL_RULES.find(r => r.match(model));
+  const rule = providerRule || modelRule;
+  return applyRule(body, rule);
+}