From 536c8a840bdade67be7694737df809861a417ba6 Mon Sep 17 00:00:00 2001
From: Vincent Koc <vincentkoc@ieee.org>
Date: Mon, 22 Jun 2026 13:29:23 +0800
Subject: [PATCH] refactor(llm): share cache retention resolution

---
 src/llm/providers/anthropic.ts          | 15 +--------------
 src/llm/providers/cache-retention.ts    | 15 +++++++++++++++
 src/llm/providers/openai-completions.ts | 11 +----------
 src/llm/providers/openai-responses.ts   | 15 +--------------
 4 files changed, 18 insertions(+), 38 deletions(-)
 create mode 100644 src/llm/providers/cache-retention.ts

diff --git a/src/llm/providers/anthropic.ts b/src/llm/providers/anthropic.ts
index 8ef3ecaa0501..bbda64673559 100644
--- a/src/llm/providers/anthropic.ts
+++ b/src/llm/providers/anthropic.ts
@@ -61,6 +61,7 @@ import {
   ANTHROPIC_OMITTED_REASONING_TEXT,
   findActiveAnthropicToolTurnAssistantIndex,
 } from "./anthropic-thinking-replay.js";
+import { resolveCacheRetention } from "./cache-retention.js";
 import { resolveCloudflareBaseUrl } from "./cloudflare.js";
 import { buildCopilotDynamicHeaders, hasCopilotVisionInput } from "./github-copilot-headers.js";
 import { adjustMaxTokensForThinking, buildBaseOptions } from "./simple-options.js";
@@ -68,20 +69,6 @@ import { transformMessages } from "./transform-messages.js";
 
 const ANTHROPIC_CACHE_CONTROL_LIMIT = 4;
 
-/**
- * Resolve cache retention preference.
- * Defaults to "short" and uses OPENCLAW_CACHE_RETENTION for backward compatibility.
- */
-function resolveCacheRetention(cacheRetention?: CacheRetention): CacheRetention {
-  if (cacheRetention) {
-    return cacheRetention;
-  }
-  if (typeof process !== "undefined" && process.env.OPENCLAW_CACHE_RETENTION === "long") {
-    return "long";
-  }
-  return "short";
-}
-
 function getCacheControl(
   model: Model<"anthropic-messages">,
   cacheRetention?: CacheRetention,
diff --git a/src/llm/providers/cache-retention.ts b/src/llm/providers/cache-retention.ts
new file mode 100644
index 000000000000..9953895eba55
--- /dev/null
+++ b/src/llm/providers/cache-retention.ts
@@ -0,0 +1,15 @@
+import type { CacheRetention } from "../types.js";
+
+/**
+ * Resolve cache retention preference.
+ * Defaults to "short" and uses OPENCLAW_CACHE_RETENTION for backward compatibility.
+ */
+export function resolveCacheRetention(cacheRetention?: CacheRetention): CacheRetention {
+  if (cacheRetention) {
+    return cacheRetention;
+  }
+  if (typeof process !== "undefined" && process.env.OPENCLAW_CACHE_RETENTION === "long") {
+    return "long";
+  }
+  return "short";
+}
diff --git a/src/llm/providers/openai-completions.ts b/src/llm/providers/openai-completions.ts
index 0fa4d9b3ea06..b4e0e3ece1c3 100644
--- a/src/llm/providers/openai-completions.ts
+++ b/src/llm/providers/openai-completions.ts
@@ -46,6 +46,7 @@ import { AssistantMessageEventStream } from "../utils/event-stream.js";
 import { headersToRecord } from "../utils/headers.js";
 import { parseStreamingJson } from "../utils/json-parse.js";
 import { sanitizeSurrogates } from "../utils/sanitize-unicode.js";
+import { resolveCacheRetention } from "./cache-retention.js";
 import { isCloudflareProvider, resolveCloudflareBaseUrl } from "./cloudflare.js";
 import { buildCopilotDynamicHeaders, hasCopilotVisionInput } from "./github-copilot-headers.js";
 import { clampOpenAIPromptCacheKey } from "./openai-prompt-cache.js";
@@ -118,16 +119,6 @@ type ChatCompletionToolWithCacheControl = OpenAI.Chat.Completions.ChatCompletion
   cache_control?: OpenAICompatCacheControl;
 };
 
-function resolveCacheRetention(cacheRetention?: CacheRetention): CacheRetention {
-  if (cacheRetention) {
-    return cacheRetention;
-  }
-  if (typeof process !== "undefined" && process.env.OPENCLAW_CACHE_RETENTION === "long") {
-    return "long";
-  }
-  return "short";
-}
-
 export const streamOpenAICompletions: StreamFunction<
   "openai-completions",
   OpenAICompletionsOptions
diff --git a/src/llm/providers/openai-responses.ts b/src/llm/providers/openai-responses.ts
index a49db54c6cdf..c11b0167483b 100644
--- a/src/llm/providers/openai-responses.ts
+++ b/src/llm/providers/openai-responses.ts
@@ -13,6 +13,7 @@ import type {
   Usage,
 } from "../types.js";
 import { AssistantMessageEventStream } from "../utils/event-stream.js";
+import { resolveCacheRetention } from "./cache-retention.js";
 import { isCloudflareProvider, resolveCloudflareBaseUrl } from "./cloudflare.js";
 import { buildCopilotDynamicHeaders, hasCopilotVisionInput } from "./github-copilot-headers.js";
 import { clampOpenAIPromptCacheKey } from "./openai-prompt-cache.js";
@@ -27,20 +28,6 @@ import { buildBaseOptions } from "./simple-options.js";
 
 const OPENAI_TOOL_CALL_PROVIDERS = new Set(["openai", "opencode"]);
 
-/**
- * Resolve cache retention preference.
- * Defaults to "short" and uses OPENCLAW_CACHE_RETENTION for backward compatibility.
- */
-function resolveCacheRetention(cacheRetention?: CacheRetention): CacheRetention {
-  if (cacheRetention) {
-    return cacheRetention;
-  }
-  if (typeof process !== "undefined" && process.env.OPENCLAW_CACHE_RETENTION === "long") {
-    return "long";
-  }
-  return "short";
-}
-
 function getCompat(model: Model<"openai-responses">): Required<OpenAIResponsesCompat> {
   return {
     sendSessionIdHeader: model.compat?.sendSessionIdHeader ?? true,