From d3398f826ae2654b52a7988830d533a354754c13 Mon Sep 17 00:00:00 2001
From: "robopipi[bot]" <336283572+robopipi[bot]@users.noreply.github.com>
Date: Tue, 27 Oct 2026 20:27:30 +0100
Subject: [PATCH] fix(ai): detect LiteLLM ContextWindowExceededError as context
 overflow (#11494)

LiteLLM's pre-call context check rejects requests that are too large
with HTTP 400 and a message like:

```
litellm.ContextWindowExceededError: litellm.BadRequestError: litellm._pre_call_checks: Context Window exceeded for given call. ... Max Input Tokens=262144, Got=265262 ...
```

None of the `OVERFLOW_PATTERNS` matched this wording, so
`isContextOverflow()` returned false. As a result, the agent session
never ran overflow compaction and retry, and the turn failed with the
400 error.

This change adds `/context window exceeded/i` to `OVERFLOW_PATTERNS` and
documents the error in the doc comment that lists providers.

Testing:
- Added a regression test in `packages/ai/test/overflow.test.ts` with
the exact error message from the issue. It fails without the pattern and
passes with it.
- `node ../../node_modules/vitest/dist/cli.js --run
test/overflow.test.ts` in `packages/ai`: 32 passed.
- `npm run check` passes.
- Running `isContextOverflow()` directly on the reported message
returned `false` before this change and returns `true` after it.

Fixes #11492

Requested by @rwachtler

Co-authored-by: triagent <triagent@localhost>
---
 packages/ai/src/utils/overflow.ts | 2 ++
 packages/ai/test/overflow.test.ts | 8 ++++++++
 2 files changed, 10 insertions(+)

diff --git a/packages/ai/src/utils/overflow.ts b/packages/ai/src/utils/overflow.ts
index 03f82f61b..36ae26b49 100644
--- a/packages/ai/src/utils/overflow.ts
+++ b/packages/ai/src/utils/overflow.ts
@@ -12,6 +12,7 @@ import type { AssistantMessage } from "../types.ts";
  * - Anthropic: "413 {\"error\":{\"type\":\"request_too_large\",\"message\":\"Request exceeds the maximum size\"}}"
  * - OpenAI: "Your input exceeds the context window of this model"
  * - OpenAI/LiteLLM: "Requested token count exceeds the model's maximum context length of 131072 tokens"
+ * - LiteLLM pre-call check: "litellm.ContextWindowExceededError: ... Context Window exceeded for given call."
  * - OpenAI-compatible: "Input length (265330) exceeds model's maximum context length (262144)."
  * - Google: "The input token count (1196265) exceeds the maximum number of tokens allowed (1048575)"
  * - xAI: "This model's maximum prompt length is 131072 but the request contains 537812 tokens"
@@ -42,6 +43,7 @@ const OVERFLOW_PATTERNS = [
 	/input is too long for requested model/i, // Amazon Bedrock
 	/exceeds the context window/i, // OpenAI (Completions & Responses API)
 	/exceeds (?:the )?(?:model'?s )?maximum context length(?: of [\d,]+ tokens?|\s*\([\d,]+\))/i, // OpenAI-compatible proxies (LiteLLM)
+	/context window exceeded/i, // LiteLLM pre-call check (ContextWindowExceededError)
 	/input token count.*exceeds the maximum/i, // Google (Gemini)
 	/maximum prompt length is \d+/i, // xAI (Grok)
 	/reduce the length of the messages/i, // Groq
diff --git a/packages/ai/test/overflow.test.ts b/packages/ai/test/overflow.test.ts
index e7ae90548..1cab5e92e 100644
--- a/packages/ai/test/overflow.test.ts
+++ b/packages/ai/test/overflow.test.ts
@@ -91,6 +91,14 @@ describe("isContextOverflow", () => {
 		expect(isContextOverflow(message, 131072)).toBe(true);
 	});
 
+	it("detects LiteLLM pre-call ContextWindowExceededError", () => {
+		// Regression for #11492.
+		const message = createErrorMessage(
+			'400: {"message":"litellm.ContextWindowExceededError: litellm.BadRequestError: litellm._pre_call_checks: Context Window exceeded for given call. No models have context window large enough for this call.\nModel=Qwen3.8-Flash-Next, Max Input Tokens=262144, Got=265262\nmodel=Qwen3.8-Flash-Next. context_window_fallbacks=None. fallbacks=[]."}',
+		);
+		expect(isContextOverflow(message, 262144)).toBe(true);
+	});
+
 	it("detects OpenAI-compatible parenthesized maximum context length errors", () => {
 		const message = createErrorMessage(
 			"Error: 400 Input length (265330) exceeds model's maximum context length (262144).",