Update .changeset/anthropic-extended-thinking.md

Co-authored-by: Kazuhiro Sera <seratch@openai.com>

Author: AryanBagade

.changeset/anthropic-extended-thinking.md

@@ -2,13 +2,4 @@
 '@openai/agents-extensions': patch
 ---
 
-Add support for Anthropic extended thinking (reasoning) in the AI SDK adapter.
-
-This feature enables capturing and preserving Claude's extended thinking blocks when using the AI SDK adapter with Anthropic models. Key changes:
-
-- Non-streaming: Extract reasoning parts from AI SDK responses and output them as `ReasoningItem` objects before tool calls (required by Anthropic API)
-- Streaming: Handle `reasoning-start`, `reasoning-delta`, and `reasoning-end` events to accumulate thinking content
-- Signature preservation: Store Anthropic's thinking signature in `providerData` for multi-turn conversation support
-- Round-trip support: Existing `itemsToLanguageV2Messages` already passes `providerData` to `providerOptions`, enabling signatures to be sent back to the API
-
-Fixes #628
+feat(agents-extensions): #628 add Anthropic extended thinking support