From 9fc8926975a36c84b7ee1f835612e23d60564396 Mon Sep 17 00:00:00 2001
From: kshitij <82637225+kshitijk4poor@users.noreply.github.com>
Date: Tue, 4 Aug 2026 11:19:15 +0530
Subject: [PATCH] perf: reuse request_input_estimate instead of recomputing
 estimate_request_tokens_rough

The output-cap error handler already computes request_input_estimate at
line 4722 via estimate_request_tokens_rough(api_messages, tools=...).
The new compression block ~50 lines below was calling the same function
with the same inputs again. Reuse the existing local.
---
 agent/conversation_loop.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/agent/conversation_loop.py b/agent/conversation_loop.py
index 4f4e57f8bf1f3..8848b64df2cb2 100644
--- a/agent/conversation_loop.py
+++ b/agent/conversation_loop.py
@@ -4775,7 +4775,7 @@ def run_conversation(
                             _overflow_input = messages
                             messages, active_system_prompt = agent._compress_context(
                                 messages, system_message,
-                                approx_tokens=estimate_request_tokens_rough(api_messages, tools=agent.tools or None),
+                                approx_tokens=request_input_estimate,
                                 task_id=effective_task_id,
                             )
                             if messages is _overflow_input and compression_skipped_due_to_lock(agent):