From 669eed6990942caa7d9a8f43996b6dcdfa553506 Mon Sep 17 00:00:00 2001
From: Thanh_Dang <ntdang23@clc.fitus.edu.vn>
Date: Thu, 25 Jun 2026 22:17:17 +0700
Subject: [PATCH] perf(agent): cache BuildFilteredTools result per run (#1282)

* fix(pipeline): defer non-tool messages until all tool results are emitted

Multi-tool responses could interleave synthetic user messages (loop
warnings, nudges) between tool results, breaking OpenAI-compatible
providers that require contiguous tool_result grouping.

Fix both sequential and parallel tool execution paths in tool_stage.go
to defer non-tool messages until after all tool results are appended.
Fix sanitizeHistory lookahead in loop_history_sanitize.go to reorder
interleaved non-tool messages in persisted session history.

Closes #1177

* perf(agent): cache BuildFilteredTools result per run

All filtering inputs (policy, disabled tools, bootstrap, channel,
orchestration mode, user MCP tools) are fixed for the lifetime of a
run. Only the final iteration differs (strips all tools). Cache the
tool definitions after the first call and reuse for iterations
0..maxIter-1, eliminating N-1 redundant 7-step policy evaluations per
agent run.
---
 internal/agent/loop_pipeline_callbacks.go | 29 +++++++++++++++++++----
 1 file changed, 25 insertions(+), 4 deletions(-)

diff --git a/internal/agent/loop_pipeline_callbacks.go b/internal/agent/loop_pipeline_callbacks.go
index 9c36b755..e2d2accc 100644
--- a/internal/agent/loop_pipeline_callbacks.go
+++ b/internal/agent/loop_pipeline_callbacks.go
@@ -199,7 +199,25 @@ func (l *Loop) makeInjectReminders(req *RunRequest) func(ctx context.Context, in
 }
 
 func (l *Loop) makeBuildFilteredTools(req *RunRequest) func(state *pipeline.RunState) ([]providers.ToolDefinition, error) {
+	// Per-run cache: all filtering inputs (policy, disabled tools, bootstrap,
+	// channel, orchestration mode, user MCP tools) are fixed for the lifetime
+	// of a run. Only the final iteration differs (strips all tools). Cache the
+	// result after the first call and reuse for iterations 0..maxIter-1.
+	var (
+		cachedToolDefs []providers.ToolDefinition
+		cacheValid     bool
+	)
 	return func(state *pipeline.RunState) ([]providers.ToolDefinition, error) {
+		maxIter := l.maxIterations
+		if req.MaxIterations > 0 && req.MaxIterations < maxIter {
+			maxIter = req.MaxIterations
+		}
+
+		// Cache hit: reuse tool defs from first call for non-final iterations.
+		if cacheValid && state.Iteration != maxIter {
+			return cachedToolDefs, nil
+		}
+
 		// Load per-user MCP tools (Notion, etc.) into registry before filtering.
 		// Servers with require_user_credentials are deferred at startup and
 		// connected per-request here with the actual user's credentials.
@@ -222,10 +240,6 @@ func (l *Loop) makeBuildFilteredTools(req *RunRequest) func(state *pipeline.RunS
 			"sender_id", state.Input.SenderID,
 			"actor_user_id", actorUserID,
 			"user_tools_count", len(userTools))
-		maxIter := l.maxIterations
-		if req.MaxIterations > 0 && req.MaxIterations < maxIter {
-			maxIter = req.MaxIterations
-		}
 		allMsgs := state.Messages.All()
 		toolDefs, _, returnedMsgs := l.buildFilteredTools(req, state.Context.HadBootstrap,
 			state.Iteration, maxIter, allMsgs, userTools)
@@ -237,6 +251,13 @@ func (l *Loop) makeBuildFilteredTools(req *RunRequest) func(state *pipeline.RunS
 				state.Messages.AppendPending(msg)
 			}
 		}
+
+		// Cache store after first successful non-final call.
+		if !cacheValid && state.Iteration != maxIter {
+			cachedToolDefs = toolDefs
+			cacheValid = true
+		}
+
 		mcpDefs := 0
 		for _, td := range toolDefs {
 			if td.Function != nil && strings.HasPrefix(strings.TrimSpace(td.Function.Name), "mcp_") {