diff --git a/.jules/bolt.md b/.jules/bolt.md
index ed3edb1..18c5dde 100644
--- a/.jules/bolt.md
+++ b/.jules/bolt.md
@@ -130,3 +130,6 @@ If your candidate site already uses any of these, the issue is fixed.
 ## 2025-05-15 - Array Materialization via sorted() on Lazy Sequences
 **Learning:** `LazyFilterSequence` does not have a custom `sorted()` implementation. Calling `sorted()` on a `LazyFilterSequence` relies on the default `Sequence` extension, which completely consumes the sequence and materializes it into an array immediately. While chaining `.lazy.filter` before `.sorted()` avoids allocating a separate intermediate array just for the filtered results, `sorted()` itself forces array materialization regardless. Adding `.lazy` specifically *just* before `.sorted()` creates confusion because it suggests the sort operation itself is lazy (which is impossible, sorting requires the whole collection). The codebase considers `.lazy.filter.sorted` an anti-pattern.
 **Action:** Do not propose adding `.lazy` before `.sorted()`. Ensure `.lazy` is only proposed when preceding operations that natively support lazy consumption like `.reduce`, `.count`, `.first`, `.contains`, or lazy `prefix`.
+## 2025-05-16 - Refactoring Filter Methods for Iteration Optimization
+**Learning:** Functions that encapsulate `.filter` logic and return a filtered array force intermediate array allocation when called within a loop, especially for potentially large datasets like process lists (~400-600 items). The `for ... where` idiom cannot be applied when the condition logic is hidden inside a method that returns the already filtered array.
+**Action:** When a method exists purely to filter an array for subsequent iteration, refactor the method into a `static func` (or regular function) that returns a `Bool` evaluating a single element. This allows the caller to use an efficient `for element in collection where isCondition(element)` pattern, avoiding intermediate memory allocations.
diff --git a/Sources/Cacheout/Memory/PredictiveEngine.swift b/Sources/Cacheout/Memory/PredictiveEngine.swift
index 725b3cd..e1ba422 100644
--- a/Sources/Cacheout/Memory/PredictiveEngine.swift
+++ b/Sources/Cacheout/Memory/PredictiveEngine.swift
@@ -15,7 +15,7 @@
 /// engine.recordAvailableMB(mb, at: timestamp)
 /// // Query predictions:
 /// let tte = engine.predictTimeToExhaustion()
-/// let growers = await engine.detectHighGrowthProcesses()
+/// // Use PredictiveEngine.isHighGrowthProcess(entry) to filter growers
 /// ```
 
 import CacheoutShared
@@ -188,23 +188,21 @@ actor PredictiveEngine {
 
     // MARK: - High-Growth Process Detection
 
-    /// Detect processes that have grown to and remain at their lifetime peak.
+    /// Check if a process has grown to and remains at its lifetime peak.
     ///
-    /// Filters for processes where:
+    /// Matches processes where:
     /// - `leakIndicator < 1.05` (within 5% of lifetime peak)
     /// - `physFootprint > 500MB` (significant memory consumer)
     ///
     /// This is NOT leak detection -- it identifies processes with sustained growth
     /// that are currently at or near their maximum footprint.
     ///
-    /// - Parameter processes: Process entries to evaluate.
-    /// - Returns: Processes matching high-growth criteria.
-    func detectHighGrowthProcesses(from processes: [ProcessEntryDTO]) -> [ProcessEntryDTO] {
-        processes.filter { entry in
-            entry.leakIndicator < Self.highGrowthMaxLeakIndicator
-                && entry.leakIndicator > 0  // Exclude zero (no footprint data)
-                && entry.physFootprint > Self.highGrowthMinFootprint
-        }
+    /// - Parameter entry: The process entry to evaluate.
+    /// - Returns: True if the process matches high-growth criteria.
+    static func isHighGrowthProcess(_ entry: ProcessEntryDTO) -> Bool {
+        entry.leakIndicator < Self.highGrowthMaxLeakIndicator
+            && entry.leakIndicator > 0  // Exclude zero (no footprint data)
+            && entry.physFootprint > Self.highGrowthMinFootprint
     }
 
     // MARK: - Process Scan Cache
diff --git a/Sources/Cacheout/Memory/RecommendationEngine.swift b/Sources/Cacheout/Memory/RecommendationEngine.swift
index f5d5300..2612c2f 100644
--- a/Sources/Cacheout/Memory/RecommendationEngine.swift
+++ b/Sources/Cacheout/Memory/RecommendationEngine.swift
@@ -189,8 +189,8 @@ struct RecommendationEngine {
         }
 
         // high_growth_process
-        let highGrowth = await predictiveEngine.detectHighGrowthProcesses(from: scanResult.processes)
-        for proc in highGrowth {
+        // ⚡ Bolt: Use for-where clause to prevent intermediate array allocation
+        for proc in scanResult.processes where PredictiveEngine.isHighGrowthProcess(proc) {
             let footprintGB = Double(proc.physFootprint) / (1024 * 1024 * 1024)
             recommendations.append(Recommendation(
                 type: .highGrowthProcess,