refactor(detekt): extract magic-number constants, wrap long lines, drop legit suppressions

Mechanical detekt cleanup across apps/core/infrastructure (behavior-preserving):
- MagicNumber 62->10: named constants + removed 'legit' MagicNumber suppressions
  (Tier level from ordinal, ReplayInferenceProvider CHARS_PER_TOKEN, ConfigLoader
  DEFAULT_* consts, capability scores, HTTP ranges, column widths, etc.)
- MaxLineLength cut to ~0 in production modules (line wraps, no logic change)
- Dropped one dead parameter (ConfigLoader.parseArray lineNum)

Structural findings (ReturnCount/LongMethod/Complexity/LargeClass) left for a
deliberate decomposition pass. Full build + detekt gate green.

core/talkie/src/main/kotlin/com/correx/core/talkie/TalkieContextBuilder.kt

@@ -108,6 +108,9 @@ class DefaultTalkieContextBuilder(
 
         /** Cap on ideas folded into router context, newest-first, so the board can't balloon L0. */
         private const val MAX_IDEAS_IN_CONTEXT = 10
+
+        /** Rough chars-per-token heuristic used when no tokenizer is available. */
+        private const val CHARS_PER_TOKEN_ESTIMATE = 4
     }
 
     override suspend fun build(
@@ -389,7 +392,7 @@ class DefaultTalkieContextBuilder(
     }
 
     private fun fallbackEstimate(content: String): Int {
-        return (content.length / 4).coerceAtLeast(1)
+        return (content.length / CHARS_PER_TOKEN_ESTIMATE).coerceAtLeast(1)
     }
 
     /**

core/talkie/src/main/kotlin/com/correx/core/talkie/TalkieFacade.kt

@@ -146,7 +146,12 @@ class DefaultTalkieFacade(
         // Rebuild state so lastRetrievedMemory reflects this turn
         state = routerRepository.getTalkieState(sessionId)
 
-        val contextPack = routerContextBuilder.build(state, config.tokenBudget, workflowSummaryProvider(), sessionProfileProvider(sessionId))
+        val contextPack = routerContextBuilder.build(
+            state,
+            config.tokenBudget,
+            workflowSummaryProvider(),
+            sessionProfileProvider(sessionId),
+        )
 
         if (contextPack.compressionMetadata.entriesDropped > 0) {
             val truncNow = Clock.System.now()
@@ -195,7 +200,8 @@ class DefaultTalkieFacade(
                     "without producing an answer (likely spent on hidden reasoning). " +
                     "Raise [router.generation] max_tokens or lower the model's reasoning budget, then ask again."
             } else {
-                "The model returned an empty response (finish reason: ${inferenceResponse.finishReason::class.simpleName}). Try again."
+                "The model returned an empty response " +
+                    "(finish reason: ${inferenceResponse.finishReason::class.simpleName}). Try again."
             }
         }