feat(pruning): enable context pruning by default with cache-ttl mode

- Change default mode from "off" to "cache-ttl" (enabled by default)
- Increase softTrimRatio from 0.25 to 0.3 to match TS defaults
- Update tests to reflect new default-enabled behavior

internal/agent/pruning.go

@@ -14,8 +14,9 @@ import (
 // Context pruning defaults matching TS DEFAULT_CONTEXT_PRUNING_SETTINGS.
 const (
 	defaultKeepLastAssistants   = 3
-	defaultSoftTrimRatio        = 0.25
+	defaultSoftTrimRatio        = 0.3
 	defaultHardClearRatio       = 0.5
+	defaultPruningMode          = "cache-ttl"
 	defaultMinPrunableToolChars = 50000
 	defaultSoftTrimMaxChars     = 6000
 	defaultSoftTrimHeadChars    = 3000
@@ -172,12 +173,16 @@ func resolvePruningSettings(cfg *config.ContextPruningConfig) *effectivePruningS
 // for non-ASCII content like Vietnamese/Chinese). When nil, falls back to the
 // legacy rune_count/charsPerTokenEstimate heuristic so existing tests pass.
 func pruneContextMessages(msgs []providers.Message, contextWindowTokens int, cfg *config.ContextPruningConfig, tc tokencount.TokenCounter, model string, stats *pipeline.PruneStats) []providers.Message {
-	// Opt-in: require explicit mode. Matches TS computeEffectiveSettings in settings.ts.
-	if cfg == nil || cfg.Mode == "" || cfg.Mode == "off" {
+	// Resolve effective mode: empty defaults to "cache-ttl" (enabled by default).
+	mode := defaultPruningMode
+	if cfg != nil && cfg.Mode != "" {
+		mode = cfg.Mode
+	}
+	if mode == "off" {
 		return msgs
 	}
-	if cfg.Mode != "cache-ttl" {
-		slog.Warn("context_pruning: unknown mode, disabled", "mode", cfg.Mode)
+	if mode != "cache-ttl" {
+		slog.Warn("context_pruning: unknown mode, disabled", "mode", mode)
 		return msgs
 	}
 	if contextWindowTokens <= 0 || len(msgs) == 0 {

internal/agent/pruning_test.go

@@ -448,8 +448,8 @@ func TestHasImportantTail_ShortContentNoMatch(t *testing.T) {
 // Lock current (buggy-vs-TS) behavior before refactoring. Subsequent phases
 // update these tests to new expected values.
 
-// makeLargeHistoryFixture creates a history with a large tool result for opt-in tests.
-// contextWindow=5000 → charWindow=20000, total ~8030 chars → ratio ~40% ≥ 25%.
+// makeLargeHistoryFixture creates a history with a large tool result for default-enabled tests.
+// contextWindow=5000 → charWindow=20000, total ~8030 chars → ratio ~40% ≥ 30% (softTrimRatio).
 func makeLargeHistoryFixture() []providers.Message {
 	return []providers.Message{
 		{Role: "user", Content: "q"},
@@ -461,21 +461,25 @@ func makeLargeHistoryFixture() []providers.Message {
 	}
 }
 
-func TestPruneContextMessages_NilCfg_NoOp(t *testing.T) {
-	// Phase 04: nil cfg → no pruning (opt-in). Matches TS design.
+func TestPruneContextMessages_NilCfg_DefaultEnabled(t *testing.T) {
+	// nil cfg defaults to "cache-ttl" mode (enabled by default).
+	// Tool result 8000 chars > 6000 threshold → trimmed to ~6000.
 	msgs := makeLargeHistoryFixture()
 	got := pruneContextMessages(msgs, 5000, nil, nil, "", nil)
-	if estimateMessageChars(got[2]) != 8000 {
-		t.Errorf("nil cfg should be no-op (opt-in). Got %d chars", estimateMessageChars(got[2]))
+	chars := estimateMessageChars(got[2])
+	if chars >= 8000 {
+		t.Errorf("nil cfg should enable pruning (default cache-ttl). Got %d chars, expected < 8000", chars)
 	}
 }
 
-func TestPruneContextMessages_EmptyMode_NoOp(t *testing.T) {
+func TestPruneContextMessages_EmptyMode_DefaultEnabled(t *testing.T) {
+	// Empty mode defaults to "cache-ttl" (enabled by default).
 	cfg := &config.ContextPruningConfig{Mode: ""}
 	msgs := makeLargeHistoryFixture()
 	got := pruneContextMessages(msgs, 5000, cfg, nil, "", nil)
-	if estimateMessageChars(got[2]) != 8000 {
-		t.Errorf("empty mode should be no-op. Got %d chars", estimateMessageChars(got[2]))
+	chars := estimateMessageChars(got[2])
+	if chars >= 8000 {
+		t.Errorf("empty mode should enable pruning (default cache-ttl). Got %d chars, expected < 8000", chars)
 	}
 }