executor: detect stream-level failures when claude exits 0

Rename streamAndParseCost → parseStream (returns float64, error).
Detect two failure modes that claude reports via exit 0:
- result message with is_error:true
- tool_result permission denial ("haven't granted it yet")

Also fix cost extraction to read total_cost_usd from the result
message (the actual field name), keeping cost_usd as legacy fallback.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

1 files changed, 85 insertions, 0 deletions

diff --git a/internal/executor/stream_test.go b/internal/executor/stream_test.go
new file mode 100644
index 0000000..10eb858
--- /dev/null
+++ b/internal/executor/stream_test.go
@@ -0,0 +1,85 @@
+package executor
+
+import (
+	"io"
+	"log/slog"
+	"strings"
+	"testing"
+)
+
+// streamLine builds a single-line stream-json message.
+func streamLine(json string) string { return json + "\n" }
+
+func TestParseStream_ResultIsError_ReturnsError(t *testing.T) {
+	input := streamLine(`{"type":"result","subtype":"error_during_execution","is_error":true,"result":"something went wrong"}`)
+	_, err := parseStream(strings.NewReader(input), io.Discard, slog.New(slog.NewTextHandler(io.Discard, nil)))
+	if err == nil {
+		t.Fatal("expected error when result.is_error=true, got nil")
+	}
+	if !strings.Contains(err.Error(), "something went wrong") {
+		t.Errorf("error should contain result text, got: %v", err)
+	}
+}
+
+func TestParseStream_PermissionDenied_ReturnsError(t *testing.T) {
+	// Simulate the permission denial tool_result that Claude Code emits
+	// when permission_mode is not bypassPermissions.
+	input := streamLine(`{"type":"user","message":{"role":"user","content":[{"type":"tool_result","is_error":true,"content":"Claude requested permissions to write to /foo/bar.go, but you haven't granted it yet.","tool_use_id":"tu_abc"}]}}`) +
+		streamLine(`{"type":"result","subtype":"success","is_error":false,"result":"I need permission","total_cost_usd":0.1}`)
+
+	_, err := parseStream(strings.NewReader(input), io.Discard, slog.New(slog.NewTextHandler(io.Discard, nil)))
+	if err == nil {
+		t.Fatal("expected error for permission denial, got nil")
+	}
+	if !strings.Contains(err.Error(), "permission") {
+		t.Errorf("error should mention permission, got: %v", err)
+	}
+}
+
+func TestParseStream_Success_ReturnsNilError(t *testing.T) {
+	input := streamLine(`{"type":"assistant","message":{"content":[{"type":"text","text":"Done."}]}}`) +
+		streamLine(`{"type":"result","subtype":"success","is_error":false,"result":"All tests pass.","total_cost_usd":0.05}`)
+
+	_, err := parseStream(strings.NewReader(input), io.Discard, slog.New(slog.NewTextHandler(io.Discard, nil)))
+	if err != nil {
+		t.Fatalf("expected nil error for success stream, got: %v", err)
+	}
+}
+
+func TestParseStream_ExtractsCostFromResultMessage(t *testing.T) {
+	input := streamLine(`{"type":"result","subtype":"success","is_error":false,"result":"done","total_cost_usd":1.2345}`)
+
+	cost, err := parseStream(strings.NewReader(input), io.Discard, slog.New(slog.NewTextHandler(io.Discard, nil)))
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if cost != 1.2345 {
+		t.Errorf("want cost 1.2345, got %f", cost)
+	}
+}
+
+func TestParseStream_ExtractsCostFromLegacyCostUSD(t *testing.T) {
+	// Some versions emit cost_usd at the top level rather than total_cost_usd.
+	input := streamLine(`{"type":"result","subtype":"success","is_error":false,"result":"done","cost_usd":0.99}`)
+
+	cost, err := parseStream(strings.NewReader(input), io.Discard, slog.New(slog.NewTextHandler(io.Discard, nil)))
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if cost != 0.99 {
+		t.Errorf("want cost 0.99, got %f", cost)
+	}
+}
+
+func TestParseStream_NonToolResultIsError_DoesNotFail(t *testing.T) {
+	// A tool_result with is_error:true that is NOT a permission denial
+	// (e.g. the tool ran but the command failed) should not mark the task failed —
+	// the agent can recover from individual tool errors.
+	input := streamLine(`{"type":"user","message":{"role":"user","content":[{"type":"tool_result","is_error":true,"content":"exit status 1","tool_use_id":"tu_xyz"}]}}`) +
+		streamLine(`{"type":"result","subtype":"success","is_error":false,"result":"Fixed it.","total_cost_usd":0.2}`)
+
+	_, err := parseStream(strings.NewReader(input), io.Discard, slog.New(slog.NewTextHandler(io.Discard, nil)))
+	if err != nil {
+		t.Fatalf("non-permission tool errors should not fail the task, got: %v", err)
+	}
+}