diff --git a/run_test.go b/run_test.go
index 6a070d68..b3ccd4d2 100644
--- a/run_test.go
+++ b/run_test.go
@@ -795,3 +795,92 @@ func Test_TestSuite_RetreiveFeatures(t *testing.T) {
 		})
 	}
 }
+
+// TestHookPanic verifies that panics in ScenarioContext.After, ScenarioContext.Before,
+// StepContext.Before, and StepContext.After hooks are handled gracefully by the godog.
+// If this test panics and fails to report gracefully, it means the panic handling in the hooks is insufficient.
+func TestHookPanic(t *testing.T) {
+	// create a temporary file
+	f, err := os.CreateTemp("", "godog_test_hook_panic_*.feature")
+	if err != nil {
+		t.Fatalf("failed to create temporary file: %v", err)
+	}
+	defer func(name string) {
+		if err := os.Remove(f.Name()); err != nil {
+			t.Fatalf("failed to remove temporary file: %v", err)
+		}
+	}(f.Name())
+
+	_, err = f.Write([]byte(`
+Feature: count
+
+  Scenario: count one two
+    When one
+    Then two
+
+  Scenario: count three four
+    When three
+    Then four
+`))
+	if err != nil {
+		t.Fatalf("failed to write to temporary file: %v", err)
+	}
+	err = f.Close()
+	if err != nil {
+		t.Fatalf("failed to close temporary file: %v", err)
+	}
+	setupSteps := func(sc *ScenarioContext) {
+		sc.Step(`^one$`, func() {})
+		sc.Step(`^two$`, func() {})
+		sc.Step(`^three$`, func() {})
+		sc.Step(`^four$`, func() {})
+	}
+	for name, test := range map[string]struct {
+		initScenarios func(*ScenarioContext)
+	}{
+		"before scenario": {
+			func(ctx *ScenarioContext) {
+				ctx.Before(func(ctx context.Context, sc *Scenario) (context.Context, error) {
+					panic("before scenario")
+				})
+				setupSteps(ctx)
+			},
+		},
+		"after scenario": {
+			func(ctx *ScenarioContext) {
+				setupSteps(ctx)
+				ctx.After(func(ctx context.Context, sc *Scenario, err error) (context.Context, error) {
+					panic("after scenario")
+				})
+			},
+		},
+		"after step": {
+			func(ctx *ScenarioContext) {
+				setupSteps(ctx)
+				ctx.StepContext().After(func(_ context.Context, _ *Step, _ StepResultStatus, _ error) (context.Context, error) {
+					panic("after step")
+				})
+			},
+		},
+		"before step": {
+			func(ctx *ScenarioContext) {
+				setupSteps(ctx)
+				ctx.StepContext().Before(func(_ context.Context, _ *Step) (context.Context, error) {
+					panic("before step")
+				})
+			},
+		},
+	} {
+		t.Run(name, func(t *testing.T) {
+			opts := Options{
+				Format: "pretty",
+				Paths:  []string{f.Name()},
+			}
+			TestSuite{
+				Name:                "HookPanic!",
+				Options:             &opts,
+				ScenarioInitializer: test.initScenarios,
+			}.Run()
+		})
+	}
+}
diff --git a/suite.go b/suite.go
index 6ca1bf53..5d50ed06 100644
--- a/suite.go
+++ b/suite.go
@@ -320,25 +320,38 @@ func (s *suite) runBeforeStepHooks(ctx context.Context, step *Step, err error) (
 	return ctx, err
 }
 
-func (s *suite) runAfterStepHooks(ctx context.Context, step *Step, status StepResultStatus, err error) (context.Context, error) {
+func (s *suite) runAfterStepHooks(ctx context.Context, step *Step, status StepResultStatus, err error) (rctx context.Context, rerr error) {
+	rerr = err
+	rctx = ctx
+	defer func() {
+		if r := recover(); r != nil {
+			pe := fmt.Errorf("panic: %v", r)
+			if rerr == nil {
+				rerr = pe
+			} else {
+				rerr = fmt.Errorf("%v; %w", pe, rerr)
+			}
+		}
+	}()
+
 	for _, f := range s.afterStepHandlers {
-		hctx, herr := f(ctx, step, status, err)
+		hctx, herr := f(rctx, step, status, rerr)
 
 		// Adding hook error to resulting error without breaking hooks loop.
 		if herr != nil {
-			if err == nil {
-				err = herr
+			if rerr == nil {
+				rerr = herr
 			} else {
-				err = fmt.Errorf("%v, %w", herr, err)
+				rerr = fmt.Errorf("%v, %w", herr, rerr)
 			}
 		}
 
 		if hctx != nil {
-			ctx = hctx
+			rctx = hctx
 		}
 	}
 
-	return ctx, err
+	return rctx, rerr
 }
 
 func (s *suite) runBeforeScenarioHooks(ctx context.Context, pickle *messages.Pickle) (context.Context, error) {
@@ -367,15 +380,26 @@ func (s *suite) runBeforeScenarioHooks(ctx context.Context, pickle *messages.Pic
 	return ctx, err
 }
 
-func (s *suite) runAfterScenarioHooks(ctx context.Context, pickle *messages.Pickle, lastStepErr error) (context.Context, error) {
-	err := lastStepErr
+func (s *suite) runAfterScenarioHooks(ctx context.Context, pickle *messages.Pickle, lastStepErr error) (rctx context.Context, err error) {
+	rctx = ctx
+	err = lastStepErr
+	defer func() {
+		if r := recover(); r != nil {
+			pe := fmt.Errorf("panic: %v", r)
+			if err == nil {
+				err = pe
+			} else {
+				err = fmt.Errorf("%v; %w", pe, err)
+			}
+		}
+	}()
 
 	hooksFailed := false
 	isStepErr := true
 
 	// run after scenario handlers
 	for _, f := range s.afterScenarioHandlers {
-		hctx, herr := f(ctx, pickle, err)
+		hctx, herr := f(rctx, pickle, err)
 
 		// Adding hook error to resulting error without breaking hooks loop.
 		if herr != nil {
@@ -394,7 +418,7 @@ func (s *suite) runAfterScenarioHooks(ctx context.Context, pickle *messages.Pick
 		}
 
 		if hctx != nil {
-			ctx = hctx
+			rctx = hctx
 		}
 	}
 
@@ -402,7 +426,7 @@ func (s *suite) runAfterScenarioHooks(ctx context.Context, pickle *messages.Pick
 		err = fmt.Errorf("after scenario hook failed: %w", err)
 	}
 
-	return ctx, err
+	return rctx, err
 }
 
 func (s *suite) maybeUndefined(ctx context.Context, text string, arg interface{}, stepType messages.PickleStepType) (context.Context, []string, error) {