diff --git a/__init__.py b/__init__.py
new file mode 100644
index 0000000..0009878
--- /dev/null
+++ b/__init__.py
@@ -0,0 +1,47 @@
+
+import importlib.util
+import os
+import sys
+import json
+
+NODE_CLASS_MAPPINGS = {}
+NODE_DISPLAY_NAME_MAPPINGS = {}
+WEB_DIRECTORY = "web"
+python = sys.executable
+
+def get_ext_dir(subpath=None, mkdir=False):
+    dir = os.path.dirname(__file__)
+    if subpath is not None:
+        dir = os.path.join(dir, subpath)
+
+    dir = os.path.abspath(dir)
+
+    if mkdir and not os.path.exists(dir):
+        os.makedirs(dir)
+    return dir
+
+def serialize(obj):
+    if isinstance(obj, (str, int, float, bool, list, dict, type(None))):
+        return obj
+    return str(obj)  
+
+
+py = get_ext_dir("py")
+files = os.listdir(py)
+all_nodes = {}
+for file in files:
+    if not file.endswith(".py"):
+        continue
+    name = os.path.splitext(file)[0]
+    imported_module = importlib.import_module(".py.{}".format(name), __name__)
+    try:
+        NODE_CLASS_MAPPINGS = {**NODE_CLASS_MAPPINGS, **imported_module.NODE_CLASS_MAPPINGS}
+        NODE_DISPLAY_NAME_MAPPINGS = {**NODE_DISPLAY_NAME_MAPPINGS, **imported_module.NODE_DISPLAY_NAME_MAPPINGS}
+        serialized_CLASS_MAPPINGS = {k: serialize(v) for k, v in imported_module.NODE_CLASS_MAPPINGS.items()}
+        serialized_DISPLAY_NAME_MAPPINGS = {k: serialize(v) for k, v in imported_module.NODE_DISPLAY_NAME_MAPPINGS.items()}
+        all_nodes[file]={"NODE_CLASS_MAPPINGS": serialized_CLASS_MAPPINGS, "NODE_DISPLAY_NAME_MAPPINGS": serialized_DISPLAY_NAME_MAPPINGS}
+    except:
+        pass
+
+
+__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS", "WEB_DIRECTORY"]
diff --git a/py/noise_injection.py b/py/noise_injection.py
new file mode 100644
index 0000000..71db31d
--- /dev/null
+++ b/py/noise_injection.py
@@ -0,0 +1,369 @@
+"""
+ZImage Noise Injection 节点
+
+通过 CFG 机制将参考图像的特征注入到生成中。
+可以让生成结果"学习"参考图像的某些特质（如水珠、纹理、质感等）。
+"""
+
+import torch
+import torch.nn.functional as F
+import logging
+import comfy.model_management
+import comfy.latent_formats
+
+
+class LGNoiseInjection:
+    """
+    LG Noise Injection - 特征注入
+    
+    工作原理：
+    参考图像的 latent 特征会被注入到 CFG 过程中，
+    模型会"学习"参考图像的某些特质并应用到生成结果上。
+    
+    适用场景：
+    - 添加水珠、汗珠等表面细节
+    - 添加纹理、材质感
+    - 添加光泽、反射效果
+    """
+    
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "model": ("MODEL", {
+                    "tooltip": "模型"
+                }),
+                "vae": ("VAE", {
+                    "tooltip": "VAE 编码器"
+                }),
+                "reference_image": ("IMAGE", {
+                    "tooltip": "参考图像（含有你想要注入的特征）"
+                }),
+                "strength": ("FLOAT", {
+                    "default": 0.15,
+                    "min": 0.0,
+                    "max": 1.0,
+                    "step": 0.01,
+                    "tooltip": "注入强度。0.1-0.2 轻微，0.2-0.4 明显"
+                }),
+                "start_percent": ("FLOAT", {
+                    "default": 0.0,
+                    "min": 0.0,
+                    "max": 1.0,
+                    "step": 0.01,
+                    "tooltip": "开始注入的采样进度"
+                }),
+                "end_percent": ("FLOAT", {
+                    "default": 0.6,
+                    "min": 0.0,
+                    "max": 1.0,
+                    "step": 0.01,
+                    "tooltip": "结束注入的采样进度"
+                }),
+            },
+            "optional": {
+                "mask": ("MASK", {
+                    "tooltip": "遮罩，白色区域会被注入特征"
+                }),
+            },
+        }
+
+    RETURN_TYPES = ("MODEL",)
+    FUNCTION = "apply"
+    CATEGORY = "advanced/model"
+    DESCRIPTION = "将参考图像的特征（如水珠、纹理等）注入到生成结果中。"
+
+    def apply(self, model, vae, reference_image, strength, start_percent, end_percent, mask=None):
+        if strength <= 0:
+            return (model,)
+        
+        m = model.clone()
+        
+        # 编码参考图像
+        ref_latent = self._encode_reference(vae, reference_image)
+        
+        # 预处理 mask
+        mask_latent = None
+        if mask is not None:
+            mask_latent = self._prepare_mask(mask)
+        
+        step_counter = [0]
+        
+        def cfg_function(args):
+            """在 CFG 合并后注入参考特征"""
+            cond = args["cond"]
+            uncond = args["uncond"]
+            cond_scale = args["cond_scale"]
+            timestep = args["timestep"]
+            
+            step_counter[0] += 1
+            
+            # 标准 CFG
+            cfg_result = uncond + cond_scale * (cond - uncond)
+            
+            # 计算进度
+            sigma = float(timestep[0]) if timestep.dim() > 0 else float(timestep)
+            progress = 1.0 - min(sigma, 1.0)
+            
+            # 检查是否在作用范围内
+            if progress < start_percent or progress > end_percent:
+                return cfg_result
+            
+            # 准备 reference
+            ref = ref_latent.to(device=cfg_result.device, dtype=cfg_result.dtype)
+            
+            if ref.shape[2:] != cfg_result.shape[2:]:
+                ref = F.interpolate(ref, size=cfg_result.shape[2:], mode='bilinear', align_corners=False)
+            
+            if ref.shape[0] != cfg_result.shape[0]:
+                if ref.shape[0] == 1:
+                    ref = ref.expand(cfg_result.shape[0], -1, -1, -1)
+                else:
+                    ref = ref[:cfg_result.shape[0]]
+            
+            # 准备 mask
+            current_mask = None
+            if mask_latent is not None:
+                current_mask = mask_latent.to(device=cfg_result.device, dtype=cfg_result.dtype)
+                # 调整 mask 尺寸到 latent 空间
+                if current_mask.shape[2:] != cfg_result.shape[2:]:
+                    current_mask = F.interpolate(current_mask, size=cfg_result.shape[2:], mode='bilinear', align_corners=False)
+                # 调整 batch size
+                if current_mask.shape[0] != cfg_result.shape[0]:
+                    if current_mask.shape[0] == 1:
+                        current_mask = current_mask.expand(cfg_result.shape[0], -1, -1, -1)
+                    else:
+                        current_mask = current_mask[:cfg_result.shape[0]]
+            
+            # 计算有效强度（线性衰减）
+            if end_percent > start_percent:
+                range_progress = (progress - start_percent) / (end_percent - start_percent)
+                decay = 1.0 - range_progress
+            else:
+                decay = 1.0
+            effective_strength = strength * decay
+            
+            # 计算特征注入方向
+            feature_direction = ref - cfg_result
+            
+            # 控制注入幅度，避免过度偏移
+            cfg_std = cfg_result.std()
+            feature_std = feature_direction.std()
+            if feature_std > cfg_std * 3:
+                feature_direction = feature_direction * (cfg_std * 3 / feature_std)
+            
+            # 应用遮罩
+            if current_mask is not None:
+                # mask: 1 = 注入区域, 0 = 保持原样
+                feature_direction = feature_direction * current_mask
+            
+            # 应用特征注入
+            injected = cfg_result + feature_direction * effective_strength
+            
+            if step_counter[0] <= 3:
+                mask_info = "with mask" if current_mask is not None else "no mask"
+                logging.warning(f"[FeatureInj] step={step_counter[0]} | progress={progress:.2f} | eff_str={effective_strength:.3f} | {mask_info}")
+            
+            return injected
+        
+        m.set_model_sampler_cfg_function(cfg_function)
+        
+        return (m,)
+    
+    def _encode_reference(self, vae, reference_image):
+        """编码参考图像"""
+        loaded_models = comfy.model_management.loaded_models(only_currently_used=True)
+        latent = vae.encode(reference_image)
+        latent = comfy.latent_formats.Flux().process_in(latent)
+        comfy.model_management.load_models_gpu(loaded_models)
+        logging.warning(f"[FeatureInj] Reference encoded: shape={latent.shape}")
+        return latent
+    
+    def _prepare_mask(self, mask):
+        """准备遮罩，转换为 latent 空间格式"""
+        # mask 输入格式: [B, H, W] 或 [H, W]
+        if mask.dim() == 2:
+            mask = mask.unsqueeze(0)  # [H, W] -> [1, H, W]
+        
+        # 添加 channel 维度: [B, H, W] -> [B, 1, H, W]
+        mask = mask.unsqueeze(1)
+        
+        logging.warning(f"[FeatureInj] Mask prepared: shape={mask.shape}")
+        return mask
+
+
+class LGNoiseInjectionLatent:
+    """
+    LG Noise Injection (Latent) - 直接使用 Latent 的特征注入
+    
+    工作原理：
+    直接输入参考 latent，其特征会被注入到 CFG 过程中。
+    如果 latent 包含 noise_mask，则自动使用该遮罩。
+    
+    适用场景：
+    - 添加水珠、汗珠等表面细节
+    - 添加纹理、材质感
+    - 添加光泽、反射效果
+    """
+    
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "model": ("MODEL", {
+                    "tooltip": "模型"
+                }),
+                "reference_latent": ("LATENT", {
+                    "tooltip": "参考 latent（含有你想要注入的特征）"
+                }),
+                "strength": ("FLOAT", {
+                    "default": 0.15,
+                    "min": 0.0,
+                    "max": 1.0,
+                    "step": 0.01,
+                    "tooltip": "注入强度。0.1-0.2 轻微，0.2-0.4 明显"
+                }),
+                "start_percent": ("FLOAT", {
+                    "default": 0.0,
+                    "min": 0.0,
+                    "max": 1.0,
+                    "step": 0.01,
+                    "tooltip": "开始注入的采样进度"
+                }),
+                "end_percent": ("FLOAT", {
+                    "default": 0.6,
+                    "min": 0.0,
+                    "max": 1.0,
+                    "step": 0.01,
+                    "tooltip": "结束注入的采样进度"
+                }),
+            },
+        }
+
+    RETURN_TYPES = ("MODEL",)
+    FUNCTION = "apply"
+    CATEGORY = "advanced/model"
+    DESCRIPTION = "直接输入 latent 进行特征注入，自动使用 latent 的 noise_mask 作为遮罩。"
+
+    def apply(self, model, reference_latent, strength, start_percent, end_percent):
+        if strength <= 0:
+            return (model,)
+        
+        m = model.clone()
+        
+        # 获取 latent samples
+        ref_latent = reference_latent["samples"]
+        
+        # 获取 noise_mask（如果存在）
+        mask_latent = None
+        if "noise_mask" in reference_latent:
+            mask_latent = self._prepare_mask(reference_latent["noise_mask"])
+            logging.warning(f"[FeatureInjLatent] Using noise_mask from latent")
+        
+        logging.warning(f"[FeatureInjLatent] Reference latent: shape={ref_latent.shape}")
+        
+        step_counter = [0]
+        
+        def cfg_function(args):
+            """在 CFG 合并后注入参考特征"""
+            cond = args["cond"]
+            uncond = args["uncond"]
+            cond_scale = args["cond_scale"]
+            timestep = args["timestep"]
+            
+            step_counter[0] += 1
+            
+            # 标准 CFG
+            cfg_result = uncond + cond_scale * (cond - uncond)
+            
+            # 计算进度
+            sigma = float(timestep[0]) if timestep.dim() > 0 else float(timestep)
+            progress = 1.0 - min(sigma, 1.0)
+            
+            # 检查是否在作用范围内
+            if progress < start_percent or progress > end_percent:
+                return cfg_result
+            
+            # 准备 reference
+            ref = ref_latent.to(device=cfg_result.device, dtype=cfg_result.dtype)
+            
+            if ref.shape[2:] != cfg_result.shape[2:]:
+                ref = F.interpolate(ref, size=cfg_result.shape[2:], mode='bilinear', align_corners=False)
+            
+            if ref.shape[0] != cfg_result.shape[0]:
+                if ref.shape[0] == 1:
+                    ref = ref.expand(cfg_result.shape[0], -1, -1, -1)
+                else:
+                    ref = ref[:cfg_result.shape[0]]
+            
+            # 准备 mask
+            current_mask = None
+            if mask_latent is not None:
+                current_mask = mask_latent.to(device=cfg_result.device, dtype=cfg_result.dtype)
+                # 调整 mask 尺寸到 latent 空间
+                if current_mask.shape[2:] != cfg_result.shape[2:]:
+                    current_mask = F.interpolate(current_mask, size=cfg_result.shape[2:], mode='bilinear', align_corners=False)
+                # 调整 batch size
+                if current_mask.shape[0] != cfg_result.shape[0]:
+                    if current_mask.shape[0] == 1:
+                        current_mask = current_mask.expand(cfg_result.shape[0], -1, -1, -1)
+                    else:
+                        current_mask = current_mask[:cfg_result.shape[0]]
+            
+            # 计算有效强度（线性衰减）
+            if end_percent > start_percent:
+                range_progress = (progress - start_percent) / (end_percent - start_percent)
+                decay = 1.0 - range_progress
+            else:
+                decay = 1.0
+            effective_strength = strength * decay
+            
+            # 计算特征注入方向
+            feature_direction = ref - cfg_result
+            
+            # 控制注入幅度，避免过度偏移
+            cfg_std = cfg_result.std()
+            feature_std = feature_direction.std()
+            if feature_std > cfg_std * 3:
+                feature_direction = feature_direction * (cfg_std * 3 / feature_std)
+            
+            # 应用遮罩
+            if current_mask is not None:
+                # mask: 1 = 注入区域, 0 = 保持原样
+                feature_direction = feature_direction * current_mask
+            
+            # 应用特征注入
+            injected = cfg_result + feature_direction * effective_strength
+            
+            if step_counter[0] <= 3:
+                mask_info = "with noise_mask" if current_mask is not None else "no mask"
+                logging.warning(f"[FeatureInjLatent] step={step_counter[0]} | progress={progress:.2f} | eff_str={effective_strength:.3f} | {mask_info}")
+            
+            return injected
+        
+        m.set_model_sampler_cfg_function(cfg_function)
+        
+        return (m,)
+    
+    def _prepare_mask(self, mask):
+        """准备遮罩，转换为 latent 空间格式"""
+        # mask 输入格式: [B, H, W] 或 [H, W]
+        if mask.dim() == 2:
+            mask = mask.unsqueeze(0)  # [H, W] -> [1, H, W]
+        
+        # 添加 channel 维度: [B, H, W] -> [B, 1, H, W]
+        if mask.dim() == 3:
+            mask = mask.unsqueeze(1)
+        
+        return mask
+
+
+NODE_CLASS_MAPPINGS = {
+    "LGNoiseInjection": LGNoiseInjection,
+    "LGNoiseInjectionLatent": LGNoiseInjectionLatent,
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "LGNoiseInjection": "🎈LG Noise Injection",
+    "LGNoiseInjectionLatent": "🎈LG Noise Injection (Latent)",
+}
diff --git a/py/sampling_offset.py b/py/sampling_offset.py
new file mode 100644
index 0000000..42626b6
--- /dev/null
+++ b/py/sampling_offset.py
@@ -0,0 +1,87 @@
+"""
+ZImage Model Sampling 节点
+
+用于调整 ZImage/Lumina2 模型的采样参数。
+关键区别：ZImage 使用 multiplier=1.0，而 SD3 使用 multiplier=1000。
+"""
+
+import torch
+import logging
+import comfy.model_sampling
+
+
+class ModelSamplingZImage:
+    """
+    ZImage/Lumina2 采样参数调整节点
+    
+    与 ModelSamplingSD3 的关键区别：
+    - ZImage 使用 multiplier=1.0（timestep 范围 0-1）
+    - SD3 使用 multiplier=1000（timestep 范围 0-1000）
+    
+    shift 参数控制噪声调度的偏移（通过 time_snr_shift 函数）：
+    - shift=1.0: 无偏移，线性调度
+    - shift>1.0: 向高噪声偏移，早期步骤更激进
+    - ZImage 默认 shift=3.0
+    """
+    
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "model": ("MODEL",),
+                "shift": ("FLOAT", {
+                    "default": 3.0, 
+                    "min": 0.0, 
+                    "max": 100.0, 
+                    "step": 0.01,
+                    "tooltip": "噪声调度偏移量。ZImage 默认 3.0。shift=1.0 为线性，>1.0 向高噪声偏移"
+                }),
+                "multiplier": ("FLOAT", {
+                    "default": 1.0,
+                    "min": 0.001,
+                    "max": 10000.0,
+                    "step": 0.001,
+                    "tooltip": "Timestep 乘数。ZImage/AuraFlow=1.0，SD3/Flux=1000"
+                }),
+            }
+        }
+
+    RETURN_TYPES = ("MODEL",)
+    FUNCTION = "patch"
+    CATEGORY = "advanced/model"
+    DESCRIPTION = "调整 ZImage/Lumina2 模型的采样参数。可设置 shift 和 multiplier。"
+
+    def patch(self, model, shift, multiplier):
+        m = model.clone()
+
+        sampling_base = comfy.model_sampling.ModelSamplingDiscreteFlow
+        sampling_type = comfy.model_sampling.CONST
+
+        class ModelSamplingAdvanced(sampling_base, sampling_type):
+            pass
+
+        model_sampling = ModelSamplingAdvanced(model.model.model_config)
+        model_sampling.set_parameters(shift=shift, multiplier=multiplier)
+        m.add_object_patch("model_sampling", model_sampling)
+        
+        # 调试输出
+        logging.info(f"[ModelSamplingZImage] Applied: shift={shift}, multiplier={multiplier}")
+        logging.info(f"[ModelSamplingZImage] sigma_min={model_sampling.sigma_min:.6f}, sigma_max={model_sampling.sigma_max:.6f}")
+        logging.info(f"[ModelSamplingZImage] sigmas[0:5]={model_sampling.sigmas[:5].tolist()}")
+        
+        # 验证 patch 是否成功
+        patched_ms = m.get_model_object("model_sampling")
+        logging.info(f"[ModelSamplingZImage] Patch verification: patched shift={patched_ms.shift}, multiplier={patched_ms.multiplier}")
+        logging.info(f"[ModelSamplingZImage] Patched sigmas[0:5]={patched_ms.sigmas[:5].tolist()}")
+        
+        return (m,)
+
+
+# 注册节点
+NODE_CLASS_MAPPINGS = {
+    "ModelSamplingZImage": ModelSamplingZImage,
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "ModelSamplingZImage": "Model Sampling ZImage",
+}
diff --git a/py/sigmas_editor.py b/py/sigmas_editor.py
new file mode 100644
index 0000000..7c6c0ca
--- /dev/null
+++ b/py/sigmas_editor.py
@@ -0,0 +1,104 @@
+"""
+Interactive Sigmas Editor Node
+Allows real-time adjustment of sigmas curve by dragging points
+"""
+
+import torch
+import numpy as np
+import json
+import os
+import folder_paths
+from server import PromptServer
+from aiohttp import web
+
+# Set web directory for custom UI
+WEB_DIRECTORY = "./web"
+
+
+class SigmasEditor:
+    """Interactive editor for adjusting sigmas curve"""
+    
+    # 类级别缓存，存储每个节点上次接收的输入sigmas（用于检测输入是否变化）
+    _last_sent_data = {}
+    
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {
+            "required": {
+                "sigmas": ("SIGMAS", {"tooltip": "Input sigmas schedule to edit"}),
+                "sigmas_adjustments": ("STRING", {
+                    "default": "[]",
+                    "multiline": False,
+                    "dynamicPrompts": False,
+                    "tooltip": "JSON array of adjusted sigma values for each step"
+                }),
+            },
+            "hidden": {
+                "unique_id": "UNIQUE_ID",
+            },
+        }
+    
+    RETURN_TYPES = ("SIGMAS",)
+    RETURN_NAMES = ("adjusted_sigmas",)
+    FUNCTION = "adjust_sigmas"
+    CATEGORY = "sampling/custom_sampling/sigmas"
+    DESCRIPTION = "Interactively adjust sigmas curve by dragging control points"
+    
+    def adjust_sigmas(self, sigmas, sigmas_adjustments="[]", unique_id=None):
+        # Convert sigmas to numpy
+        if isinstance(sigmas, torch.Tensor):
+            sigmas_np = sigmas.cpu().numpy()
+        else:
+            sigmas_np = np.array(sigmas)
+        
+        # Parse adjusted sigma values from JSON
+        try:
+            adjusted_values = json.loads(sigmas_adjustments)
+        except:
+            adjusted_values = []
+        
+        # If no adjustments or length mismatch, use original sigmas
+        if len(adjusted_values) != len(sigmas_np):
+            adjusted_sigmas = sigmas_np.copy()
+        else:
+            # Use the adjusted sigma values directly
+            adjusted_sigmas = np.array(adjusted_values, dtype=np.float64)
+        
+        # Ensure last sigma is still 0 if original was 0
+        if sigmas_np[-1] == 0:
+            adjusted_sigmas[-1] = 0
+        
+        result_tensor = torch.FloatTensor(adjusted_sigmas)
+        
+        # Send sigmas data to frontend via PromptServer (只有输入sigmas改变时才发送)
+        if unique_id is not None:
+            # 只根据输入的sigmas创建缓存键（不包括adjustments）
+            current_sigmas_key = tuple(sigmas_np.tolist())
+            
+            # 检查是否与上次输入的sigmas相同
+            last_sigmas_key = self._last_sent_data.get(unique_id)
+            
+            # 只有输入sigmas改变时才发送数据到前端
+            if last_sigmas_key != current_sigmas_key:
+                PromptServer.instance.send_sync("sigmas_editor_update", {
+                    "node_id": unique_id,
+                    "sigmas_data": {
+                        "original": sigmas_np.tolist(),
+                        "adjusted": adjusted_sigmas.tolist(),
+                    }
+                })
+                
+                # 更新缓存（只缓存输入的sigmas）
+                self._last_sent_data[unique_id] = current_sigmas_key
+        
+        return (result_tensor,)
+
+
+NODE_CLASS_MAPPINGS = {
+    "SigmasEditor": SigmasEditor,
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "SigmasEditor": "Sigmas Editor 🎚️",
+}
+
diff --git a/py/timestep_noise.py b/py/timestep_noise.py
new file mode 100644
index 0000000..431d490
--- /dev/null
+++ b/py/timestep_noise.py
@@ -0,0 +1,211 @@
+"""
+ZImage/Lumina2 采样扰动节点
+
+这个节点可以在采样过程中注入噪声扰动，
+打破模型的同质化输出，使不同种子能产生更明显的差异。
+"""
+
+import torch
+import torch.nn.functional as F
+import logging
+
+
+class ZImageTimestepNoise:
+    """
+    对 timestep/sigma 添加噪声扰动
+    这会改变模型对当前去噪步骤的感知，产生不同的输出
+    
+    支持两种模式：
+    - sigma: 适用于传统扩散模型，使用乘性噪声
+    - flow: 适用于 Flow Matching 模型（如 ZImage/Lumina2），使用加性噪声
+    """
+    
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "model": ("MODEL",),
+                "sigmas": ("SIGMAS",),
+                "mode": (["sigma", "flow"], {
+                    "default": "flow",
+                    "tooltip": "sigma: 传统扩散模型（乘性噪声）; flow: Flow Matching 模型（加性噪声）"
+                }),
+                "noise_strength": ("FLOAT", {
+                    "default": 0.05, 
+                    "min": 0.0, 
+                    "max": 2.0, 
+                    "step": 0.01,
+                    "tooltip": "噪声强度"
+                }),
+                "seed": ("INT", {
+                    "default": 0, 
+                    "min": 0, 
+                    "max": 0xffffffffffffffff,
+                    "tooltip": "噪声种子"
+                }),
+                "start_percent": ("FLOAT", {
+                    "default": 0.0,
+                    "min": 0.0,
+                    "max": 1.0,
+                    "step": 0.01,
+                    "tooltip": "开始应用噪声的采样进度 (0.0 = 开始)"
+                }),
+                "end_percent": ("FLOAT", {
+                    "default": 0.5,
+                    "min": 0.0,
+                    "max": 1.0,
+                    "step": 0.01,
+                    "tooltip": "停止应用噪声的采样进度 (1.0 = 结束)"
+                }),
+            },
+            "optional": {
+                "mask": ("MASK",),
+            }
+        }
+
+    RETURN_TYPES = ("MODEL",)
+    FUNCTION = "patch"
+    CATEGORY = "advanced/model"
+    DESCRIPTION = "对 timestep 添加噪声扰动，改变模型对去噪步骤的感知。sigma 模式适用于传统扩散模型，flow 模式适用于 Flow Matching 模型（如 ZImage/Lumina2）。可选遮罩限制影响区域。"
+
+    def patch(self, model, sigmas, mode, noise_strength, seed, start_percent, end_percent, mask=None):
+        m = model.clone()
+        
+        if noise_strength <= 0:
+            return (m,)
+        
+        # 将 sigmas 转换为 list 以便查找
+        sigma_list = sigmas.tolist()
+        total_steps = len(sigma_list) - 1  # 最后一个是 0
+        
+        stored_mode = mode
+        stored_strength = noise_strength
+        stored_seed = seed
+        stored_start = start_percent
+        stored_end = end_percent
+        stored_sigmas = sigma_list
+        stored_total_steps = total_steps
+        stored_mask = mask
+        step_counter = [0]
+        
+        def unet_wrapper(apply_model_func, args):
+            input_x = args["input"]
+            timestep = args["timestep"]
+            c = args["c"]
+            
+            step_counter[0] += 1
+            current_step = step_counter[0]
+            
+            # 获取当前 sigma 值
+            sigma_val = float(timestep[0]) if timestep.dim() > 0 else float(timestep)
+            
+            # 在 sigma_list 中查找当前 sigma 的位置来确定进度
+            min_diff = float('inf')
+            matched_idx = 0
+            for i, s in enumerate(stored_sigmas):
+                diff = abs(s - sigma_val)
+                if diff < min_diff:
+                    min_diff = diff
+                    matched_idx = i
+            
+            # 计算进度
+            progress = matched_idx / stored_total_steps if stored_total_steps > 0 else 0.0
+            progress = max(0.0, min(1.0, progress))
+            
+            # 检查是否在指定范围内
+            in_range = progress >= stored_start and progress <= stored_end
+            
+            # 准备遮罩（如果有）
+            latent_mask = None
+            if stored_mask is not None:
+                mask_tensor = stored_mask
+                if mask_tensor.dim() == 2:
+                    mask_tensor = mask_tensor.unsqueeze(0)
+                latent_h, latent_w = input_x.shape[2], input_x.shape[3]
+                latent_mask = F.interpolate(
+                    mask_tensor.unsqueeze(1),
+                    size=(latent_h, latent_w),
+                    mode='bilinear',
+                    align_corners=False
+                )
+                if latent_mask.shape[0] == 1 and input_x.shape[0] > 1:
+                    latent_mask = latent_mask.expand(input_x.shape[0], -1, -1, -1)
+                latent_mask = latent_mask.to(device=input_x.device, dtype=input_x.dtype)
+            
+            # 计算噪声（无论是否在范围内都计算，用于调试显示）
+            generator = torch.Generator(device=timestep.device)
+            generator.manual_seed(stored_seed + current_step)
+            
+            t_orig = float(timestep[0]) if timestep.dim() > 0 else float(timestep)
+            
+            if stored_mode == "sigma":
+                noise_factor = 1.0 + (torch.rand(1, generator=generator, device=timestep.device).item() - 0.5) * stored_strength
+                noisy_timestep = timestep * noise_factor
+            else:  # flow 模式
+                # 计算可用空间
+                headroom_up = 1.0 - t_orig    # 向上的空间
+                headroom_down = t_orig - 0.0  # 向下的空间
+                
+                # 确定实际可用的噪声范围
+                actual_up = min(stored_strength, headroom_up)
+                actual_down = min(stored_strength, headroom_down)
+                
+                # 生成 [0, 1] 随机数，然后映射到 [-actual_down, +actual_up]
+                raw = torch.rand(1, generator=generator, device=timestep.device).item()
+                total_range = actual_down + actual_up
+                
+                if total_range > 1e-6:
+                    # 映射到可用范围：raw=0 -> -actual_down, raw=1 -> +actual_up
+                    actual_delta = raw * total_range - actual_down
+                else:
+                    actual_delta = 0.0
+                
+                noisy_timestep = timestep + actual_delta
+                # 安全 clamp
+                noisy_timestep = torch.clamp(noisy_timestep, 0.0, 1.0)
+            
+            # 获取数值用于日志
+            t_noisy = float(noisy_timestep[0]) if noisy_timestep.dim() > 0 else float(noisy_timestep)
+            delta = t_noisy - t_orig
+            
+            # 调试输出
+            if current_step <= 5:
+                status = "✓ APPLIED" if in_range else "✗ SKIPPED"
+                has_mask = "mask=YES" if latent_mask is not None else "mask=NO"
+                if stored_mode == "flow":
+                    logging.info(f"[ZImageTimestepNoise] step={current_step}/{stored_total_steps} | progress={progress:.2f} | range=[{stored_start:.2f}, {stored_end:.2f}] | {status}")
+                    logging.info(f"[ZImageTimestepNoise]   timestep: {t_orig:.6f} -> {t_noisy:.6f} (delta={delta:+.6f}) | headroom=[↓{headroom_down:.3f}, ↑{headroom_up:.3f}] | {has_mask}")
+                else:
+                    logging.info(f"[ZImageTimestepNoise] step={current_step}/{stored_total_steps} | progress={progress:.2f} | range=[{stored_start:.2f}, {stored_end:.2f}] | {status}")
+                    logging.info(f"[ZImageTimestepNoise]   timestep: {t_orig:.6f} -> {t_noisy:.6f} (delta={delta:+.6f}) | factor={noise_factor:.4f} | {has_mask}")
+            
+            if not in_range:
+                # 不在范围内，使用原始 timestep
+                return apply_model_func(input_x, timestep, **c)
+            
+            # 在范围内，应用噪声
+            if latent_mask is not None:
+                # 计算两种 timestep 下的结果并混合
+                result_original = apply_model_func(input_x, timestep, **c)
+                result_noisy = apply_model_func(input_x, noisy_timestep, **c)
+                result = result_original * (1 - latent_mask) + result_noisy * latent_mask
+                return result
+            else:
+                # 无遮罩，直接使用扰动后的 timestep
+                return apply_model_func(input_x, noisy_timestep, **c)
+            
+            return apply_model_func(input_x, timestep, **c)
+        
+        m.set_model_unet_function_wrapper(unet_wrapper)
+        
+        return (m,)
+
+
+# 注册节点
+NODE_CLASS_MAPPINGS = {
+    "ZImageTimestepNoise": ZImageTimestepNoise,
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "ZImageTimestepNoise": "ZImage Timestep Noise",
+}
diff --git a/web/sigmas_editor.js b/web/sigmas_editor.js
new file mode 100644
index 0000000..0c86e69
--- /dev/null
+++ b/web/sigmas_editor.js
@@ -0,0 +1,483 @@
+import { app } from "../../scripts/app.js";
+import { api } from "../../scripts/api.js";
+
+app.registerExtension({
+    name: "SigmasEditor.Interactive",
+    async beforeRegisterNodeDef(nodeType, nodeData) {
+        if (nodeData.name === "SigmasEditor") {
+            
+            // 扩展节点的构造函数
+            const onNodeCreated = nodeType.prototype.onNodeCreated;
+            nodeType.prototype.onNodeCreated = function() {
+                const result = onNodeCreated?.apply(this, arguments);
+                
+                // 初始化节点数据
+                this.sigmas_data = null;
+                this.adjustments = []; // 存储实际调整后的sigma值
+                this.dragging_point = -1;
+                this.isAdjusting = false;
+                
+                // 允许调整节点大小
+                this.resizable = true;
+                
+                // 设置初始大小（宽度自动，高度400）
+                this.size = [this.size[0] || 600, 400];
+                
+                // 设置WebSocket监听
+                this.setupWebSocket();
+                
+                return result;
+            };
+            
+            // 添加WebSocket设置方法
+            nodeType.prototype.setupWebSocket = function() {
+                const messageHandler = (event) => {
+                    const data = event.detail;
+                    
+                    if (!data || !data.node_id || !data.sigmas_data) {
+                        return;
+                    }
+                    
+                    // 通过node_id查找对应的节点
+                    const targetNode = app.graph.getNodeById(parseInt(data.node_id));
+                    
+                    // 检查是否是当前节点
+                    if (targetNode && targetNode === this) {
+                        this.sigmas_data = data.sigmas_data.original;
+                        // 从后端接收调整后的值，如果为空则使用原始值
+                        this.adjustments = data.sigmas_data.adjusted || data.sigmas_data.original.slice();
+                        
+                        // 第一次执行时，触发 sigmas_adjustments 组件更新
+                        this.updateAdjustmentsWidget();
+                        
+                        // 立即更新画布（强制重新计算尺寸）
+                        if (this.canvas) {
+                            this.updateCanvas(true);
+                        } else {
+                            // 如果画布还没创建，等待一下再尝试
+                            setTimeout(() => {
+                                if (this.canvas) {
+                                    this.updateCanvas(true);
+                                }
+                            }, 100);
+                        }
+                        this.setDirtyCanvas(true, true);
+                    }
+                };
+                
+                api.addEventListener("sigmas_editor_update", messageHandler);
+                
+                // 存储handler引用以便后续清理
+                this._sigmasEditorMessageHandler = messageHandler;
+            };
+            
+            // 添加节点时的处理
+            const onAdded = nodeType.prototype.onAdded;
+            nodeType.prototype.onAdded = function() {
+                const result = onAdded?.apply(this, arguments);
+                
+                if (!this.canvasContainer && this.id !== undefined && this.id !== -1) {
+                    // 创建画布容器（高度由节点自适应）
+                    const container = document.createElement("div");
+                    container.style.position = "relative";
+                    container.style.width = "100%";
+                    container.style.height = "100%";
+                    container.style.minHeight = "300px";
+                    container.style.backgroundColor = "#1e1e1e";
+                    container.style.borderRadius = "8px";
+                    container.style.overflow = "hidden";
+                    
+                    // 创建画布
+                    const canvas = document.createElement("canvas");
+                    canvas.style.width = "100%";
+                    canvas.style.height = "100%";
+                    canvas.style.cursor = "crosshair";
+                    
+                    container.appendChild(canvas);
+                    this.canvas = canvas;
+                    this.canvasContainer = container;
+                    
+                    // 添加鼠标事件监听
+                    this.addCanvasEventListeners();
+                    
+                    // 添加DOM组件
+                    this.widgets ||= [];
+                    this.widgets_up = true;
+                    
+                    requestAnimationFrame(() => {
+                        if (this.widgets) {
+                            this.canvasWidget = this.addDOMWidget("sigmas_canvas", "canvas", container);
+                            
+                            // 初始化时强制计算画布尺寸
+                            this.updateCanvas(true);
+                            this.setDirtyCanvas(true, true);
+                        }
+                    });
+                }
+                
+                return result;
+            };
+            
+            // 添加画布事件监听器
+            nodeType.prototype.addCanvasEventListeners = function() {
+                const canvas = this.canvas;
+                
+                // 获取缩放后的鼠标坐标
+                const getScaledMousePos = (e) => {
+                    const rect = canvas.getBoundingClientRect();
+                    // 计算缩放比例
+                    const scaleX = canvas.width / rect.width;
+                    const scaleY = canvas.height / rect.height;
+                    
+                    // 鼠标相对于canvas显示区域的位置
+                    const displayX = e.clientX - rect.left;
+                    const displayY = e.clientY - rect.top;
+                    
+                    // 转换为实际canvas坐标
+                    const canvasX = displayX * scaleX;
+                    const canvasY = displayY * scaleY;
+                    
+                    return { x: canvasX, y: canvasY };
+                };
+                
+                // 鼠标按下
+                canvas.addEventListener("mousedown", (e) => {
+                    const pos = getScaledMousePos(e);
+                    const pointIdx = this.findNearestPoint(pos.x, pos.y);
+                    
+                    if (pointIdx !== -1) {
+                        this.dragging_point = pointIdx;
+                        this.isAdjusting = true;
+                        this.updateCanvas(); // 只重绘，不resize
+                    }
+                });
+                
+                // 鼠标移动
+                canvas.addEventListener("mousemove", (e) => {
+                    if (this.dragging_point !== -1 && this.isAdjusting) {
+                        const pos = getScaledMousePos(e);
+                        this.updatePointAdjustment(this.dragging_point, pos.y);
+                        this.updateAdjustmentsWidget(); // 实时同步更新组件值
+                        this.updateCanvas(); // 只重绘，不resize
+                    }
+                });
+                
+                // 鼠标释放
+                canvas.addEventListener("mouseup", (e) => {
+                    if (this.dragging_point !== -1) {
+                        this.dragging_point = -1;
+                        this.isAdjusting = false;
+                        
+                        // 更新adjustments widget
+                        this.updateAdjustmentsWidget();
+                        this.updateCanvas(); // 只重绘，不resize
+                    }
+                });
+                
+                // 鼠标离开
+                canvas.addEventListener("mouseleave", (e) => {
+                    if (this.dragging_point !== -1) {
+                        this.dragging_point = -1;
+                        this.isAdjusting = false;
+                        this.updateAdjustmentsWidget();
+                        this.updateCanvas(); // 只重绘，不resize
+                    }
+                });
+            };
+            
+            // 查找最近的点
+            nodeType.prototype.findNearestPoint = function(mouseX, mouseY) {
+                if (!this.sigmas_data || this.sigmas_data.length === 0) return -1;
+                
+                const paddingLeft = 60;
+                const paddingTop = 50;
+                const paddingRight = 20;
+                const paddingBottom = 50;
+                
+                const canvas = this.canvas;
+                const chartWidth = canvas.width - paddingLeft - paddingRight;
+                const chartHeight = canvas.height - paddingTop - paddingBottom;
+                const chartX = paddingLeft;
+                const chartY = paddingTop;
+                
+                const steps = this.sigmas_data.length;
+                
+                let closestDist = Infinity;
+                let closestIdx = -1;
+                
+                for (let i = 0; i < steps; i++) {
+                    const x = chartX + (chartWidth / (steps - 1)) * i;
+                    // adjustments存储的是实际调整后的sigma值
+                    const adjustedValue = this.adjustments[i] !== undefined ? this.adjustments[i] : this.sigmas_data[i];
+                    const y = chartY + chartHeight - (adjustedValue * chartHeight);
+                    
+                    const dist = Math.sqrt(Math.pow(mouseX - x, 2) + Math.pow(mouseY - y, 2));
+                    if (dist < 15 && dist < closestDist) {
+                        closestDist = dist;
+                        closestIdx = i;
+                    }
+                }
+                
+                return closestIdx;
+            };
+            
+            // 更新点的调整值
+            nodeType.prototype.updatePointAdjustment = function(pointIdx, mouseY) {
+                if (!this.sigmas_data || pointIdx < 0 || pointIdx >= this.sigmas_data.length) return;
+                
+                const paddingTop = 50;
+                const paddingBottom = 50;
+                
+                const canvas = this.canvas;
+                const chartHeight = canvas.height - paddingTop - paddingBottom;
+                const chartY = paddingTop;
+                
+                // 计算新的sigma值（范围0-1）
+                const clampedY = Math.max(chartY, Math.min(chartY + chartHeight, mouseY));
+                const newSigmaValue = (chartY + chartHeight - clampedY) / chartHeight;
+                
+                // 直接存储调整后的sigma值，限制在0-1范围内
+                this.adjustments[pointIdx] = Math.max(0.0, Math.min(1.0, newSigmaValue));
+            };
+            
+            // 向上取整到指定小数位
+            nodeType.prototype.ceilToFixed = function(value, decimals) {
+                const multiplier = Math.pow(10, decimals);
+                return Math.ceil(value * multiplier) / multiplier;
+            };
+            
+            // 更新adjustments widget（向上取整到4位小数）
+            nodeType.prototype.updateAdjustmentsWidget = function() {
+                const widget = this.widgets?.find(w => w.name === "sigmas_adjustments");
+                if (widget) {
+                    // 对所有调整后的值进行向上取整到4位小数，并格式化为字符串保留末尾的0
+                    const formattedValues = this.adjustments.map(v => {
+                        const rounded = this.ceilToFixed(v, 4);
+                        return rounded.toFixed(4);
+                    });
+                    // 手动构建 JSON 数组字符串，保留4位小数格式
+                    widget.value = '[' + formattedValues.join(', ') + ']';
+                }
+            };
+            
+            // 更新画布
+            nodeType.prototype.updateCanvas = function(forceResize = false) {
+                if (!this.canvas) return;
+                
+                requestAnimationFrame(() => {
+                    const canvas = this.canvas;
+                    const ctx = canvas.getContext("2d");
+                    
+                    // 只在必要时重新设置画布尺寸
+                    if (forceResize || !this._canvasInitialized) {
+                        const rect = canvas.getBoundingClientRect();
+                        
+                        // 确保画布有合理的尺寸
+                        const width = rect.width > 0 ? rect.width : 600;
+                        const height = rect.height > 0 ? rect.height : 300;
+                        
+                        // 只有尺寸真正改变时才重新设置
+                        if (canvas.width !== width || canvas.height !== height) {
+                            canvas.width = width;
+                            canvas.height = height;
+                        }
+                        
+                        this._canvasInitialized = true;
+                    }
+                    
+                    // 增加padding以容纳坐标轴标签
+                    const paddingLeft = 60;
+                    const paddingRight = 20;
+                    const paddingTop = 50;
+                    const paddingBottom = 50;
+                    
+                    const chartWidth = canvas.width - paddingLeft - paddingRight;
+                    const chartHeight = canvas.height - paddingTop - paddingBottom;
+                    const chartX = paddingLeft;
+                    const chartY = paddingTop;
+                    
+                    // 清空画布
+                    ctx.fillStyle = "#1e1e1e";
+                    ctx.fillRect(0, 0, canvas.width, canvas.height);
+                    
+                    // 如果没有数据，显示提示
+                    if (!this.sigmas_data || this.sigmas_data.length === 0) {
+                        ctx.fillStyle = "#999";
+                        ctx.font = "14px Arial";
+                        ctx.textAlign = "center";
+                        ctx.fillText("Connect Sigmas Input & Execute Workflow", canvas.width / 2, canvas.height / 2);
+                        return;
+                    }
+                    
+                    // 绘制图表区域
+                    ctx.fillStyle = "#2a2a2a";
+                    ctx.fillRect(chartX, chartY, chartWidth, chartHeight);
+                    
+                    // 绘制网格
+                    ctx.strokeStyle = "#444";
+                    ctx.lineWidth = 1;
+                    for (let i = 0; i <= 10; i++) {
+                        const y = chartY + (chartHeight / 10) * i;
+                        ctx.beginPath();
+                        ctx.moveTo(chartX, y);
+                        ctx.lineTo(chartX + chartWidth, y);
+                        ctx.stroke();
+                    }
+                    
+                    const steps = this.sigmas_data.length;
+                    
+                    // 确保adjustments数组长度正确，初始化为原始sigma值
+                    if (this.adjustments.length !== steps) {
+                        this.adjustments = this.sigmas_data.slice();
+                    }
+                    
+                    // 绘制Y轴刻度和标签（Sigma值 0-1，刻度0.1）
+                    ctx.fillStyle = "#999";
+                    ctx.font = "10px Arial";
+                    ctx.textAlign = "right";
+                    for (let i = 0; i <= 10; i++) {
+                        const value = 1.0 - (i / 10);
+                        const y = chartY + (chartHeight / 10) * i;
+                        ctx.fillText(value.toFixed(1), chartX - 5, y + 3);
+                    }
+                    
+                    // Y轴标签
+                    ctx.save();
+                    ctx.translate(15, chartY + chartHeight / 2);
+                    ctx.rotate(-Math.PI / 2);
+                    ctx.textAlign = "center";
+                    ctx.font = "12px Arial";
+                    ctx.fillStyle = "#ccc";
+                    ctx.fillText("Sigma Value", 0, 0);
+                    ctx.restore();
+                    
+                    // 绘制X轴刻度和标签
+                    ctx.fillStyle = "#999";
+                    ctx.font = "10px Arial";
+                    ctx.textAlign = "center";
+                    
+                    // 根据步数决定显示哪些刻度
+                    let stepInterval = 1;
+                    if (steps > 30) stepInterval = 2;
+                    if (steps > 50) stepInterval = 5;
+                    if (steps > 100) stepInterval = 10;
+                    
+                    for (let i = 0; i < steps; i += stepInterval) {
+                        const x = chartX + (chartWidth / (steps - 1)) * i;
+                        ctx.fillText(i.toString(), x, chartY + chartHeight + 15);
+                    }
+                    // 确保显示最后一步
+                    if ((steps - 1) % stepInterval !== 0) {
+                        const x = chartX + chartWidth;
+                        ctx.fillText((steps - 1).toString(), x, chartY + chartHeight + 15);
+                    }
+                    
+                    // X轴标签
+                    ctx.font = "12px Arial";
+                    ctx.fillStyle = "#ccc";
+                    ctx.fillText("Steps", chartX + chartWidth / 2, chartY + chartHeight + 30);
+                    
+                    // 绘制曲线
+                    ctx.strokeStyle = "#4a9eff";
+                    ctx.lineWidth = 2;
+                    ctx.beginPath();
+                    
+                    for (let i = 0; i < steps; i++) {
+                        const x = chartX + (chartWidth / (steps - 1)) * i;
+                        // adjustments存储的是实际调整后的sigma值
+                        const adjustedValue = this.adjustments[i] !== undefined ? this.adjustments[i] : this.sigmas_data[i];
+                        // 限制在0-1范围内
+                        const clampedValue = Math.max(0, Math.min(1, adjustedValue));
+                        const y = chartY + chartHeight - (clampedValue * chartHeight);
+                        
+                        if (i === 0) {
+                            ctx.moveTo(x, y);
+                        } else {
+                            ctx.lineTo(x, y);
+                        }
+                    }
+                    ctx.stroke();
+                    
+                    // 绘制控制点
+                    for (let i = 0; i < steps; i++) {
+                        const x = chartX + (chartWidth / (steps - 1)) * i;
+                        const adjustedValue = this.adjustments[i] !== undefined ? this.adjustments[i] : this.sigmas_data[i];
+                        const clampedValue = Math.max(0, Math.min(1, adjustedValue));
+                        const y = chartY + chartHeight - (clampedValue * chartHeight);
+                        
+                        ctx.fillStyle = this.dragging_point === i ? "#ff6b6b" : "#4a9eff";
+                        ctx.beginPath();
+                        ctx.arc(x, y, 5, 0, Math.PI * 2);
+                        ctx.fill();
+                    }
+                    
+                    // 绘制标题
+                    ctx.fillStyle = "#ccc";
+                    ctx.font = "14px Arial";
+                    ctx.textAlign = "center";
+                    ctx.fillText("Sigmas Schedule Editor", canvas.width / 2, 20);
+                    
+                    // 显示当前拖拽点的信息
+                    if (this.dragging_point !== -1) {
+                        const orig = this.sigmas_data[this.dragging_point];
+                        const adjusted = this.adjustments[this.dragging_point];
+                        const multiplier = orig > 0 ? (adjusted / orig) : 1.0;
+                        
+                        // 向上取整到4位小数显示
+                        const origCeil = this.ceilToFixed(orig, 4);
+                        const adjustedCeil = this.ceilToFixed(adjusted, 4);
+                        
+                        ctx.fillStyle = "#ff6b6b";
+                        ctx.textAlign = "center";
+                        ctx.font = "11px Arial";
+                        ctx.fillText(
+                            `Step ${this.dragging_point}: Original=${origCeil.toFixed(4)}, Adjusted=${adjustedCeil.toFixed(4)}, Multiplier=${multiplier.toFixed(2)}x`,
+                            canvas.width / 2,
+                            35
+                        );
+                    }
+                });
+            };
+            
+            // 监听节点尺寸变化
+            const onResize = nodeType.prototype.onResize;
+            nodeType.prototype.onResize = function(size) {
+                const result = onResize?.apply(this, arguments);
+                
+                // 节点尺寸改变时，强制重新计算画布尺寸
+                if (this.canvas) {
+                    this._canvasInitialized = false; // 重置标志
+                    this.updateCanvas(true);
+                }
+                
+                return result;
+            };
+            
+            // 节点移除时的处理
+            const onRemoved = nodeType.prototype.onRemoved;
+            nodeType.prototype.onRemoved = function() {
+                const result = onRemoved?.apply(this, arguments);
+                
+                // 清理画布
+                if (this && this.canvas) {
+                    const ctx = this.canvas.getContext("2d");
+                    if (ctx) {
+                        ctx.clearRect(0, 0, this.canvas.width, this.canvas.height);
+                    }
+                    this.canvas = null;
+                }
+                if (this) {
+                    this.canvasContainer = null;
+                }
+                
+                // 清理事件监听器
+                if (this._sigmasEditorMessageHandler) {
+                    api.removeEventListener("sigmas_editor_update", this._sigmasEditorMessageHandler);
+                    this._sigmasEditorMessageHandler = null;
+                }
+                
+                return result;
+            };
+        }
+    }
+});