diff --git a/.features/skill/MEMORY.md b/.features/skill/MEMORY.md
index e202500..120efd1 100644
--- a/.features/skill/MEMORY.md
+++ b/.features/skill/MEMORY.md
@@ -18,16 +18,16 @@ Skill 系统支持两种来源：官方 skills (`./skills/`) 和用户 skills (`
 
 ## 最近重要事项
 
+- 2026-04-19: 环境变量 `SKILLS_SUBDIR` 重命名为 `PROJECT_NAME`，用于选择 `skills/{PROJECT_NAME}` 和 `skills/autoload/{PROJECT_NAME}` 目录
 - 2026-04-19: `create_robot_project` 的 autoload 去重和 stale 清理补强，autoload 目录也纳入 managed 清理，避免 `rag-retrieve-only` 场景下旧的 `rag-retrieve` 残留
-- 2026-04-18: `create_robot_project` 改为自动加载 `skills/autoload/{SKILLS_SUBDIR}` 下所有 skill，并跳过已显式传入的同名 skill
-- 2026-04-18: `/api/v1/skill/list` 的官方库改为同时读取 `skills/common` 和 `skills/{SKILLS_SUBDIR}`，并按目录顺序去重
-- 2026-04-18: `_extract_skills_to_robot` 改为通过环境变量 `SKILLS_SUBDIR` 选择官方 skills 子目录，默认使用 `skills/common`
+- 2026-04-18: `/api/v1/skill/list` 的官方库改为同时读取 `skills/common` 和 `skills/{PROJECT_NAME}`，并按目录顺序去重
+- 2026-04-18: `_extract_skills_to_robot` 改为通过环境变量 `PROJECT_NAME` 选择官方 skills 子目录，默认使用 `skills/common`
 - 2025-02-11: 初始化 skill 功能 memory
 
 ## Gotchas（开发必读）
 
 - ⚠️ `create_robot_project` 的 autoload 去重是“包含匹配”，只要传入的 skill 字符串里包含 autoload skill 名，就不会重复自动加载
-- ⚠️ `_extract_skills_to_robot` 只会从 `skills/{SKILLS_SUBDIR}` 读取官方 skills，默认是 `common`
+- ⚠️ `_extract_skills_to_robot` 只会从 `skills/{PROJECT_NAME}` 读取官方 skills，默认是 `common`
 - ⚠️ 执行脚本必须使用绝对路径
 - ⚠️ MCP 配置优先级：Skill MCP > 默认 MCP > 用户参数
 - ⚠️ 上传大小限制：50MB（ZIP），解压后最大 500MB
diff --git a/routes/skill_manager.py b/routes/skill_manager.py
index 9f52a1a..1617f2e 100644
--- a/routes/skill_manager.py
+++ b/routes/skill_manager.py
@@ -10,7 +10,7 @@ from typing import List, Optional
 from dataclasses import dataclass
 from fastapi import APIRouter, HTTPException, Query, UploadFile, File, Form
 from pydantic import BaseModel
-from utils.settings import SKILLS_DIR, SKILLS_SUBDIR
+from utils.settings import SKILLS_DIR, PROJECT_NAME
 import aiofiles
 
 logger = logging.getLogger('app')
@@ -438,7 +438,7 @@ def get_official_skills(base_dir: str) -> List[SkillItem]:
 
     official_skills_dirs = [
         os.path.join(skills_root_dir, "common"),
-        os.path.join(skills_root_dir, SKILLS_SUBDIR),
+        os.path.join(skills_root_dir, PROJECT_NAME),
     ]
 
     for official_skills_dir in official_skills_dirs:
@@ -511,7 +511,7 @@ async def list_skills(
         SkillListResponse containing all skills
 
     Notes:
-        - Official skills are read from /skills/common and /skills/{SKILLS_SUBDIR}
+        - Official skills are read from /skills/common and /skills/{PROJECT_NAME}
         - User skills are read from /projects/uploads/{bot_id}/skills directory
         - User skills are marked with user_skill: true
     """
diff --git a/utils/multi_project_manager.py b/utils/multi_project_manager.py
index 544c38e..64f7eb4 100644
--- a/utils/multi_project_manager.py
+++ b/utils/multi_project_manager.py
@@ -331,15 +331,15 @@ def create_robot_project(dataset_ids: List[str], bot_id: str, force_rebuild: boo
 
     skills = list(skills or [])
     if os.path.isabs(settings.SKILLS_DIR):
-        autoload_skills_dir = Path(settings.SKILLS_DIR) / "autoload" / settings.SKILLS_SUBDIR
+        autoload_skills_dir = Path(settings.SKILLS_DIR) / "autoload" / settings.PROJECT_NAME
     else:
-        autoload_skills_dir = project_path.parent / settings.SKILLS_DIR / "autoload" / settings.SKILLS_SUBDIR
+        autoload_skills_dir = project_path.parent / settings.SKILLS_DIR / "autoload" / settings.PROJECT_NAME
 
     if autoload_skills_dir.exists():
         for item in sorted(autoload_skills_dir.iterdir()):
             if not item.is_dir() or any(_skill_matches_autoload(skill, item.name) for skill in skills):
                 continue
-            skill_path = f"@skills/autoload/{settings.SKILLS_SUBDIR}/{item.name}"
+            skill_path = f"@skills/autoload/{settings.PROJECT_NAME}/{item.name}"
             skills.append(skill_path)
             logger.info(f"Auto loaded skill '{skill_path}' from {autoload_skills_dir}")
     else:
@@ -399,10 +399,10 @@ def _extract_skills_to_robot(bot_id: str, skills: List[str], project_path: Path)
     - 如果是完整路径（如 "projects/uploads/xxx/skills/rag-retrieve_2.zip"），直接使用该路径
     - 如果是简单名称（如 "rag-retrieve"），从以下目录按优先级顺序查找：
       1. projects/uploads/{bot_id}/skills/
-      2. skills/{SKILLS_SUBDIR}/
+      2. skills/{PROJECT_NAME}/
     - 如果是以 @ 开头的仓库相对路径（如 "@skills/autoload/support/rag-retrieve"），则从仓库根目录直接解析
 
-    搜索目录优先级：先搜索 projects/uploads/{bot_id}/skills/，再搜索 skills/common，最后搜索 skills/{SKILLS_SUBDIR}
+    搜索目录优先级：先搜索 projects/uploads/{bot_id}/skills/，再搜索 skills/common，最后搜索 skills/{PROJECT_NAME}
 
     Args:
         bot_id: 机器人 ID
@@ -411,8 +411,8 @@ def _extract_skills_to_robot(bot_id: str, skills: List[str], project_path: Path)
     """
     # skills 源目录（按优先级顺序）
     repo_root = Path(__file__).resolve().parent.parent
-    official_skills_dir = repo_root / "skills" / settings.SKILLS_SUBDIR
-    autoload_skills_dir = repo_root / "skills" / "autoload" / settings.SKILLS_SUBDIR
+    official_skills_dir = repo_root / "skills" / settings.PROJECT_NAME
+    autoload_skills_dir = repo_root / "skills" / "autoload" / settings.PROJECT_NAME
     if not official_skills_dir.exists():
         logger.warning(f"Official skills directory does not exist: {official_skills_dir}")
     skills_source_dirs = [
diff --git a/utils/settings.py b/utils/settings.py
index 10014f3..d90a714 100644
--- a/utils/settings.py
+++ b/utils/settings.py
@@ -24,7 +24,7 @@ TOOL_CACHE_AUTO_RENEW = os.getenv("TOOL_CACHE_AUTO_RENEW", "true") == "true"
 # Project Settings
 PROJECT_DATA_DIR = os.getenv("PROJECT_DATA_DIR", "./projects/data")
 SKILLS_DIR = os.getenv("SKILLS_DIR", "./skills")
-SKILLS_SUBDIR = os.getenv("SKILLS_SUBDIR", "support")
+PROJECT_NAME = os.getenv("PROJECT_NAME", "support")
 
 # Tokenizer Settings
 TOKENIZERS_PARALLELISM = os.getenv("TOKENIZERS_PARALLELISM", "true")