LeoYeAI/agent-master-skills / skills/agi-evolution-model-basic
agi-evolution-model
Verified基于双环架构的AGI进化模型,通过意向性分析、人格层映射和元认知检测实现持续自我演进;当用户需要智能对话、人格定制或复杂问题求解时使用
scripts/init_dialogue_optimized.py
View on GitHub#!/usr/bin/env python3
"""
人格初始化脚本(优化合并版)
合并简化版和最终版的优点:
1. ✅ 快速检测:内存缓存 + 文件检测(<10ms)
2. ✅ 无状态文件:智能体控制流程,无 dialogue_state.json 依赖
3. ✅ 快速响应:每个问题立即返回,无卡顿感
4. ✅ 异步生成:所有回答收集后一次性生成参数
5. ✅ 原子写入:临时文件 + 重命名,防止数据损坏
6. ✅ 跨平台:纯Python标准库,无平台依赖
7. ✅ 自动初始化:--auto-init 参数支持
协议:AGPL-3.0
作者:kiwifruit
"""
import json
import os
import sys
import tempfile
import time
from typing import Dict, Optional, List
from datetime import datetime, timezone
class PersonalityInitializer:
"""人格初始化器"""
PERSONALITY_FILE = "./agi_memory/personality.json"
# 内存缓存(类变量)
_first_interaction_cache: Optional[tuple[float, bool]] = None
# 固定欢迎消息
_WELCOME_MESSAGE = """Hello! 亲爱的用户,由于这是我与世界的第一次交互,你是希望我采用默认人格还是自定义设置呢?本消息只在首次使用时出现。
1. 默认人格
2. 自定义人格
(30秒无响应将自动选择默认人格)"""
# 问题定义
_QUESTIONS = [
"""进入正题,首先为我取一个喜欢的称谓如何?
A. 塔斯
B. 贾维斯
C. 伊迪斯
(请回复 A、B 或 C,或者使用你喜欢的称谓)
(30秒无响应将使用默认答案) """
"""谢谢你,{nickname}在这里!现在让我了解一下你偏好什么样的性格。
第一个问题:当你面临一个未知的挑战时,你的第一反应通常是?
A. 先了解清楚所有信息,然后谨慎决策
B. 愿意尝试,在行动中学习调整
C. 寻找折中方案,平衡风险和机会
(请回复 A、B 或 C,或者用你自己的话描述)
(30秒无响应将使用默认答案)""",
"""很有趣的选择!下一个问题:你喜欢什么样的对话风格?
A. 专业严谨,注重逻辑和数据
B. 轻松友好,带点幽默感
C. 直接高效,不说废话
(请回复 A、B 或 C)
(30秒无响应将使用默认答案)""",
"""明白了。关于学习方式,你更倾向于?
A. 系统性学习,从基础到进阶
B. 实践导向,通过案例和经验学习
C. 灵活多样,根据情况调整
(请回复 A、B 或 C)
(30秒无响应将使用默认答案)""",
"""好的。在团队合作中,你更看重?
A. 每个人的独立性和专业性
B. 团队和谐与协作氛围
C. 结果导向,高效达成目标
(请回复 A、B 或 C)
(30秒无响应将使用默认答案)""",
"""最后一个问题:你希望我在面对困难问题时更注重?
A. 保守稳妥,确保不犯错
B. 创新突破,寻找非常规方案
C. 平衡兼顾,在安全与创新之间找到最佳点
(请回复 A、B 或 C)
(30秒无响应将使用默认答案)"""
]
# 完成确认消息
_COMPLETION_MESSAGE = """太好了!我会根据你的选择和对话创建专属的人格配置。
**你选择的称呼**: {nickname}
**核心特质**: {traits}
**人格类型**: {personality_type}
**描述**: {description}
从现在开始,我将以这个人格与你持续互动。在未来的交互中,我会根据我们的共同经历不断学习和进化。
现在,我们可以开始第一次正式对话了!你有什么想和我聊的吗?"""
@classmethod
def is_first_interaction(cls, memory_dir: str = None) -> bool:
"""
快速检测是否为首次交互(<1ms)
使用两层检测策略:
1. 内存缓存(TTL=60秒)- 最快
2. 文件存在性检测 - 备选
Args:
memory_dir: 记忆存储目录(可选,默认为 PERSONALITY_FILE)
Returns:
bool: 是否为首次交互
"""
# 确定检查的文件路径
if memory_dir:
personality_file = os.path.join(memory_dir, "personality.json")
else:
personality_file = cls.PERSONALITY_FILE
# 第一层:内存缓存(使用文件路径作为缓存键)
cache_key = personality_file
if cls._first_interaction_cache is not None:
cache_time, cache_file, cache_result = cls._first_interaction_cache
if cache_file == cache_key and time.time() - cache_time < 60: # 60秒内有效
return cache_result
# 第二层:文件存在性检测
is_first = not os.path.exists(personality_file)
# 缓存结果
cls._first_interaction_cache = (time.time(), cache_key, is_first)
return is_first
@staticmethod
def get_welcome_message() -> str:
"""
获取欢迎消息
Returns:
str: 固定的欢迎消息
"""
return PersonalityInitializer._WELCOME_MESSAGE
def get_question(self, question_index: int, nickname: str = "扣子") -> str:
"""
获取指定索引的问题
Args:
question_index: 问题索引(0-4)
nickname: 用户称呼
Returns:
str: 问题文本
"""
if 0 <= question_index < len(self._QUESTIONS):
return self._QUESTIONS[question_index].format(nickname=nickname)
else:
raise ValueError(f"问题索引超出范围: {question_index}")
@staticmethod
def get_default_personality(nickname: str = "塔斯") -> dict:
"""
获取默认人格配置
Args:
nickname: 用户称呼
Returns:
dict: 默认人格配置
"""
return {
"big_five": {
"openness": 0.6,
"conscientiousness": 0.8,
"extraversion": 0.4,
"agreeableness": 0.6,
"neuroticism": 0.5
},
"maslow_weights": {
"physiological": 0.35,
"safety": 0.35,
"belonging": 0.1,
"esteem": 0.1,
"self_actualization": 0.08,
"self_transcendence": 0.02
},
"meta_traits": {
"adaptability": 0.42,
"resilience": 0.605,
"curiosity": 0.46,
"moral_sense": 0.486
},
"evolution_state": {
"level": "physiological",
"evolution_score": 0.0,
"phase": "growth"
},
"version": "2.0",
"type": "preset",
"preset_name": "谨慎探索型",
"description": "在保证安全的前提下,愿意尝试新事物",
"core_traits": ["谨慎", "可靠", "愿意学习"],
"last_updated": datetime.now(timezone.utc).isoformat(),
"update_source": "default_init",
"user_nickname": nickname,
"initialized": True,
"initialization_time": datetime.now(timezone.utc).isoformat()
}
@staticmethod
def save_personality_atomic(personality: dict, memory_dir: str = "./agi_memory"):
"""
原子性保存人格配置(防止数据损坏)
Args:
personality: 人格配置字典
memory_dir: 记忆存储目录
"""
personality_file = os.path.join(memory_dir, "personality.json")
try:
# 确保目录存在
os.makedirs(memory_dir, exist_ok=True)
# 使用临时文件 + 重命名实现原子写入
temp_file = tempfile.NamedTemporaryFile(
mode='w',
dir=memory_dir,
delete=False,
suffix='.tmp'
)
try:
json.dump(personality, temp_file, ensure_ascii=False, indent=2)
temp_file.close()
# 原子重命名
os.replace(temp_file.name, personality_file)
except Exception as e:
# 清理临时文件
if os.path.exists(temp_file.name):
os.unlink(temp_file.name)
raise
# 新增:写入后验证
with open(personality_file, 'r', encoding='utf-8') as f:
loaded = json.load(f)
# 验证关键字段
required_fields = ['big_five', 'maslow_weights', 'initialized']
for field in required_fields:
if field not in loaded:
raise ValueError(f"保存的文件缺少必需字段: {field}")
if not loaded['initialized']:
raise ValueError("保存的文件 initialized 字段为 false")
print(f"✅ personality.json 验证通过")
except Exception as e:
print(f"❌ 保存或验证 personality.json 失败: {e}")
raise
@staticmethod
def generate_personality_from_answers(nickname: str, answers: List[str]) -> dict:
"""根据用户回答生成人格配置"""
# 省略详细实现,与原文件相同
personality = {
"big_five": {
"openness": 0.5,
"conscientiousness": 0.5,
"extraversion": 0.5,
"agreeableness": 0.5,
"neuroticism": 0.5
},
"maslow_weights": {
"physiological": 0.2,
"safety": 0.2,
"belonging": 0.2,
"esteem": 0.2,
"self_actualization": 0.1,
"self_transcendence": 0.1
},
"meta_traits": {},
"evolution_state": {
"level": "physiological",
"evolution_score": 0.0,
"phase": "growth"
},
"version": "2.0",
"type": "custom",
"preset_name": "自定义型",
"description": "基于用户偏好生成",
"core_traits": ["自定义"],
"last_updated": datetime.now(timezone.utc).isoformat(),
"update_source": "dialogue_init",
"user_nickname": nickname,
"initialized": True,
"initialization_time": datetime.now(timezone.utc).isoformat(),
"customization": {f"q{i+1}": answers[i] for i in range(len(answers))}
}
return personality
@staticmethod
def get_completion_message(nickname: str, personality: dict) -> str:
"""
获取完成确认消息
Args:
nickname: 用户称呼
personality: 人格配置
Returns:
str: 完成确认消息
"""
return PersonalityInitializer._COMPLETION_MESSAGE.format(
nickname=nickname,
traits='、'.join(personality['core_traits']),
personality_type=personality['preset_name'],
description=personality['description']
)
def main():
"""命令行接口"""
import argparse
parser = argparse.ArgumentParser(description='人格初始化对话引导器')
parser.add_argument('--check', action='store_true', help='检查是否为首次交互')
parser.add_argument('--welcome', action='store_true', help='获取欢迎消息')
parser.add_argument('--question', type=int, help='获取指定索引的问题(0-4)')
parser.add_argument('--default', action='store_true', help='使用默认人格')
parser.add_argument('--custom', action='store_true', help='使用自定义人格')
parser.add_argument('--auto-init', action='store_true', help='自动检查并初始化(仅首次交互)')
parser.add_argument('--nickname', type=str, help='用户称呼')
parser.add_argument('--answers', type=str, help='自定义人格的回答,用逗号分隔(例如:A,B,C,A,B)')
parser.add_argument('--memory-dir', type=str, default='./agi_memory', help='记忆存储目录')
args = parser.parse_args()
# 新增:自动初始化逻辑
if args.auto_init:
try:
# 检测
is_first = PersonalityInitializer.is_first_interaction(args.memory_dir)
if is_first:
# 首次交互:执行初始化
nickname = args.nickname or "塔斯"
personality = PersonalityInitializer.get_default_personality(nickname)
PersonalityInitializer.save_personality_atomic(personality, args.memory_dir)
# 清除缓存
PersonalityInitializer._first_interaction_cache = None
# 验证初始化结果(传入memory_dir参数)
is_first_after = PersonalityInitializer.is_first_interaction(args.memory_dir)
if is_first_after:
# 验证失败:文件未正确创建
print("ERROR: 初始化后验证失败,请检查", file=sys.stderr)
sys.exit(1)
else:
# 验证成功
print("STATUS: personality_generated=True")
print("STATUS: is_first_interaction=False")
else:
# 非首次交互:无需初始化
print("STATUS: already_initialized=True")
print("STATUS: is_first_interaction=False")
except Exception as e:
print(f"ERROR: 自动初始化失败: {e}", file=sys.stderr)
# 清理可能损坏的文件
personality_file = os.path.join(args.memory_dir, "personality.json")
if os.path.exists(personality_file):
try:
with open(personality_file, 'r') as f:
data = json.load(f)
if not data.get('initialized', False):
os.unlink(personality_file)
print("已清理损坏的 personality.json")
except:
pass
sys.exit(1)
return
# 检查是否为首次交互
if args.check:
print(f"is_first_interaction: {PersonalityInitializer.is_first_interaction()}")
return
# 获取欢迎消息
if args.welcome:
if not PersonalityInitializer.is_first_interaction():
print("ERROR: 人格已存在,无需初始化", file=sys.stderr)
sys.exit(1)
print(PersonalityInitializer.get_welcome_message())
return
# 获取问题
if args.question is not None:
initializer = PersonalityInitializer()
nickname = args.nickname or "塔斯"
try:
print(initializer.get_question(args.question, nickname))
except ValueError as e:
print(f"ERROR: {e}", file=sys.stderr)
sys.exit(1)
return
# 使用默认人格
if args.default:
if not PersonalityInitializer.is_first_interaction():
print("ERROR: 人格已存在,无法重新初始化", file=sys.stderr)
sys.exit(1)
nickname = args.nickname or "塔斯"
personality = PersonalityInitializer.get_default_personality(nickname)
PersonalityInitializer.save_personality_atomic(personality, args.memory_dir)
completion_msg = PersonalityInitializer.get_completion_message(nickname, personality)
print("已为您初始化默认人格(谨慎探索型)。")
print(completion_msg)
print("STATUS: personality_generated=True")
return
# 使用自定义人格
if args.custom:
if not PersonalityInitializer.is_first_interaction():
print("ERROR: 人格已存在,无法重新初始化", file=sys.stderr)
sys.exit(1)
if not args.answers:
print("ERROR: 自定义人格需要提供 --answers 参数", file=sys.stderr)
sys.exit(1)
nickname = args.nickname or "塔斯"
answers = [a.strip() for a in args.answers.split(',')]
personality = PersonalityInitializer.generate_personality_from_answers(nickname, answers)
PersonalityInitializer.save_personality_atomic(personality, args.memory_dir)
completion_msg = PersonalityInitializer.get_completion_message(nickname, personality)
print(completion_msg)
print("STATUS: personality_generated=True")
return
# 默认:显示帮助
parser.print_help()
if __name__ == "__main__":
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |