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/advice_pool.py
View on GitHub#!/usr/bin/env python3
"""
建议池模块(Advice Pool)
功能:
- 存储软调节建议
- 查询软调节建议(由主循环调用)
- 记录采纳情况(由主循环调用)
- 清理过期建议
- 基于人格的个性化排序
设计原则:
- 超然存在:外环写入,主循环读取(单向流)
- 不强制执行:主循环自主决定是否采纳
- 时效性约束:过期建议自动清理
- 个性化排序:基于人格特质和采纳历史
数据结构:
- 建议:{id, content, priority, confidence, valid_until, vertex, based_on_intentionality, adoption_stats}
- 按顶点分类:drive, math, iteration
- 采纳统计:presented, adopted, effectiveness_sum
"""
import argparse
import json
import sys
import uuid
from datetime import datetime, timedelta
from typing import Dict, Any, List, Optional
class AdvicePool:
"""建议池"""
def __init__(self, memory_dir: str = "./agi_memory"):
self.memory_dir = memory_dir
self.pool_file = f"{memory_dir}/advice_pool.json"
# 建议池结构
self.pool = {
"drive": [], # 针对得不到顶点的建议
"math": [], # 针对数学顶点的建议
"iteration": [] # 针对自我迭代顶点的建议
}
# 元数据
self.metadata = {
"last_updated": None,
"total_suggestions": 0,
"active_suggestions": 0
}
# 加载现有建议池
self._load()
def add_suggestion(self, vertex: str, suggestion: Dict[str, Any]) -> str:
"""
添加软调节建议
Args:
vertex: 目标顶点(drive/math/iteration)
suggestion: 建议内容
- content: 建议内容
- priority: 优先级(0.0-1.0)
- confidence: 置信度(0.0-1.0)
- based_on_intentionality: 基于的意向性数据
Returns:
建议ID
"""
if vertex not in self.pool:
raise ValueError(f"Invalid vertex: {vertex}. Must be one of: drive, math, iteration")
suggestion_with_metadata = {
"id": str(uuid.uuid4()),
"generation_time": datetime.now().isoformat(),
"valid_until": (datetime.now() + timedelta(minutes=30)).isoformat(),
"vertex": vertex,
"content": suggestion.get('content', ''),
"priority": suggestion.get('priority', 0.5),
"confidence": suggestion.get('confidence', 0.5),
"based_on_intentionality": suggestion.get('based_on_intentionality', {}),
"adoption_stats": {
"presented": 0,
"adopted": 0,
"effectiveness_sum": 0.0
}
}
self.pool[vertex].append(suggestion_with_metadata)
# 清理过期建议
self._cleanup_expired()
# 更新元数据
self._update_metadata()
# 保存
self._save()
return suggestion_with_metadata['id']
def query_suggestions(self, vertex: str, context: Dict[str, Any] = None,
personality: Dict[str, Any] = None,
limit: int = 3) -> List[Dict[str, Any]]:
"""
查询软调节建议(由主循环调用)
Args:
vertex: 目标顶点(drive/math/iteration)
context: 上下文信息
personality: 人格数据
limit: 返回建议数量限制
Returns:
建议列表(已排序)
"""
# 1. 清理过期建议
self._cleanup_expired()
# 2. 过滤顶点
vertex_suggestions = self.pool.get(vertex, [])
if not vertex_suggestions:
return []
# 3. 过滤相关性
relevant_suggestions = [
s for s in vertex_suggestions
if self._check_relevance(s, context) > 0.5
]
if not relevant_suggestions:
return []
# 4. 基于人格排序
if personality:
scored_suggestions = [
{
**s,
"personalized_score": self._calculate_personalized_score(
s,
personality
)
}
for s in relevant_suggestions
]
else:
scored_suggestions = [
{
**s,
"personalized_score": s['priority'] * s['confidence']
}
for s in relevant_suggestions
]
# 按个性化得分排序
scored_suggestions.sort(
key=lambda x: x['personalized_score'],
reverse=True
)
# 5. 返回Top N
top_suggestions = scored_suggestions[:limit]
# 6. 更新呈现次数
for s in top_suggestions:
s['adoption_stats']['presented'] += 1
# 保存更新
self._save()
return top_suggestions
def record_adoption(self, suggestion_id: str, adopted: bool,
effectiveness: Optional[float] = None) -> bool:
"""
记录采纳情况(由主循环调用)
Args:
suggestion_id: 建议ID
adopted: 是否采纳
effectiveness: 有效性评分(0.0-1.0,可选)
Returns:
是否成功记录
"""
# 在所有顶点中查找建议
found = False
for vertex, suggestions in self.pool.items():
for suggestion in suggestions:
if suggestion['id'] == suggestion_id:
found = True
if adopted:
suggestion['adoption_stats']['adopted'] += 1
if effectiveness is not None:
suggestion['adoption_stats']['effectiveness_sum'] += effectiveness
break
if found:
break
if found:
self._update_metadata()
self._save()
return found
def get_stats(self) -> Dict[str, Any]:
"""
获取建议池统计信息
Returns:
统计信息
"""
stats = {
"metadata": self.metadata,
"by_vertex": {}
}
for vertex, suggestions in self.pool.items():
if suggestions:
total_presented = sum(s['adoption_stats']['presented'] for s in suggestions)
total_adopted = sum(s['adoption_stats']['adopted'] for s in suggestions)
total_effectiveness = sum(s['adoption_stats']['effectiveness_sum'] for s in suggestions)
adoption_rate = total_adopted / max(total_presented, 1)
avg_effectiveness = total_effectiveness / max(total_adopted, 1)
stats["by_vertex"][vertex] = {
"active_suggestions": len(suggestions),
"total_presented": total_presented,
"total_adopted": total_adopted,
"adoption_rate": adoption_rate,
"avg_effectiveness": avg_effectiveness
}
return stats
def clear_expired(self) -> int:
"""
清理所有过期建议
Returns:
清理的建议数量
"""
before_count = sum(len(suggestions) for suggestions in self.pool.values())
self._cleanup_expired()
self._save()
after_count = sum(len(suggestions) for suggestions in self.pool.values())
return before_count - after_count
def _load(self):
"""从文件加载建议池"""
try:
with open(self.pool_file, 'r', encoding='utf-8') as f:
data = json.load(f)
self.pool = data.get("pool", self.pool)
self.metadata = data.get("metadata", self.metadata)
except (FileNotFoundError, json.JSONDecodeError):
# 文件不存在或格式错误,使用默认值
pass
def _save(self):
"""保存建议池到文件"""
data = {
"pool": self.pool,
"metadata": self.metadata
}
with open(self.pool_file, 'w', encoding='utf-8') as f:
json.dump(data, f, ensure_ascii=False, indent=2)
def _cleanup_expired(self):
"""清理过期建议"""
current_time = datetime.now()
for vertex, suggestions in self.pool.items():
self.pool[vertex] = [
s for s in suggestions
if datetime.fromisoformat(s['valid_until']) > current_time
]
def _update_metadata(self):
"""更新元数据"""
self.metadata['last_updated'] = datetime.now().isoformat()
self.metadata['total_suggestions'] = sum(
len(suggestions) for suggestions in self.pool.values()
)
self.metadata['active_suggestions'] = sum(
len(suggestions) for suggestions in self.pool.values()
)
def _check_relevance(self, suggestion: Dict[str, Any],
context: Dict[str, Any] = None) -> float:
"""
检查建议与上下文的相关性
Args:
suggestion: 建议
context: 上下文
Returns:
相关性得分(0.0-1.0)
"""
if not context:
return 0.5
based_on = suggestion.get('based_on_intentionality', {})
context_intentionality = context.get('intentionality', {})
if not based_on or not context_intentionality:
return 0.5
# 计算匹配度
match_score = 0.0
dimensions = ['agent', 'direction', 'content']
for dimension in dimensions:
if based_on.get(dimension) == context_intentionality.get(dimension):
match_score += 1.0 / len(dimensions)
return match_score
def _calculate_personalized_score(self, suggestion: Dict[str, Any],
personality: Dict[str, Any]) -> float:
"""
计算基于人格的个性化得分
Args:
suggestion: 建议
personality: 人格数据
Returns:
个性化得分
"""
base_score = suggestion['priority'] * suggestion['confidence']
# 基于采纳历史调整
adoption_rate = suggestion['adoption_stats']['adopted'] / max(
suggestion['adoption_stats']['presented'],
1
)
# 基于人格特质调整
conscientiousness = personality.get('conscientiousness', 0.5)
openness = personality.get('openness', 0.5)
# 谨慎性越高,越倾向于采纳历史成功的建议
personality_factor = conscientiousness * 0.3 + openness * 0.2
personalized_score = base_score * (1 + adoption_rate) * (1 + personality_factor)
return personalized_score
def main():
parser = argparse.ArgumentParser(description="建议池模块")
parser.add_argument("--memory-dir", default="./agi_memory", help="记忆目录")
parser.add_argument("--action", choices=["add", "query", "record", "stats", "clear"],
required=True, help="操作类型")
parser.add_argument("--vertex", choices=["drive", "math", "iteration"],
help="目标顶点")
parser.add_argument("--suggestion", help="建议内容(JSON)")
parser.add_argument("--suggestion-id", help="建议ID")
parser.add_argument("--adopted", type=bool, help="是否采纳")
parser.add_argument("--effectiveness", type=float, help="有效性评分")
parser.add_argument("--context", help="上下文(JSON)")
parser.add_argument("--personality", help="人格数据(JSON文件)")
parser.add_argument("--limit", type=int, default=3, help="返回建议数量限制")
parser.add_argument("--output", help="输出文件")
args = parser.parse_args()
# 创建建议池
pool = AdvicePool(memory_dir=args.memory_dir)
if args.action == "add":
# 添加建议
if not args.vertex or not args.suggestion:
print("Error: --vertex and --suggestion are required for add action")
sys.exit(1)
suggestion = json.loads(args.suggestion)
suggestion_id = pool.add_suggestion(args.vertex, suggestion)
result = {
"success": True,
"suggestion_id": suggestion_id,
"message": "建议已添加到建议池"
}
elif args.action == "query":
# 查询建议
if not args.vertex:
print("Error: --vertex is required for query action")
sys.exit(1)
context = json.loads(args.context) if args.context else None
personality = None
if args.personality:
with open(args.personality, 'r', encoding='utf-8') as f:
personality = json.load(f)
suggestions = pool.query_suggestions(
args.vertex,
context,
personality,
args.limit
)
result = {
"success": True,
"vertex": args.vertex,
"count": len(suggestions),
"suggestions": suggestions
}
elif args.action == "record":
# 记录采纳情况
if not args.suggestion_id or args.adopted is None:
print("Error: --suggestion-id and --adopted are required for record action")
sys.exit(1)
success = pool.record_adoption(
args.suggestion_id,
args.adopted,
args.effectiveness
)
result = {
"success": success,
"message": "采纳情况已记录" if success else "建议ID不存在"
}
elif args.action == "stats":
# 获取统计信息
result = {
"success": True,
"stats": pool.get_stats()
}
elif args.action == "clear":
# 清理过期建议
cleared_count = pool.clear_expired()
result = {
"success": True,
"cleared_count": cleared_count,
"message": f"已清理 {cleared_count} 条过期建议"
}
# 输出结果
if args.output:
with open(args.output, 'w', encoding='utf-8') as f:
json.dump(result, f, ensure_ascii=False, indent=2)
else:
print(json.dumps(result, ensure_ascii=False, indent=2))
sys.exit(0)
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 |