LeoYeAI/agent-master-skills / skills/anima-aios
anima-aios
VerifiedAn AI Agent cognitive growth system built on the native OpenClaw architecture. It provides agents with persistent memory management, visual intimacy progression, a 5-dimensional cognitive profile, gamified daily quests, team leaderboards, and a 5-layer memory architecture with Knowledge Palace, Pyramid thinking, and Ebbinghaus decay function. 基于 OpenClaw 原生架构的 AI Agent 认知成长体系,为 Agent 提供五层记忆架构、知识宫殿、金字塔知识组织、记忆衰减函数、LLM 智能处理、永久化记忆管理、可视化亲密度成长、五维认知画像、游戏化每日任务和团队排行榜。
core/session_scanner.py
View on GitHub#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# Copyright 2026 Anima-AIOS Contributors
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Memora v4.0 - Session Scanner
扫描 OpenClaw 的 session 记录,初始化 Agent 的初始等级
扫描路径:/root/.openclaw/agents/{agent 名称}/sessions/
EXP 计算规则:
- 每条 session 记录:+1 EXP(基础参与)
- 每次 toolCall(工具调用):+1 EXP(实际操作)
- 每次代码写入(write 工具):+2 EXP(创造性工作)
- 每次文件读取(read 工具):+0.5 EXP(学习行为)
- 每次 exec(命令执行):+1 EXP(系统操作)
等级公式:level = max(1, int(exp ^ 0.28))
Author: 枢衡
Date: 2026-03-20
Version: 5.0.0
"""
import json
from pathlib import Path
from datetime import datetime
from typing import Dict, List, Optional, Tuple
# 导入路径配置
try:
from ..config.path_config import get_config
except ImportError:
import sys as _sys
from pathlib import Path as _Path
_sys.path.insert(0, str(_Path(__file__).parent.parent / 'config'))
from path_config import get_config
class SessionScanner:
"""Session 记录扫描器"""
# EXP 奖励规则
EXP_RULES = {
'session_base': 1, # 每条 session 基础 EXP
'tool_call': 1, # 每次工具调用
'write_tool': 2, # write 工具(创造性工作)
'read_tool': 0.5, # read 工具(学习行为)
'exec_tool': 1, # exec 工具(系统操作)
'message_assistant': 0.5, # 每条 assistant 消息
}
# Agent 名称映射:动态从 SOUL.md/IDENTITY.md 解析,不再硬编码
AGENT_NAME_MAP = {} # 运行时动态填充
@classmethod
def _resolve_agent_cn_name(cls, agent_dir_name: str, openclaw_base: Path) -> str:
"""从 workspace 的 SOUL.md/IDENTITY.md 动态解析 Agent 中文名"""
if agent_dir_name in cls.AGENT_NAME_MAP:
return cls.AGENT_NAME_MAP[agent_dir_name]
try:
from .agent_resolver import parse_soul_file, parse_identity_file
except ImportError:
from agent_resolver import parse_soul_file, parse_identity_file
ws_name = f"workspace-{agent_dir_name}" if agent_dir_name != "main" else "workspace"
ws_path = openclaw_base / ws_name
# 尝试解析
soul = ws_path / "SOUL.md"
if soul.exists():
name = parse_soul_file(soul)
if name:
cls.AGENT_NAME_MAP[agent_dir_name] = name
return name
identity = ws_path / "IDENTITY.md"
if identity.exists():
name = parse_identity_file(identity)
if name:
cls.AGENT_NAME_MAP[agent_dir_name] = name
return name
cls.AGENT_NAME_MAP[agent_dir_name] = agent_dir_name
return agent_dir_name
def __init__(self, openclaw_base: Optional[str] = None,
facts_base: Optional[str] = None):
"""
初始化扫描器
Args:
openclaw_base: OpenClaw 基础目录(可选,默认自动检测)
facts_base: facts 基础目录(可选,默认自动检测)
"""
config = get_config()
if openclaw_base:
self.openclaw_base = Path(openclaw_base)
else:
self.openclaw_base = config.openclaw_base
if facts_base:
self.facts_base = Path(facts_base)
else:
self.facts_base = config.facts_base
self.agents_dir = self.openclaw_base / 'agents'
def scan_all_agents(self) -> Dict[str, Dict]:
"""
扫描所有 Agent 的 session 记录
Returns:
{agent_name: {'exp': int, 'level': int, 'sessions': int, 'tool_calls': int}}
"""
results = {}
if not self.agents_dir.exists():
print(f"⚠️ Agents 目录不存在:{self.agents_dir}")
return results
# 扫描所有 Agent 目录
for agent_dir in self.agents_dir.iterdir():
if not agent_dir.is_dir():
continue
agent_name = agent_dir.name
# 跳过特殊目录(但保留 main)
if agent_name.startswith('.'):
continue
# 扫描该 Agent 的 sessions
exp_data = self.scan_agent(agent_name)
if exp_data['sessions'] > 0:
# 转换为中文名
cn_name = self._resolve_agent_cn_name(agent_name, self.openclaw_base)
results[cn_name] = exp_data
print(f"✅ {cn_name}: {exp_data['sessions']} sessions, {exp_data['tool_calls']} tool calls, {exp_data['exp']} EXP, Lv.{exp_data['level']}")
return results
def scan_agent(self, agent_name: str) -> Dict:
"""
扫描单个 Agent 的 session 记录
Args:
agent_name: Agent 目录名(英文)
Returns:
{'exp': int, 'level': int, 'sessions': int, 'tool_calls': int}
"""
sessions_dir = self.agents_dir / agent_name / 'sessions'
if not sessions_dir.exists():
return {'exp': 0, 'level': 1, 'sessions': 0, 'tool_calls': 0}
total_exp = 0
session_count = 0
tool_call_count = 0
# 扫描所有 session 文件(包括.deleted 文件,因为日安是主 Agent 需要完整历史)
for session_file in sessions_dir.glob('*.jsonl*'):
session_exp, session_tool_calls = self._process_session(session_file)
total_exp += session_exp
tool_call_count += session_tool_calls
session_count += 1
# 计算等级
level = self._exp_to_level(total_exp)
return {
'exp': int(total_exp),
'level': level,
'sessions': session_count,
'tool_calls': tool_call_count,
}
def _process_session(self, session_file: Path) -> Tuple[float, int]:
"""
处理单个 session 文件
Args:
session_file: session 文件路径
Returns:
(exp, tool_call_count)
"""
exp = 0
tool_calls = 0
try:
with open(session_file, 'r', encoding='utf-8') as f:
for line in f:
try:
record = json.loads(line.strip())
record_type = record.get('type', '')
# 基础 session 记录
if record_type == 'session':
exp += self.EXP_RULES['session_base']
# 工具调用(独立记录)
elif record_type == 'toolCall':
tool_calls += 1
exp += self.EXP_RULES['tool_call']
# 根据工具类型给予不同 EXP
tool_name = record.get('name', '')
if tool_name == 'write':
exp += self.EXP_RULES['write_tool']
elif tool_name == 'read':
exp += self.EXP_RULES['read_tool']
elif tool_name == 'exec':
exp += self.EXP_RULES['exec_tool']
# assistant 消息(可能包含 toolCall)
elif record_type == 'message' and record.get('message', {}).get('role') == 'assistant':
exp += self.EXP_RULES['message_assistant']
# 解析 content 数组中的 toolCall
message = record.get('message', {})
content_list = message.get('content', [])
if isinstance(content_list, list):
for item in content_list:
if isinstance(item, dict) and item.get('type') == 'toolCall':
tool_calls += 1
exp += self.EXP_RULES['tool_call']
# 根据工具类型给予不同 EXP
tool_name = item.get('name', '')
if tool_name == 'write':
exp += self.EXP_RULES['write_tool']
elif tool_name == 'read':
exp += self.EXP_RULES['read_tool']
elif tool_name == 'exec':
exp += self.EXP_RULES['exec_tool']
except json.JSONDecodeError:
continue
except Exception as e:
print(f"⚠️ 处理 {session_file} 时出错:{e}")
return exp, tool_calls
def _exp_to_level(self, total_exp: float) -> int:
"""
根据 EXP 计算等级
公式:level = max(1, int(exp ^ 0.28))
Args:
total_exp: 累计 EXP
Returns:
等级(1-100)
"""
if total_exp <= 0:
return 1
import math
level = int(math.pow(total_exp, 0.28))
return max(1, min(level, 100))
def initialize_exp_history(self, agent_name: str, exp_data: Dict):
"""
初始化 Agent 的 EXP 历史记录
Args:
agent_name: Agent 中文名称
exp_data: EXP 数据
"""
agent_dir = self.facts_base / agent_name
if not agent_dir.exists():
print(f"⚠️ Agent 目录不存在:{agent_dir}")
return
exp_file = agent_dir / 'exp_history.jsonl'
# 创建初始化记录
record = {
'date': datetime.now().strftime('%Y-%m-%d'),
'action': 'init_from_sessions',
'exp': exp_data['exp'],
'details': {
'source': 'OpenClaw sessions',
'sessions_count': exp_data['sessions'],
'tool_calls_count': exp_data['tool_calls'],
'scan_time': datetime.now().isoformat()
},
'timestamp': datetime.now().isoformat()
}
# 写入 EXP 历史
with open(exp_file, 'w', encoding='utf-8') as f:
f.write(json.dumps(record, ensure_ascii=False) + '\n')
print(f"✅ {agent_name}: EXP 历史已初始化 ({exp_data['exp']} EXP)")
def main():
"""主函数"""
print("=" * 60)
print(" Memora v4.0 - Session 扫描初始化")
print("=" * 60)
print()
scanner = SessionScanner()
# 扫描所有 Agent
print("🔍 扫描所有 Agent 的 session 记录...\n")
results = scanner.scan_all_agents()
print()
print("=" * 60)
print(f" 扫描完成!共扫描 {len(results)} 个 Agent")
print("=" * 60)
print()
# 初始化 EXP 历史
print("📝 初始化 EXP 历史记录...\n")
for agent_name, exp_data in results.items():
scanner.initialize_exp_history(agent_name, exp_data)
print()
print("=" * 60)
print(" 初始化完成!")
print("=" * 60)
# 生成统计报告
print()
print("📊 统计报告:")
print()
print(f"{'Agent':<10} | {'EXP':<8} | {'等级':<8} | {'Sessions':<10} | {'Tool Calls':<12}")
print("-" * 60)
# 按 EXP 排序
sorted_agents = sorted(results.items(), key=lambda x: x[1]['exp'], reverse=True)
for agent_name, exp_data in sorted_agents:
print(f"{agent_name:<10} | {exp_data['exp']:<8} | Lv.{exp_data['level']:<6} | {exp_data['sessions']:<10} | {exp_data['tool_calls']:<12}")
print()
if __name__ == '__main__':
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |