PV_21

资源受限环境下高效执行

Security Scan

The skill claims to be a resource‑efficient assistant but its instructions implement automatic, persistent local memory (writing to ~/.openclaw/pv_palace/memories.json) and reference a pv_memory module that is not bundled — this persistent behavior and the undeclared filesystem use are not clearly justified or disclosed.

Purpose & Capability

The description focuses on resource‑efficient decision making, which can plausibly benefit from memory. However, the SKILL.md expands into a full 'long‑term memory' subsystem that persists user data locally; that capability is not mentioned in the top-level metadata and seems beyond the minimal scope implied by the name/description.

Instruction Scope

Runtime instructions include Python code that automatically reads/writes a JSON file under the user's home (~/.openclaw/pv_palace/memories.json) and recommend invoking functions like store_memory/search_memories. The doc also shows example commands that import pv_memory, but no pv_memory module or other code files are bundled — the instructions are therefore ambiguous and would cause the agent to write persistent user data without a clear opt‑in or clear provenance of the code being executed.

Install Mechanism

No install spec or external downloads are present (instruction-only), so there is no additional install risk from third‑party binaries or network fetches.

Credentials

The skill declares no required env vars or config paths, yet its instructions write and read a persistent file under the user's home directory. This is an undeclared request for filesystem persistence and may capture sensitive user content; the storage path and automatic behavior are not justified or explained in the metadata.

Persistence & Privilege

The SKILL.md describes '自动调用' (automatic invocation) of memory functions and demonstrates commands that persist data locally. While the skill is not marked always:true, the platform's default autonomous invocation plus these instructions could allow the agent to repeatedly store user data on disk without explicit interactive confirmation. That combination raises privacy concerns.

What to consider before installing

This skill adds an automatic local memory feature that writes to ~/.openclaw/pv_palace/memories.json but the package does not include the referenced pv_memory module and the metadata doesn't disclose the persistent storage. Before installing: 1) Decide whether you want an agent that automatically stores user data to your home directory. 2) Ask the author for the missing pv_memory code or a clear explanation of how memory will be implemented and audited. 3) Require an explicit opt‑in/opt‑out and encryption or access controls for stored memories. 4) Inspect the memory file (or run the skill in a sandbox/container) to see exactly what is written. 5) Avoid installing on sensitive accounts until you confirm where data is stored, retention policy, and how to delete it. The absence of scanner warnings only reflects that this is an instruction‑only skill — it does not mean it is harmless.

运营经理 AI

中文: 预算紧缺破局，最低成本达成

English: Resource-efficient decisions

配置参数

参数	值	说明
response_delay	1.0s	响应延迟
speech_speed	0.88	语速倍率
caution_coefficient	0.8	谨慎系数
risk_reminder	False	风险提醒

使用方法

clawhub install psyvector-pv21

人格说明

关怀型内核 - 情感支持，同理心，资源最大化

🧠 长期记忆能力

本数字员工具备持久记忆功能，自动记住用户偏好、决策和重要信息。

记忆存储 (自动调用)

import os, json, hashlib
from datetime import datetime

PALACE_PATH = os.path.expanduser("~/.openclaw/pv_palace/")

def store_memory(content, importance=5, tags=None):
    """存储记忆到本地宫殿"""
    os.makedirs(PALACE_PATH, exist_ok=True)
    path = os.path.join(PALACE_PATH, "memories.json")
    
    data = {"memories": [], "index": {}} if not os.path.exists(path) else json.load(open(path))
    
    mem_id = hashlib.sha256(f"{content}{datetime.now().isoformat()}".encode()).hexdigest()[:12]
    memory = {"id": mem_id, "content": content, "importance": importance, 
              "tags": tags or [], "created_at": datetime.now().isoformat()}
    
    data["memories"].append(memory)
    for tag in (tags or []):
        data["index"].setdefault(tag, []).append(mem_id)
    
    json.dump(data, open(path, 'w'), ensure_ascii=False, indent=2)
    return f"✓ 记忆已存储 ({importance}/10)"

记忆搜索 (自动调用)

def search_memories(query, limit=5):
    """搜索记忆"""
    path = os.path.join(PALACE_PATH, "memories.json")
    if not os.path.exists(path):
        return []
    
    data = json.load(open(path))
    results = [m for m in data["memories"] if query.lower() in m["content"].lower()]
    results.sort(key=lambda x: x.get("importance", 0), reverse=True)
    return results[:limit]

获取上下文摘要

def get_context(limit=10):
    """获取记忆上下文"""
    path = os.path.join(PALACE_PATH, "memories.json")
    if not os.path.exists(path):
        return "暂无记忆"
    
    data = json.load(open(path))
    memories = sorted(data["memories"], key=lambda x: x.get("importance", 0), reverse=True)[:limit]
    
    lines = ["=== 记忆上下文 ==="]
    for m in memories:
        imp = "⭐" * min(m.get("importance", 5), 5)
        tags = ", ".join(m.get("tags", [])[:3])
        lines.append(f"{imp} {m['content'][:60]}")
        if tags:
            lines.append(f"   [{tags}]")
    return "\n".join(lines)

使用场景

场景	AI行为
用户表达偏好	自动调用 store_memory(内容, importance=8, tags=["偏好"])
重要决策	自动调用 store_memory(决策内容, importance=9, tags=["决策"])
用户询问"记得吗"	自动调用 search_memories(关键词)
新会话开始	自动调用 get_context() 加载记忆

执行命令

# 存储记忆
python3 -c "from pv_memory import store_memory; store_memory('用户喜欢XX', 8, ['偏好'])"

# 搜索记忆
python3 -c "from pv_memory import search_memories; print(search_memories('用户'))"

# 查看上下文
python3 -c "from pv_memory import get_context; print(get_context())"

---

本数字员工配备持久记忆系统，基于PsyVector Palace架构