Yaoyao Memory Homo

六层架构记忆系统 - 让 AI 跨会话保持上下文、沉淀知识、持续进化 【本地存储】SQLite 数据库 【全文搜索】FTS5 搜索引擎 【模块化】核心+可选模块，按需安装

What you should check before installing or enabling auto features: - This skill includes a large codebase (many scripts) — review auto_updater.py, safe_extension_loader.py, vector_extension_manager.py, auto_updater/self_improver and any code that performs network requests or dynamic code loading before enabling automatic updates or extension loading. - Cloud sync (IMA/Samba/SFTP) and LLM/Embedding integrations are optional. Only configure API keys/credentials if you trust the endpoints. Prefer storing credentials with strict file permissions (600) and verify where the skill will read them (~/.config/ima, ~/.openclaw/credentials/secrets.env). - If you plan to run the dashboard/api_server, inspect api_server.py to confirm it only exposes local interfaces or is properly authenticated before binding to network ports. - Consider disabling auto-update/self-improvement and native-extension loading until you audit the updater/source URLs and the code that verifies update integrity (signatures/hashes). - Run the skill in a restricted environment (sandbox/container or a non-sensitive account) first, and keep backups of your ~/.openclaw workspace and databases before enabling cloud sync or running migration scripts. If you want, I can list specific files to inspect first (auto_updater.py, safe_extension_loader.py, install_modules.py, sync_ima.py, api_server.py) and summarize suspicious patterns found inside them.

Type: OpenClaw Skill Name: yaoyao-memory-v2 Version: 4.0.1 The skill bundle is an extremely comprehensive memory system that includes high-risk capabilities such as a local HTTP API server (api_server.py) and a shell command execution utility (shell_embed.py). While these features are aligned with the stated goal of advanced memory management and include defensive measures like command whitelisting and a governance layer (governance.py), they significantly expand the agent's attack surface. Specifically, the API server allows for remote configuration and data access, and the shell utility—despite using shell=False and a whitelist—could potentially be abused to interact with the host system in unintended ways. The presence of a modular installer (install_modules.py) that performs its own security scans suggests a sophisticated but potentially fragile security model, making the bundle 'suspicious' due to the inherent risks of its powerful administrative tools.