Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only memory guide that tells the agent how to use existing MemOS memory tools, with privacy-relevant behavior that is expected for its stated purpose.
Install this only if you are comfortable with the agent using stored conversation history to answer requests. For sensitive topics, ask the agent to confirm before searching memory, and require explicit approval before it installs any learned skill for future sessions.
The skill directs the agent to use long-term memory whenever the user refers to past chats, preferences, or history, which is a broad trigger and can cause memory retrieval in many ordinary conversations. That increases unnecessary access to stored personal context and raises privacy risk, especially when the user did not explicitly ask for memory lookup in the current turn.
The skill instructs the agent to search and use prior conversation memory but provides no requirement to disclose that behavior or warn the user when personal history is being consulted. This can lead to non-transparent processing of sensitive historical data and surprise the user with information sourced from prior chats.
Skill grants unrestricted tool access without appropriate constraints. An agent with unfettered tool access can perform arbitrary actions including file modification, network requests, and code execution.
## How memory is provided each turn
- **Automatic recall (hook):** At the start of each turn, the system runs a memory search using the user's current message and injects relevant past memories into your context. You do not need to call any tool for that.
- **When that is not enough:** If the user's message is very long, vague, or the automatic search returns **no memories**, you should **generate your own short, focused query** and call `memory_search` yourself. For example:
- User sent a long paragraph → extract 1–2 key topics or a short question and search with that.
- Auto-recall said "no memories" or you see no memory block → call `memory_search` with a query you derive (e.g. the user's name, a topic they often mention, or a rephrased question).
No suspicious patterns detected.