Scan memory content before storage for poisoning attacks
AI agents call scan_memory to retrieve information from ZugaShield without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs security analysis and detection (scanning) of memory content to identify threats, but does not modify, delete, execute code, or move resources. It retrieves and inspects data for threat patterns—a quintessential Read operation. The security context (poison detection) does not elevate the category beyond Read, as the tool itself performs no side effects on the scanned content.
From the tool's definition Tool name 'scan_memory' and description 'Scan memory content before storage for poisoning attacks' indicate inspection/analysis of data without modification or execution.
Attacks that exploit this kind of access
Scan memory content before storage for poisoning attacks. It is categorised as a Read tool in the ZugaShield MCP Server, which means it retrieves data without modifying state.
Register the ZugaShield MCP server in PolicyLayer and add a rule for scan_memory: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches ZugaShield. Nothing to install.
scan_memory is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the scan_memory rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for scan_memory. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
scan_memory is provided by the ZugaShield MCP server (zuga-technologies/zugashield). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →