AI agents call get_xrefs_to to retrieve information from GhidraMCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This is a Read operation. Cross-reference retrieval is a fundamental code analysis operation that queries existing data without modifying, deleting, or executing anything. Even in a reverse-engineering context, gathering xref metadata poses minimal risk—it informs analysis but does not change the target binary or trigger external operations.
From the tool's definition Tool name 'get_xrefs_to' indicates retrieval of cross-references (xrefs) to a target; the sibling tools like 'get_xrefs_from', 'get_function_xrefs', 'list_classes', and 'list_data_items' are all informational queries with no side effects.
Attacks that exploit this kind of access
get_xrefs_to. It is categorised as a Read tool in the GhidraMCP MCP Server, which means it retrieves data without modifying state.
Register the Ghidra MCP server in PolicyLayer and add a rule for get_xrefs_to: 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 GhidraMCP. Nothing to install.
get_xrefs_to 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 get_xrefs_to 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 get_xrefs_to. 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.
get_xrefs_to is provided by the Ghidra MCP server (pr0cf5/ghidramcp). 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 →