AI agents call zig_references to retrieve information from Zig without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool searches for and retrieves references to a symbol in Zig source code. It does not modify, execute, or delete anything—it only queries and returns information. This is a standard IDE feature that performs static analysis to locate where a symbol is used, making it a low-severity Read operation with high confidence.
From the tool's definition Tool name 'zig_references' and description 'Find all references to a symbol' indicates a query/search operation with no side effects. The verb 'find' and the readonly nature of symbol reference lookup confirm this is a Read operation.
Attacks that exploit this kind of access
Find all references to a symbol at a specific position in a Zig file via ZLS. It is categorised as a Read tool in the Zig MCP Server, which means it retrieves data without modifying state.
Register the Zig MCP server in PolicyLayer and add a rule for zig_references: 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 Zig. Nothing to install.
zig_references 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 zig_references 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 zig_references. 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.
zig_references is provided by the Zig MCP server (sadopc/mcp-server-zig). 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 →