AI agents call suggest_next to retrieve information from Substrate without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool queries contextual information to generate suggestions. This is fundamentally a read operation: it retrieves or generates guidance based on existing state without causing side effects, executing external commands, modifying data, or performing destructive actions. The low severity reflects that misuse would provide unhelpful suggestions rather than compromise systems or data.
From the tool's definition Tool description states 'Get smart suggestions for next tool based on current context' - this is a retrieval operation that provides recommendations without modifying, executing, or deleting any data.
Attacks that exploit this kind of access
Get smart suggestions for next tool based on current context. It is categorised as a Read tool in the Substrate MCP Server, which means it retrieves data without modifying state.
Register the Substrate MCP server in PolicyLayer and add a rule for suggest_next: 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 Substrate. Nothing to install.
suggest_next 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 suggest_next 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 suggest_next. 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.
suggest_next is provided by the Substrate MCP server (ivan-saorin/substrate). 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 →