AI agents call kg_explain to retrieve information from Predicate without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
kg_explain is fundamentally a read operation that retrieves and returns inference traces from the knowledge graph. It answers questions about existing claims and their provenance without creating, modifying, or deleting any data. The tool's purpose is to make the reasoning process auditable and transparent, which aligns with read-only data retrieval.
From the tool's definition Tool description states 'Return one valid inference trace for a claim, with cited provenance' — this is a retrieval and explanation operation that queries the knowledge graph to provide reasoning traces and citations.
Risk signalsBulk/mass operation — affects multiple targets
Attacks that exploit this kind of access
Return one valid inference trace for a claim, with cited provenance for every asserted premise. It is categorised as a Read tool in the Predicate MCP Server, which means it retrieves data without modifying state.
Register the Predicate MCP server in PolicyLayer and add a rule for kg_explain: 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 Predicate. Nothing to install.
kg_explain 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 kg_explain 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 kg_explain. 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.
kg_explain is provided by the Predicate MCP server (nordicagents/predicate). 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 →