Score how much the 9 models disagree on a prompt (0..1).
AI agents call disagreement_score to retrieve information from OpenClaw Consensus without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and analyzes consensus data from 9 LLM models to produce a scoring metric. It is a query/calculation operation with no side effects, data modifications, or irreversible actions. The most severe risk is a Read operation: an agent could misuse it to identify prompts where models disagree to exploit uncertainty, but this requires no elevated permissions and causes no direct harm.
From the tool's definition Tool description: 'Score how much the 9 models disagree on a prompt (0..1)'. The tool computes and returns a disagreement metric without modifying any data, executing commands, or triggering external operations.
Attacks that exploit this kind of access
Score how much the 9 models disagree on a prompt (0..1). It is categorised as a Read tool in the OpenClaw Consensus MCP Server, which means it retrieves data without modifying state.
Register the OpenClaw Consensus MCP server in PolicyLayer and add a rule for disagreement_score: 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 OpenClaw Consensus. Nothing to install.
disagreement_score 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 disagreement_score 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 disagreement_score. 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.
disagreement_score is provided by the OpenClaw Consensus MCP server (MICONNM/openclaw-consensus-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
disagreement_score is one line of OpenClaw Consensus's registry record.
The record carries the whole server: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.
Teams ship this data inside their own products. See what a licence covers →