get_prediction
Get details about a specific prediction
This record as markdown: /tools/deepfates-mcp-replicate/get-prediction.md
What get_prediction does on Replicate MCP Server
AI agents call get_prediction to retrieve information from Replicate MCP Server without modifying anything. It is typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Why get_prediction is rated Low
This tool performs a read-only operation that retrieves prediction metadata or status from the Replicate API. No data is created, modified, deleted, or executed. The blast radius of misuse is minimal—an attacker could only view prediction details they query, which poses a low information disclosure risk. This clearly falls into the Read category.
From the tool's definition Tool name 'get_prediction' and description 'Get details about a specific prediction' indicate a query/retrieval operation with no side effects. It retrieves information about an existing prediction without modifying, deleting, or executing any operations.
Attacks that exploit this kind of access
The rule that runs get_prediction safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Replicate MCP Server, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For get_prediction, this is the rule to start with:
get_prediction is read-only, so it stays allowed. Everything else on the server is denied unless you say otherwise.
The button opens the PolicyLayer dashboard: create your workspace, connect Replicate MCP Server, apply this rule, and every get_prediction call is checked against it from then on.
Questions about get_prediction
Get details about a specific prediction. It is categorised as a Read tool in the Replicate MCP Server MCP Server, which means it retrieves data without modifying state.
Register the Replicate MCP Server MCP server in PolicyLayer and add a rule for get_prediction: 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 Replicate MCP Server. Nothing to install.
get_prediction 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_prediction 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_prediction. 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_prediction is provided by the Replicate MCP Server MCP server (deepfates/mcp-replicate). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Replicate MCP Server, and thousands of servers like it.
This server
Across the catalogue