Describe one sObject (standard or custom): fields, labels, types, relationships.
AI agents call describe_sobject to retrieve information from Salesforce without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves metadata about Salesforce objects with no ability to modify, delete, or execute operations. It is a pure read operation that returns structural information about sObjects. The read-only nature of the server and the descriptive nature of the operation confirm this is a Read category risk.
From the tool's definition Tool name is 'describe_sobject' and description states it describes sObject metadata (fields, labels, types, relationships). Server description explicitly states 'read-only access' and the tool performs no mutations.
Attacks that exploit this kind of access
Describe one sObject (standard or custom): fields, labels, types, relationships. It is categorised as a Read tool in the Salesforce MCP Server, which means it retrieves data without modifying state.
Register the Salesforce MCP server in PolicyLayer and add a rule for describe_sobject: 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 Salesforce. Nothing to install.
describe_sobject 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 describe_sobject 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 describe_sobject. 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.
describe_sobject is provided by the Salesforce MCP server (pmankar-netchex/salesforce-mcp). 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 →