AI agents call get_budget_balances to retrieve information from Actual without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool only retrieves and returns existing budget data (category balances). It performs no write, delete, or execution operations. The action is passive data retrieval with no side effects. Even in a financial context (Actual Budget), read-only access to balance information poses minimal risk—the severity is low because the tool cannot move money, modify accounts, or execute transactions.
From the tool's definition Tool name 'get_budget_balances' and description 'Returns budget category balances' indicate a retrieval operation with no modification capability. Sibling tools include 'get_transactions' and 'list_accounts', which are clearly read operations.
Attacks that exploit this kind of access
Returns budget category balances. It is categorised as a Read tool in the Actual MCP Server, which means it retrieves data without modifying state.
Register the Actual MCP server in PolicyLayer and add a rule for get_budget_balances: 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 Actual. Nothing to install.
get_budget_balances 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_budget_balances 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_budget_balances. 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_budget_balances is provided by the Actual MCP server (slushpupie/actual-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 →