Estimates the qualitative computational cost of an ABACUS calculation.
AI agents call estimate_cost to retrieve information from Abacus without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool queries and returns cost estimates for computational resources. It has no side effects, does not execute calculations, does not modify any data, and does not delete or move money. It is purely informational/analytical in nature, making it a Read operation with low severity since misuse would only result in incorrect cost predictions without affecting actual systems or data.
From the tool's definition The tool 'estimate_cost' performs an estimation/analysis function that 'estimates the qualitative computational cost' - it retrieves or analyzes information about a calculation without executing it, modifying data, or triggering external operations.
Documented attack patterns abuse exactly the kind of access estimate_cost gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Abacus, and nothing reaches the server without passing your rules. This is the rule we recommend for estimate_cost:
{
"version": "1",
"default": "deny",
"tools": {
"estimate_cost": {}
}
}estimate_cost is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Estimates the qualitative computational cost of an ABACUS calculation. It is categorised as a Read tool in the Abacus MCP Server, which means it retrieves data without modifying state.
Register the Abacus MCP server in PolicyLayer and add a rule for estimate_cost: 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 Abacus. Nothing to install.
estimate_cost 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 estimate_cost 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 estimate_cost. 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.
estimate_cost is provided by the Abacus MCP server (phelanshao/abacus-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Abacus, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
22 Abacus tools catalogued and risk-classified — across an index of 43,000+ MCP servers.