Return total cost (USD), token counts broken down by input / cache read / cache create / output, and turn count for a given session.
AI agents call get_session_cost to retrieve information from AgentWatch without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool reads pre-computed cost and token metrics from a session record. It has no side effects, does not execute code, does not modify data, and does not involve financial transactions (it merely reports costs, not moving money). The blast radius of misuse is minimal—an attacker could query session costs to gather intelligence about usage patterns, but cannot alter records or trigger operations.
From the tool's definition Tool description states 'Return total cost (USD), token counts... for a given session' — it retrieves and queries session accounting data without modifying or deleting it.
Documented attack patterns abuse exactly the kind of access get_session_cost gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and AgentWatch, and nothing reaches the server without passing your rules. This is the rule we recommend for get_session_cost:
{
"version": "1",
"default": "deny",
"tools": {
"get_session_cost": {}
}
}get_session_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.
Return total cost (USD), token counts broken down by input / cache read / cache create / output, and turn count for a given session. It is categorised as a Read tool in the AgentWatch MCP Server, which means it retrieves data without modifying state.
Register the AgentWatch MCP server in PolicyLayer and add a rule for get_session_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 AgentWatch. Nothing to install.
get_session_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 get_session_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 get_session_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.
get_session_cost is provided by the AgentWatch MCP server (mishanefedov/agentwatch). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from AgentWatch, 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.
5 AgentWatch tools catalogued and risk-classified — across an index of 43,000+ MCP servers.