multivon-mcp

Official

Overview Schema Source Related Servers Discussions Admin

eval_tool_call_accuracy

Check whether an agent called the expected tool or sequence by comparing expected versus actual calls, with optional argument, order, and unexpected-call penalization.

Instructions

Evaluate whether an agent called the expected tool or tool sequence.

Pure deterministic — no LLM judge needed. Two compatible modes:

Single-call mode compares expected_tool / actual_tool and optional argument dictionaries exactly.
Trace mode consumes expected_tool_calls plus the canonical agent_trace returned by eval_ingest_trace. It can require order and optionally penalize unexpected calls.

Args: expected_tool: Single tool name the agent should have called. actual_tool: Single tool name the agent actually called. expected_arguments: Expected arguments for single-call mode. actual_arguments: Actual arguments for single-call mode. expected_tool_calls: Expected names for trace mode. An empty list explicitly asserts that the agent should call no tools. agent_trace: Canonical step dictionaries returned by eval_ingest_trace. require_order: In trace mode, require expected names in order. penalize_unexpected: In trace mode, lower the score for calls not present in expected_tool_calls.

Returns: {"score": float, "passed": bool, "reason": str, "evaluator": "tool_call_accuracy"}, or an error dict when the arguments do not form either mode.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`actual_tool`	No
`agent_trace`	No
`expected_tool`	No
`require_order`	No
`actual_arguments`	No
`expected_arguments`	No
`expected_tool_calls`	No
`penalize_unexpected`	No

Output Schema

TableJSON Schema

Name	Required	Description	Default
No arguments

Schema Changelog

Changes observed during successful MCP inspections.

Changed11 schema fields changedv0.4.0Sep 19, 2026

addedInput schema / properties / actual_tool / anyOf

Added value: +[
+  {
+    "type": "string"
+  },
+  {
+    "type": "null"
+  }
+]

addedInput schema / properties / actual_tool / default
```
Added value: +null
```
removedInput schema / properties / actual_tool / type
```
Removed value: -"string"
```

addedInput schema / properties / agent_trace

Added value: +{
+  "anyOf": [
+    {
+      "items": {
+        "additionalProperties": true,
+        "type": "object"
+      },
+      "type": "array"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "title": "Agent Trace"
+}

addedInput schema / properties / expected_tool / anyOf

Added value: +[
+  {
+    "type": "string"
+  },
+  {
+    "type": "null"
+  }
+]

addedInput schema / properties / expected_tool / default
```
Added value: +null
```
removedInput schema / properties / expected_tool / type
```
Removed value: -"string"
```

addedInput schema / properties / expected_tool_calls

Added value: +{
+  "anyOf": [
+    {
+      "items": {
+        "type": "string"
+      },
+      "type": "array"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "title": "Expected Tool Calls"
+}

addedInput schema / properties / penalize_unexpected

Added value: +{
+  "default": false,
+  "title": "Penalize Unexpected",
+  "type": "boolean"
+}

addedInput schema / properties / require_order

Added value: +{
+  "default": false,
+  "title": "Require Order",
+  "type": "boolean"
+}

removedInput schema / required

Removed value: -[
-  "expected_tool",
-  "actual_tool"
-]

First observedv0.1.0May 17, 2026

TDQS

A4.8/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries full behavioral disclosure responsibility, and it does very well. It explicitly states 'Pure deterministic — no LLM judge needed', explains exact argument comparison, describes order-sensitive behavior, optional unexpected-call penalization, the empty-list assertion, and the return/error contract.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is front-loaded with the core purpose and then organized into mode bullets, Args, and Returns. Despite covering eight parameters and two modes, it contains no filler and each sentence adds necessary information.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a complex evaluation tool with zero annotations and zero schema descriptions, the description is remarkably complete. It covers both modes, the relationship to eval_ingest_trace, special cases like empty expected_tool_calls, default-order behavior, penalization semantics, return schema, and the error condition for invalid argument combinations.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0%, so the description must compensate, and it does: every parameter is named and given operational meaning. Expected/actual arguments are tied to single-call mode, expected_tool_calls and agent_trace to trace mode, and require_order/penalize_unexpected to boolean behavioral switches.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with a specific verb and resource: 'Evaluate whether an agent called the expected tool or tool sequence.' It clearly distinguishes this as an evaluation tool for tool-call accuracy, not a generic evaluator or a tool being evaluated. The mention of two modes further pins down what the tool does.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description gives clear context for when each mode applies: single-call mode for comparing one expected/actual pair, trace mode for consuming the canonical agent_trace from eval_ingest_trace. It does not explicitly name alternatives or state 'use this instead of X', but the mode breakdown and deterministic no-LLM-judge note give sufficient decision guidance.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.