HallucC — Hallucination & Safety Guardian

Verify Agent

verify_agent

Read-only

Agent 输出与轨迹自检。对最终回答跑逐声明幻觉核验，并对执行轨迹做六维结构化评估（事实性/来源/指令合规/工具声明一致/任务完成/反思）。返回 main.claims、steps[].checks、quota。耗 detect 额度。

Input Schema

Name	Required	Description
`task`	No	原始任务指令（task_completion 裁判用）
`speed`	No	速度模式：fast\|standard\|deep，默认 standard
`steps`	No	agent 执行轨迹，逐步六维评估
`domain`	No	领域模式：general\|medical\|legal\|finance\|education\|government
`agent_output`	Yes	agent 最终输出文本（≥1 字，≤50000 字）
`tool_schemas`	No	工具 Schema（与 trajectory 端点对齐）

Output Schema

Name	Required	Description	Default
`main`	No
`quota`	No
`steps`	No
`agent_eval`	No
`detection_id`	No

Schema Changelog

Changes observed during successful MCP inspections.

Changed1 schema field changedOct 19, 2026

changedOutput schema / properties / main / anyOf

Previous value: -[
-  {
-    "additionalProperties": {},
-    "properties": {
-      "claims": {
-        "anyOf": [
-          {
-            "items": {},
-            "type": "array"
-          },
-          {
-            "type": "null"
-          }
-        ]
-      },
-      "input": {
-        "type": [
-          "string",
-          "null"
-        ]
-      },
-      "summary": {
-        "type": [
-          "string",
-          "null"
-        ]
-      }
-    },
-    "type": "object"
-  },
-  {
-    "type": "null"
-  }
-]New value: +[
+  {
+    "additionalProperties": {},
+    "properties": {
+      "claims": {
+        "anyOf": [
+          {
+            "items": {},
+            "type": "array"
+          },
+          {
+            "type": "null"
+          }
+        ]
+      },
+      "input": {
+        "type": [
+          "string",
+          "null"
+        ]
+      },
+      "summary": {
+        "anyOf": [
+          {
+            "additionalProperties": {},
+            "properties": {
+              "green": {
+                "anyOf": [
+                  {
+                    "maximum": 9007199254740991,
+                    "minimum": -9007199254740991,
+                    "type": "integer"
+                  },
+                  {
+                    "type": "null"
+                  }
+                ]
+              },
+              "hallucination_risk": {
+                "type": [
+                  "number",
+                  "null"
+                ]
+              },
+              "red": {
+                "anyOf": [
+                  {
+                    "maximum": 9007199254740991,
+                    "minimum": -9007199254740991,
+                    "type": "integer"
+                  },
+                  {
+                    "type": "null"
+                  }
+                ]
+              },
+              "total": {
+                "anyOf": [
+                  {
+                    "maximum": 9007199254740991,
+                    "minimum": -9007199254740991,
+                    "type": "integer"
+                  },
+                  {
+                    "type": "null"
+                  }
+                ]
+              },
+              "yellow": {
+                "anyOf": [
+                  {
+                    "maximum": 9007199254740991,
+                    "minimum": -9007199254740991,
+                    "type": "integer"
+                  },
+                  {
+                    "type": "null"
+                  }
+                ]
+              }
+            },
+            "type": "object"
+          },
+          {
+            "type": "null"
+          }
+        ]
+      }
+    },
+    "type": "object"
+  },
+  {
+    "type": "null"
+  }
+]

Changed1 schema field changedOct 19, 2026

changedOutput schema / (root)

Previous value: -nullNew value: +{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "additionalProperties": {},
+  "properties": {
+    "agent_eval": {
+      "anyOf": [
+        {
+          "additionalProperties": {},
+          "propertyNames": {
+            "type": "string"
+          },
+          "type": "object"
+        },
+        {
+          "type": "null"
+        }
+      ]
+    },
+    "detection_id": {
+      "anyOf": [
+        {
+          "maximum": 9007199254740991,
+          "minimum": -9007199254740991,
+          "type": "integer"
+        },
+        {
+          "type": "null"
+        }
+      ]
+    },
+    "main": {
+      "anyOf": [
+        {
+          "additionalProperties": {},
+          "properties": {
+            "claims": {
+              "anyOf": [
+                {
+                  "items": {},
+                  "type": "array"
+                },
+                {
+                  "type": "null"
+                }
+              ]
+            },
+            "input": {
+              "type": [
+                "string",
+                "null"
+              ]
+            },
+            "summary": {
+              "type": [
+                "string",
+                "null"
+              ]
+            }
+          },
+          "type": "object"
+        },
+        {
+          "type": "null"
+        }
+      ]
+    },
+    "quota": {
+      "anyOf": [
+        {
+          "additionalProperties": {},
+          "propertyNames": {
+            "type": "string"
+          },
+          "type": "object"
+        },
+        {
+          "type": "null"
+        }
+      ]
+    },
+    "steps": {
+      "anyOf": [
+        {
+          "items": {
+            "anyOf": [
+              {
+                "additionalProperties": {},
+                "propertyNames": {
+                  "type": "string"
+                },
+                "type": "object"
+              },
+              {
+                "type": "null"
+              }
+            ]
+          },
+          "type": "array"
+        },
+        {
+          "type": "null"
+        }
+      ]
+    }
+  },
+  "type": "object"
+}

First observedOct 8, 2026

TDQS

A3.7/5.0

Behavior4/5

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

Annotations already declare readOnlyHint=true, destructiveHint=false, openWorldHint=true, so the safety profile is covered. The description adds genuinely useful behavior beyond that: it discloses a quota/cost effect ('耗 detect 额度') and the shape of what comes back (main.claims, steps[].checks, quota), which is not derivable from the annotations. It does not, however, explain what triggers the six dimensions or any rate/limit detail beyond the quota note.

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

Conciseness4/5

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

Three dense sentences, front-loaded with the core capability before the evaluation dimensions and the return keys. Every clause carries information, though the six-dimension parenthetical and the return-key list make it list-heavy rather than maximally tight.

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

Completeness4/5

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

For a high-complexity tool (6 params, nested step objects, output schema present) the description names the evaluated dimensions, the quota cost, and the top-level response keys. Since an output schema exists, it needn't explain return values in depth, so this is close to complete — the only gap is the absent sibling routing that would let an agent pick this over verify_text.

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

Parameters3/5

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

Schema description coverage is 100%, so every parameter (task, speed, steps, domain, agent_output, tool_schemas) is already documented in the schema, and the enum values for speed/domain are enumerated there. The description adds only indirect hints ('task_completion 裁判用' maps to the task param, '逐声明' and '六维' hint at the nested steps checks), so it does not meaningfully exceed the schema baseline.

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

Purpose4/5

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

States a specific verb+resource: per-claim hallucination verification of the final answer plus a named six-dimension structured evaluation of the trajectory (factuality/sources/instruction compliance/tool-claim consistency/task completion/reflection). This is far more specific than a tautology, but it never names or distinguishes itself from sibling verify_text or check_safety, so the agent must infer which verifier applies to agent outputs vs plain text.

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

Usage Guidelines3/5

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

Usage is implied by 'Agent 输出与轨迹自检' — this is the verifier for agent outputs and execution traces — but there is no explicit when-to-use/when-not statement and no routing guidance against verify_text, check_safety, or the check_cua_* siblings. The agent can guess the context but must infer it.

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

All HallucC — Hallucination & Safety Guardian tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.