Agent Reliability

Answer a question about Agent Reliability

answer

Read-onlyIdempotent

Answer a question from the corpus, or refuse. Returns only the claims that bear on the question, each with the sources it cites and its editorial confidence. When the corpus cannot answer, answered is false and abstention_reason plus missing_topics say what was not covered — a refusal is a real result here, not an error. Use this when the user asked a question in words; use search when you want to see the candidates yourself.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "question": "<question>"
    },
    "name": "answer"
  }
}

Input Schema

Name	Required	Description	Default
`question`	Yes	The question to answer, in the words the user asked it.

Output Schema

Name	Required	Description
`claims`	No
`entity`	No
`margin`	No	Relative gap between the best candidate and the runner-up. A small margin means the corpus cannot tell them apart.
`answered`	Yes
`coverage`	No	Share of the question's information the winning object accounts for, weighted by term rarity.
`candidates`	Yes	Near misses, so a refusal is still a lead. Present whether or not the question was answered.
`match_score`	No
`matched_fields`	No
`missing_topics`	Yes	Meaningful words in the question that appear nowhere in the winning object — the gap, named.
`match_threshold`	Yes	The coverage this corpus requires before it will answer at all.
`match_confidence`	No	How well the answer fits the QUESTION — retrieval confidence, not editorial. Never confuse it with a claim's own confidence, which is how much the corpus stands behind the statement.
`matched_question`	No	The stored question this was matched to, when the answer came from one.
`abstention_reason`	No	Why the corpus declined: empty_query, no_candidates, below_coverage, ambiguous_candidates or no_intent_match.

Schema Changelog

Changes observed during successful MCP inspections.

Changed14 schema fields changedSep 6, 2026

addedOutput schema / properties / abstention_reason

Added value: +{
+  "description": "Why the corpus declined: empty_query, no_candidates, below_coverage, ambiguous_candidates or no_intent_match.",
+  "type": "string"
+}

addedOutput schema / properties / candidates

Added value: +{
+  "description": "Near misses, so a refusal is still a lead. Present whether or not the question was answered.",
+  "items": {
+    "additionalProperties": false,
+    "properties": {
+      "coverage": {
+        "type": "number"
+      },
+      "id": {
+        "type": "string"
+      },
+      "score": {
+        "type": "number"
+      }
+    },
+    "required": [
+      "id",
+      "score",
+      "coverage"
+    ],
+    "type": "object"
+  },
+  "type": "array"
+}

addedOutput schema / properties / claims / items / properties / relevance
```
Added value: +{
+  "type": "number"
+}
```

changedOutput schema / properties / claims / items / required

Previous value: -[
-  "text",
-  "confidence",
-  "sources"
-]New value: +[
+  "text",
+  "confidence",
+  "sources",
+  "relevance"
+]

addedOutput schema / properties / coverage

Added value: +{
+  "description": "Share of the question's information the winning object accounts for, weighted by term rarity.",
+  "type": "number"
+}

addedOutput schema / properties / entity / properties / x

Added value: +{
+  "additionalProperties": {
+    "additionalProperties": {},
+    "propertyNames": {
+      "type": "string"
+    },
+    "type": "object"
+  },
+  "description": "Indexed instance-specific attributes, grouped by namespace — the same fields api/index.json publishes. Absent when the instance declares none.",
+  "propertyNames": {
+    "type": "string"
+  },
+  "type": "object"
+}

addedOutput schema / properties / margin

Added value: +{
+  "description": "Relative gap between the best candidate and the runner-up. A small margin means the corpus cannot tell them apart.",
+  "type": "number"
+}

addedOutput schema / properties / match_confidence

Added value: +{
+  "description": "How well the answer fits the QUESTION — retrieval confidence, not editorial. Never confuse it with a claim's own confidence, which is how much the corpus stands behind the statement.",
+  "type": "number"
+}

addedOutput schema / properties / match_score
```
Added value: +{
+  "type": "number"
+}
```

addedOutput schema / properties / match_threshold

Added value: +{
+  "description": "The coverage this corpus requires before it will answer at all.",
+  "type": "number"
+}

addedOutput schema / properties / matched_fields

Added value: +{
+  "items": {
+    "type": "string"
+  },
+  "type": "array"
+}

addedOutput schema / properties / matched_question / description

Added value: +"The stored question this was matched to, when the answer came from one."

addedOutput schema / properties / missing_topics

Added value: +{
+  "description": "Meaningful words in the question that appear nowhere in the winning object — the gap, named.",
+  "items": {
+    "type": "string"
+  },
+  "type": "array"
+}

changedOutput schema / required

Previous value: -[
-  "answered"
-]New value: +[
+  "answered",
+  "match_threshold",
+  "missing_topics",
+  "candidates"
+]

Changed1 schema field changedSep 1, 2026
- addedInput schema / properties / question / description
```
Added value: +"The question to answer, in the words the user asked it."
```
First observedAug 23, 2026

TDQS

A4.7/5.0

Behavior5/5

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

The description goes beyond the readOnly/idempotent annotations by explaining refusal semantics in detail: 'answered is false and abstention_reason plus missing_topics say what was not covered — a refusal is a real result here, not an error.' It also discloses that the tool returns only claims bearing on the question with sources and editorial confidence, which is useful behavioral context.

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?

Three sentences, all high-value: the core action, the return format, and routing to search. There is no redundant filler, and the most important scoping information is front-loaded.

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 single-parameter tool with an output schema, the description is complete. It covers the key edge case (corpus cannot answer), names the output fields for that case, and provides sibling differentiation. Nothing essential is missing for correct invocation.

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 coverage is 100% and the parameter description already states 'The question to answer, in the words the user asked it.' The tool description reinforces the verbatim natural-language requirement ('when the user asked a question in words') but does not add material new semantics beyond the schema, so baseline 3 applies.

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: 'Answer a question from the corpus, or refuse.' It clearly distinguishes itself from the sibling search tool by stating 'Use this when the user asked a question in words; use search when you want to see the candidates yourself.' This differentiates it without ambiguity.

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

Usage Guidelines5/5

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

The description explicitly states when to use the tool and names the alternative: 'Use this when the user asked a question in words; use search when you want to see the candidates yourself.' It also clarifies that a refusal is a legitimate outcome, preventing the agent from misinterpreting abstention as an error.

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

All Agent Reliability 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.