gradeNot graded
get_companion_vocabularyAnswer
Returns reusable, versioned vocabulary derived deterministically from stored factual relations for one subject. Inspect labels and choose relevantVocabularyIds yourself for query_companion_memory. Companion does not interpret your question, plan retrieval, evaluate sufficiency, or automatically refresh vocabulary. Source attribution is not objective truth.
View tool JSON
{
"name": "get_companion_vocabulary",
"kind": "answer",
"impl": "imperative",
"description": "Returns reusable, versioned vocabulary derived deterministically from stored factual relations for one subject. Inspect labels and choose relevantVocabularyIds yourself for query_companion_memory. Companion does not interpret your question, plan retrieval, evaluate sufficiency, or automatically refresh vocabulary. Source attribution is not objective truth.",
"inputSchema": {
"type": "object",
"properties": {
"subjectId": {
"type": "string",
"description": "Exact subject identifier; defaults to the current active subject when omitted."
}
},
"additionalProperties": false
},
"page": "/app/index.html"
}query_companion_memoryAnswer
The external agent must choose relevantVocabularyIds from reusable vocabulary. Returns only those stored factual relations and their endpoint nodes/source citations by exact ID; no model call, question interpretation, ranking, automatic expansion, vocabulary refresh or sufficiency decision. Unknown IDs are listed explicitly. Empty IDs return no evidence. SOURCE_EXPLICIT and SOURCE_STRONGLY_IMPLIED preserve attribution, not objective truth; AGENT_INFERRED is excluded. The agent owns subsequent reasoning.
View tool JSON
{
"name": "query_companion_memory",
"kind": "answer",
"impl": "imperative",
"description": "The external agent must choose relevantVocabularyIds from reusable vocabulary. Returns only those stored factual relations and their endpoint nodes/source citations by exact ID; no model call, question interpretation, ranking, automatic expansion, vocabulary refresh or sufficiency decision. Unknown IDs are listed explicitly. Empty IDs return no evidence. SOURCE_EXPLICIT and SOURCE_STRONGLY_IMPLIED preserve attribution, not objective truth; AGENT_INFERRED is excluded. The agent owns subsequent reasoning.",
"inputSchema": {
"type": "object",
"properties": {
"subjectId": {
"type": "string",
"minLength": 1,
"description": "Exact subject identifier from the vocabulary."
},
"relevantVocabularyIds": {
"type": "array",
"items": {
"type": "string",
"minLength": 1
},
"description": "Opaque vocabulary IDs selected by the external agent. Duplicates are collapsed; an empty array requests no evidence."
}
},
"required": [
"subjectId",
"relevantVocabularyIds"
],
"additionalProperties": false
},
"page": "/app/index.html"
}