Tool tree
gradeNot graded
recall_memoriesAction
Search the user's memories on this page by meaning. Read-only. Returns ranked matches; pending and superseded memories are excluded. If nothing matches, ask the user and then use remember.
View tool JSON
{
"name": "recall_memories",
"kind": "act",
"impl": "imperative",
"description": "Search the user's memories on this page by meaning. Read-only. Returns ranked matches; pending and superseded memories are excluded. If nothing matches, ask the user and then use remember.",
"inputSchema": {
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "What to look for, in natural language."
},
"limit": {
"type": "number",
"description": "Max results, default 8, max 25."
}
},
"required": [
"query"
]
}
}rememberAction
Propose a new memory. The card appears on the canvas and WAITS for the human to approve, edit, or reject it; this call resolves with their decision, or with status pending if the human has not decided within about a minute (the card stays on the canvas). Only store durable facts, preferences, decisions, or notes the user would want kept.
View tool JSON
{
"name": "remember",
"kind": "act",
"impl": "imperative",
"description": "Propose a new memory. The card appears on the canvas and WAITS for the human to approve, edit, or reject it; this call resolves with their decision, or with status pending if the human has not decided within about a minute (the card stays on the canvas). Only store durable facts, preferences, decisions, or notes the user would want kept.",
"inputSchema": {
"type": "object",
"properties": {
"content": {
"type": "string",
"description": "The memory text, 1 to 2000 characters."
},
"kind": {
"type": "string",
"enum": [
"fact",
"preference",
"decision",
"note"
]
},
"tags": {
"type": "array",
"items": {
"type": "string"
}
}
},
"required": [
"content"
]
}
}correct_memoryAction
Propose a replacement for a stale or wrong memory by id. The human approves or rejects; on approval the old memory is superseded and stays visible in the card's history. Resolves with the decision.
View tool JSON
{
"name": "correct_memory",
"kind": "act",
"impl": "imperative",
"description": "Propose a replacement for a stale or wrong memory by id. The human approves or rejects; on approval the old memory is superseded and stays visible in the card's history. Resolves with the decision.",
"inputSchema": {
"type": "object",
"properties": {
"id": {
"type": "string",
"description": "Id of the memory to correct."
},
"replacement": {
"type": "string",
"description": "Corrected text, 1 to 2000 characters."
}
},
"required": [
"id",
"replacement"
]
}
}forget_memoryAction
Ask to remove a memory by id. The human must confirm; the memory is tombstoned (its text is cleared, the id is kept). Resolves with forgotten or declined, or pending when the human has not decided yet.
View tool JSON
{
"name": "forget_memory",
"kind": "act",
"impl": "imperative",
"description": "Ask to remove a memory by id. The human must confirm; the memory is tombstoned (its text is cleared, the id is kept). Resolves with forgotten or declined, or pending when the human has not decided yet.",
"inputSchema": {
"type": "object",
"properties": {
"id": {
"type": "string"
},
"reason": {
"type": "string"
}
},
"required": [
"id"
]
}
}load_contextAnswer
Pull a compact markdown digest of the user's active memories into your context, optionally focused on a topic. Read-only. Use this at the start of a task to know what is already established.
View tool JSON
{
"name": "load_context",
"kind": "answer",
"impl": "imperative",
"description": "Pull a compact markdown digest of the user's active memories into your context, optionally focused on a topic. Read-only. Use this at the start of a task to know what is already established.",
"inputSchema": {
"type": "object",
"properties": {
"topic": {
"type": "string",
"description": "Optional topic to focus the digest."
}
}
}
}