Checkpoint research task
checkpoint_research_taskAfter each meaningful step, save the updated plan, cursor, new findings with exact evidence ids, and tool calls used; pass expectedVersion from the latest read.
When to use it
After each meaningful step, save the updated plan, cursor, new findings with exact evidence ids, and tool calls used; pass expectedVersion from the latest read.
List existing Threads first, inspect the case file, then add exact evidence, propose a revision, or start one bounded run and poll it to a terminal state.
The updated task and new version. Budget limits are enforced here and return a visible error when exceeded.
Avoid using it when
Do not checkpoint hidden reasoning or unverified claims; store findings as evidence-linked observations.
Advanced input contract
Agent hosts use this JSON Schema to validate the exact input shape.
View full registry notes
Save progress on a research task: replace the plan or cursor, append evidence-linked findings, and report tool calls used. Pass expectedVersion from your latest read; budgets are enforced and exceeded limits return a visible error. Use when: After each meaningful step, save the updated plan, cursor, new findings with exact evidence ids, and tool calls used; pass expectedVersion from the latest read. Avoid when: Do not checkpoint hidden reasoning or unverified claims; store findings as evidence-linked observations.
View JSON Schema
{
"type": "object",
"properties": {
"taskId": {
"type": "string",
"format": "uuid"
},
"expectedVersion": {
"type": "integer",
"minimum": 1
},
"plan": {
"type": "array",
"maxItems": 30,
"items": {
"type": "object",
"properties": {
"id": {
"type": "string",
"minLength": 1,
"maxLength": 80
},
"label": {
"type": "string",
"minLength": 1,
"maxLength": 300
},
"status": {
"type": "string",
"enum": [
"pending",
"running",
"complete",
"failed",
"skipped"
]
}
},
"required": [
"id",
"label"
],
"additionalProperties": false
}
},
"cursor": {
"type": "object",
"additionalProperties": true
},
"findings": {
"type": "array",
"maxItems": 50,
"items": {
"type": "object",
"properties": {
"summary": {
"type": "string",
"minLength": 1,
"maxLength": 4000
},
"stance": {
"type": "string",
"enum": [
"supports",
"challenges",
"context"
],
"default": "context"
},
"evidence": {
"type": "array",
"maxItems": 20,
"items": {
"type": "object",
"properties": {
"sourceType": {
"type": "string",
"enum": [
"file",
"document",
"artifact",
"claim",
"thread",
"observation",
"url"
]
},
"sourceId": {
"type": "string",
"minLength": 1,
"maxLength": 200
},
"url": {
"type": "string",
"format": "uri",
"maxLength": 2000
},
"excerpt": {
"type": "string",
"maxLength": 2000
}
},
"required": [
"sourceType"
],
"additionalProperties": false
}
}
},
"required": [
"summary"
],
"additionalProperties": false
}
},
"toolCallsUsed": {
"type": "integer",
"minimum": 0,
"maximum": 5000
},
"note": {
"type": "string",
"maxLength": 2000
}
},
"required": [
"taskId"
],
"additionalProperties": false
}