Campaign Briefs / MCP
Revise icp rubric
Revise a brief ICP rubric from QA corrections on a canary sample. Pass the sample returned by get-icp-tuning-sample, operator corrections, and the current rubric. Returns { revisedRubric, rationale } without saving.
tools/callBackend AI, consent requiredParameters
workspaceIdOptional · string- Optional workspace ID or slug. Required when this credential can access multiple workspaces and no workspace was configured on the client.
briefIdRequired · string- Campaign brief id.
sampleRequired · array- Sample rows from get-icp-tuning-sample.
correctionsRequired · array- Operator judgments: rowId plus verdict fit, unfit, or skip.
currentRubricRequired · object- The rubric to revise. Usually the rubric returned by get-icp-tuning-sample.
allowHostedAiOptional · boolean- Must be true when this operation uses Pipeline hosted inference.
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "revise-icp-rubric",
"arguments": {
"briefId": "00000000-0000-4000-8000-000000000000",
"sample": [
{
"rowId": "row_01"
}
],
"corrections": [
{
"rowId": "row_01",
"verdict": "fit"
}
],
"currentRubric": {
"naturalLanguageCriteria": "Founder-led B2B software companies"
},
"allowHostedAi": true
}
}
}Full input schema
{
"type": "object",
"properties": {
"workspaceId": {
"type": "string",
"minLength": 1,
"description": "Optional workspace ID or slug. Required when this credential can access multiple workspaces and no workspace was configured on the client."
},
"briefId": {
"type": "string",
"description": "Campaign brief id."
},
"sample": {
"type": "array",
"description": "Sample rows from get-icp-tuning-sample."
},
"corrections": {
"type": "array",
"description": "Operator judgments: rowId plus verdict fit, unfit, or skip."
},
"currentRubric": {
"type": "object",
"description": "The rubric to revise. Usually the rubric returned by get-icp-tuning-sample."
},
"allowHostedAi": {
"type": "boolean",
"description": "Must be true when this operation uses Pipeline hosted inference."
}
},
"required": [
"briefId",
"sample",
"corrections",
"currentRubric"
]
}