The interview of a model: who answers which input and how far
interviewTracks ownership, progress, and corrections for model elicitation, letting agents claim inputs, resume stopped runs, audit answers, and accept or reopen steps.
Instructions
Track the interview of a model: ownership, progress, review and corrections.
progress names every input key with its state, owner, answered parts and
publication; call it before handing a shock to an agent. claim gives one key to
one agent; a live claim of another agent refuses a second writer. release drops
the caller's claim. restart resumes an agent whose run stopped — a network error,
a timeout, a crashed client: it renews that agent's claim and returns the first
missing part, so the same agent continues where the saved parts end and nothing is
asked twice. audit lists the advisory findings of one published answer or
of every one. accept records that the warnings were reviewed, with the reason;
reopen asks named parts again, keeps the old answer as a version and makes the
key missing until it is complete again. history lists the versions of one answer
and journal the last events of the interview.
This tool never writes an answer: the numbers come only through elicit. The
orchestrating agent owns the process: it hands one shock to one subagent, reads the
warnings and either accepts the answer or reopens the steps they name.
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": {
"action": "<action>",
"as_of": "<as_of>",
"model_id": "<model_id>"
},
"name": "interview"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| key | No | The input key, one shock or market: named by every action that touches one answer, left out by progress and by the whole-model audit. | |
| agent | No | The name of the answering agent: it owns the claim and is written to the journal; required by claim. | |
| as_of | Yes | The calculation date, YYYY-MM-DD: the interview of this date. | |
| limit | No | How many last journal events journal returns, 50 by default. | |
| steps | No | The parts to ask again, by step id from the question; required by reopen and refused when a step is unknown. | |
| action | Yes | progress, claim, release, restart, audit, accept, reopen, history or journal: the state of an elicitation, not its answers. | |
| reason | No | Why a published answer is accepted or reopened: written to the journal and to the accepted record. | |
| minutes | No | How many minutes a claim stays alive without renewal, 45 by default; a stopped agent does not block its shock forever. | |
| model_id | Yes | The macro model that asks the questions, for example expert.elicited.rw; model(kind="catalog") lists the installed ones. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||