v4

latestOpenAPI 3.1.02026-08-08349225311.7 KB

runtime_mutations

ckg

Add Evidence Endpoint

Add a notebook-captured evidence claim for an untracked variable.

Kernel-only on both mounts: this local route requires the kernel API bearer token; the cloud mirror requires an internal principal.

Rejections (unsupported value type, no session for the notebook, bad variable name) come back in rejected_reason rather than via HTTP error so the client can surface them in-cell without raising.

post/api/evidence

Request body

notebook_idstring required

Notebook the value was produced in.

descriptionstring required

Brief narrative of what the value shows.

variable_namestring required

Notebook variable name.

{"stackTrail":"components:schemas:AddEvidenceRequest:properties:value","oasType":"schema","type":"unknown","title":"Value","description":"The value to record. Same shape as qualia.submit_variable."}

cell_idstring nullable

Producing cell id when known; cited as the source.

task_idstring nullable

Optional task to associate the evidence claim with.

recall_priority'global' | 'high' | 'medium' | 'low'

How readily a claim should enter an agent's context.

This is a query-independent prior authored by the agent that files the claim. It is not relevance, confidence, or truth: a Low claim that is highly relevant to the current work must still be able to outrank an irrelevant Global one once the recall system ranks candidates.

GLOBAL is the only value with delivery semantics rather than ranking semantics: it is injected unconditionally into every agent in the workspace. Every claim type carries a priority — evidence, figures, and findings can all be the thing future agents most need to see.

Response

Successful Response

claim_idstring nullable

Created claim id, or null when the request was rejected.

rejected_reasonstring nullable

Why the request was rejected, when claim_id is null.