Operations
Operations

Override run verdict

Set a manual verdict on an eval run (success / failure / review_required) without overwriting the engine's computed result, it writes a separate override row, one per run. Use this when the automated pass/fail is wrong or a run needs a human to look. The run must belong to the current agent.

POST
/operations/override_run_verdict

Authorization

bearerAuth
AuthorizationBearer <token>

A personal access token from Settings > API tokens.

In: header

Header Parameters

Idempotency-Key?string

A key you choose, such as a UUID, that makes this POST safe to retry. Within 24 hours the same key and request answers with the first successful answer and runs nothing, the same key with a different request answers 422 idempotency_key_reused, and a key whose request is still running answers 409 idempotency_key_in_use. A failed request keeps nothing, so the same key runs it again.

Length1 <= length <= 255

Request Body

application/json

TypeScript Definitions

Use the request body type in TypeScript.

Response Body

application/json

application/problem+json

application/problem+json

application/problem+json

application/problem+json

application/problem+json

application/problem+json

application/problem+json

curl -X POST "https://example.com/operations/override_run_verdict" \  -H "Content-Type: application/json" \  -d '{    "runId": "string",    "verdict": "success",    "agent": "string"  }'
{}
{
  "type": "http://example.com",
  "title": "string",
  "status": 0,
  "detail": "string",
  "code": "string",
  "errors": [
    {
      "path": "string",
      "message": "string"
    }
  ],
  "estimate": {},
  "retryAfterSeconds": 0
}
{
  "type": "http://example.com",
  "title": "string",
  "status": 0,
  "detail": "string",
  "code": "string",
  "errors": [
    {
      "path": "string",
      "message": "string"
    }
  ],
  "estimate": {},
  "retryAfterSeconds": 0
}
{
  "type": "http://example.com",
  "title": "string",
  "status": 0,
  "detail": "string",
  "code": "string",
  "errors": [
    {
      "path": "string",
      "message": "string"
    }
  ],
  "estimate": {},
  "retryAfterSeconds": 0
}
{
  "type": "http://example.com",
  "title": "string",
  "status": 0,
  "detail": "string",
  "code": "string",
  "errors": [
    {
      "path": "string",
      "message": "string"
    }
  ],
  "estimate": {},
  "retryAfterSeconds": 0
}
{
  "type": "http://example.com",
  "title": "string",
  "status": 0,
  "detail": "string",
  "code": "string",
  "errors": [
    {
      "path": "string",
      "message": "string"
    }
  ],
  "estimate": {},
  "retryAfterSeconds": 0
}
{
  "type": "http://example.com",
  "title": "string",
  "status": 0,
  "detail": "string",
  "code": "string",
  "errors": [
    {
      "path": "string",
      "message": "string"
    }
  ],
  "estimate": {},
  "retryAfterSeconds": 0
}
{
  "type": "http://example.com",
  "title": "string",
  "status": 0,
  "detail": "string",
  "code": "string",
  "errors": [
    {
      "path": "string",
      "message": "string"
    }
  ],
  "estimate": {},
  "retryAfterSeconds": 0
}

Optimise agent prompt POST

Start the Optimise-Prompt loop for this agent: diagnose the runs the agent is FAILING on the selected evaluators, make scoped edits to the system prompt, apply each edit, and re-validate over the SAME scenarios, looping until the agent passes or a stop condition fires. An overfitting gate scrubs every edit BEFORE it's applied (it rejects edits that embed scenario IDs, transcript-cloned phrasing, hardcoded test data, or hyper-narrow single-case clauses) so the loop can't game its own validation set. On a 100% pass it sweeps the full eval set before declaring success. The loop runs in the background and writes an improvement session you poll; it returns the sessionId. It NEVER auto-applies an edit without a re-run proving the pass rate went up, and it NEVER publishes: a winning edit is left as a proposed session for an operator to accept. Set selectedEvalIds to the failing evaluators; maxIterations (default 10) caps the attempts; mode/autoMode controls whether the run is attended.

Permanently delete agent POST

Permanently delete this agent, IRREVERSIBLE. Removes the agent and cascades to every version, tool, variable, eval, and the call-transcript links. There is no recovery. Only an archived agent can be deleted: call `archive_agent` first, which stops its calls and unassigns its numbers (they stay in the workspace), the same two steps as the Archive or delete page. Prefer archiving alone (reversible) unless the agent must be erased for good. Over MCP this requires `confirm: true`.