---
title: "Get Agent Execution Status"
method: GET
path: "/admin/agent-executions/{execution_id}"
tags: ["admin"]
---

# Get Agent Execution Status

`GET /admin/agent-executions/{execution_id}`

## Request

- Base URL: `https://api.claimkit.co` (the document declares no server; this is the origin it was published from)
- URL: `https://api.claimkit.co/admin/agent-executions/{execution_id}`
- Auth: none declared

Return the terminal status of an agent execution (admin only).

The regenerate-review flow polls this to learn the *actual* workflow outcome.
The answer-count-based ``/claims/{id}/generate/status`` can't be trusted for a
regenerate: the answers already exist, so it reports ``completed`` as soon as
the claim leaves ``ai_processing`` — even when the run failed and reverted.
The execution row carries the real ``failed``/``completed`` verdict.

## Path parameters

- `execution_id` integer, required

## Response `200`

Successful Response

- AgentExecutionStatusResponse
  - `id` integer, required
  - `status` string, required
  - `error_message` string, nullable

## Other responses

- `422`: Validation Error

---

[API](https://skmtc.dev/claimkit/apis/claimkit-api.md) - [All operations](https://skmtc.dev/claimkit/apis/claimkit-api/llms.txt) - [OpenAPI document](https://skmtc.dev/claimkit/apis/claimkit-api/revisions/94dc85ba5326?raw)