---
search:
  tags:
    - V1
    - PATCH
seo:
  description: >-
    Edit fields, refresh evidence, comment on, or change the… Reference for the
    PATCH /v1/projects/{projectId}/issues/{issueId} endpoint in the Judgeval
    API.
sidebar:
  label: Update issue
  badge: PATCH
title: Update issue
type: openapi-operation
---
Edit fields, refresh evidence, comment on, or change the review status of one issue.

`PATCH /v1/projects/{projectId}/issues/{issueId}`

**Request body** (`application/json`, required)

- `title` (string) — Short issue title.
- `diagnosis` (string) — 1–2 short sentences describing the observed failure, with inline trace and relevant signal citations. Omit fixes and investigation history.
- `impact_summary` (string | null) — 2–3 Markdown bullets on impact, scope, frequency, trends, and material uncertainty. Quantify how often contributing labels were applied, and any other judge signals, with inline citations. Combine related facts; omit immaterial unknowns and labeled checklists.
- `priority` (string)
- `production_activity` (string)
- `agent_reasoning` (string | null) — 1–3 Markdown bullets suggesting evidence-backed remediation. Omit repeated evidence and claims of an exact cause or fix.
- `cited_items` (object[]) — Cite every supporting trace with a one-sentence failure description in its \`label\` field and an optional short quote. Add a span or trace-behavior citation for each contributing location, using the same trace\_id and real span\_id when available; use trace-only evidence only when no deeper location is reliable. For contributing label or judge signals, include their citation with the stable name in its \`label\` field and matching trace-behavior verdicts with tool-returned behavior\_id/judge\_id and value. Cite the label itself with type "behavior" (behavior\_id) and its verdict on a trace with type "trace-behavior". Put changing metrics and confidence in impact\_summary.
- `user_status` (string)
- `judge_id` (string) — Monitoring Judge ID; required for monitored status.
- `behavior_id` (string) — ID of the label the production judge applies; required for monitored status. The field name remains behavior\_id.
- `message` (string) — Timeline comment, alone or with other changes.

Request body example:

```json
{
  "title": "string",
  "diagnosis": "string",
  "impact_summary": "string",
  "priority": "low",
  "production_activity": "new",
  "agent_reasoning": "string",
  "cited_items": [
    {
      "type": "string",
      "property1": null,
      "property2": null
    }
  ],
  "user_status": "needs_review",
  "judge_id": "string",
  "behavior_id": "string",
  "message": "string"
}
```

**Responses**

- `200` — Response for status 200

Response example, 200:

```json
{
  "issue": {
    "id": "string",
    "organization_id": "string",
    "project_id": "string",
    "issue_key": "string",
    "title": "string",
    "diagnosis": "string",
    "impact_summary": "string",
    "priority": "low",
    "user_status": "needs_review",
    "production_activity": "new",
    "first_seen_at": "string",
    "last_seen_at": "string",
    "evidence_count": 0,
    "agent_reasoning": "string",
    "monitoring_judge_id": "string",
    "monitoring_behavior_id": "string",
    "created_at": "string",
    "updated_at": "string"
  }
}
```