> ## Documentation Index
> Fetch the complete documentation index at: https://docs.60db.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# List Healing Changes

> Changes that self healing drafted, applied or measured for an analyzer

Returns the newest 50 changes self healing made for the AI agents this analyzer watches, with the evidence and the instructions before and after. See [Self healing](https://docs.60db.ai/features/call-analysis.md#self-healing).

> [!NOTE]
>
> Needs view access to the analyzer. `can_manage` is true when the caller can also approve, dismiss and undo (manage access plus permission to edit agents). Not available to API keys.

## Request

### Headers

**Authorization** `string` *required*

Bearer token of a signed-in user (dashboard session). API keys cannot call healing endpoints

### Path

**agentId** `string` *required*

Analyzer ID (UUID)

## Response

**200**

```json theme={null}
{
  "success": true,
  "data": {
    "items": [
      {
        "id": "<change-uuid>",
        "agent": { "id": "<ai-agent-uuid>", "name": "Front desk" },
        "status": "proposed",
        "applied_by": null,
        "created_at": "2026-11-13T09:00:00.000Z",
        "applied_at": null,
        "decided_at": null,
        "problems": [
          { "category": "wrong_information", "severity": "high", "reason": "Quoted the old price", "count": 3, "call_ids": ["<call-uuid>"] }
        ],
        "changes": ["State the current price from the price list"],
        "previous_script": "…",
        "new_script": "…",
        "base_hash": "<sha256 of previous_script>",
        "verification": {
          "target_calls": 10,
          "baseline": { "calls": 20, "problem_calls": 3, "rate": 0.15 },
          "after": null,
          "outcome": null
        },
        "can_manage": true
      }
    ],
    "pending_review_count": 1
  }
}
```

| `status` | Meaning |
| - | - |
| `proposed` | Waiting for review |
| `verifying` | Applied; measuring the next `target_calls` calls |
| `improved` / `no_change` | Measured; the change stays |
| `rolled_back` | Undone automatically because the problem got worse |
| `superseded` | The agent was edited by someone else, so the change was neither applied nor undone |
| `undone` / `dismissed` | A person undid or dismissed it |
| `failed` | The update of the agent was refused |

`applied_by` is `auto` or `person`. `call_ids` are call IDs you can open in call analysis.

## Errors

| Status | Code | When |
| - | - | - |
| 404 | `NOT_FOUND` | The analyzer does not exist or is not visible |
| 503 | `AUTHZ_UNAVAILABLE` | The permission check could not be made |


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.