## Get a line reputation audit

`client.PhoneNumbers.GetReputationAudit(ctx, auditID, query) (*ReputationAudit, error)`

**get** `/v3/phone_numbers/{phoneNumber}/reputation_audit/{auditId}`

Returns the audit's status and, once complete, the report. Audits are
scoped to the line in the URL — an `auditId` started on a different
line returns `404`.

### Parameters

- `auditID string`

- `query PhoneNumberGetReputationAuditParams`

  - `PhoneNumber param.Field[string]`

    The line in E.164 format.

### Returns

- `type ReputationAudit struct{…}`

  - `AuditID string`

  - `Status ReputationAuditStatus`

    `pending` until the report is ready — poll until `complete` or `error`.

    - `const ReputationAuditStatusPending ReputationAuditStatus = "pending"`

    - `const ReputationAuditStatusComplete ReputationAuditStatus = "complete"`

    - `const ReputationAuditStatusError ReputationAuditStatus = "error"`

  - `Error string`

    Present only when `status` is `error`. Short, generic reason safe to display.

  - `GeneratedAt Time`

    When the report was generated; signals reflect the line at this moment.

  - `Phone string`

    The line audited, E.164.

  - `Report ReputationReport`

    Present only when `status` is `complete`.

    - `ActionItems []ReputationActionItem`

      Ordered by `priority`; 1 = do first.

      - `Detail string`

      - `ExpectedImpact ReputationActionItemExpectedImpact`

        - `const ReputationActionItemExpectedImpactHigh ReputationActionItemExpectedImpact = "high"`

        - `const ReputationActionItemExpectedImpactMedium ReputationActionItemExpectedImpact = "medium"`

        - `const ReputationActionItemExpectedImpactLow ReputationActionItemExpectedImpact = "low"`

      - `Priority int64`

        1 = do first

      - `Title string`

    - `Drivers []ReputationDriver`

      Ranked, highest impact first.

      - `Key ReputationDriverKey`

        Stable driver-category identifier — what is dragging the line, or one
        of its conversations, down.

        - `low_engagement` — The conversation is one-sided: several messages
          sent, few or no replies back. Pause or rework outreach where
          recipients are not replying, and lead with messages that invite a
          response. Conversation-level: it appears on
          `evidence.unhealthy_chats[].driver_keys`, never in `drivers`.
        - `overall_conversation_health` — A large share of the line's active
          conversations are trending unhealthy. Fix the unhealthy conversations
          first — review their content and timing, and whether recipients are
          engaging.
        - `volume_spike` — The line's daily sending volume jumped far above its
          own normal level while few recipients were replying, or exceeded the
          recommended daily volume for a single line. Ramp volume gradually
          instead of spiking, prioritize people who have already engaged with
          you, and spread sustained high volume across additional lines.
        - `new_conversation_rate` — The line is starting too many brand-new
          conversations in a single day. Spread new conversations out over time
          instead of starting many at once.
        - `opt_out_handling` — Recipients asked this line to stop. Honor every
          stop request immediately: send nothing further to that recipient
          unless they opt back in. Every send to them is rejected with `403`
          (error code `2024`), including a final courtesy message — to send
          one telling them they can reply to resume, set
          `override_optout: true` on that single request.
        - `flagged` — The line is currently restricted and its messages may not
          be reaching recipients. Move active traffic to a healthy line now,
          and let this one recover before sending more.
        - `other` — Fallback for a signal without dedicated partner copy.

        - `const ReputationDriverKeyLowEngagement ReputationDriverKey = "low_engagement"`

        - `const ReputationDriverKeyOverallConversationHealth ReputationDriverKey = "overall_conversation_health"`

        - `const ReputationDriverKeyVolumeSpike ReputationDriverKey = "volume_spike"`

        - `const ReputationDriverKeyNewConversationRate ReputationDriverKey = "new_conversation_rate"`

        - `const ReputationDriverKeyOptOutHandling ReputationDriverKey = "opt_out_handling"`

        - `const ReputationDriverKeyFlagged ReputationDriverKey = "flagged"`

        - `const ReputationDriverKeyOther ReputationDriverKey = "other"`

      - `Metric string`

        A specific observed figure when available; otherwise a short qualitative note.

      - `Summary string`

        One plain-English sentence.

    - `Evidence ReputationEvidence`

      The specific conversations behind the drivers, so partners can verify every claim against their own send logs. Each `chat_id` can be fetched via `GET /v3/chats/{chatId}` — its current health appears there.

      - `OptOutChats []ReputationOptOutChat`

        Worst first — most messages sent after the stop request; honor these immediately.

        - `ChatID string`

        - `MessagesAfterStop int64`

          Outbound messages sent after the recipient asked to stop.

      - `UnhealthyChats []ReputationUnhealthyChat`

        Up to 15, worst first.

        - `ChatID string`

        - `DriverKeys []ReputationDriverKey`

          What is dragging this conversation down, in the same vocabulary as the report's drivers. Each key's meaning and the fix for it are documented on `ReputationDriverKey`.

          - `const ReputationDriverKeyLowEngagement ReputationDriverKey = "low_engagement"`

          - `const ReputationDriverKeyOverallConversationHealth ReputationDriverKey = "overall_conversation_health"`

          - `const ReputationDriverKeyVolumeSpike ReputationDriverKey = "volume_spike"`

          - `const ReputationDriverKeyNewConversationRate ReputationDriverKey = "new_conversation_rate"`

          - `const ReputationDriverKeyOptOutHandling ReputationDriverKey = "opt_out_handling"`

          - `const ReputationDriverKeyFlagged ReputationDriverKey = "flagged"`

          - `const ReputationDriverKeyOther ReputationDriverKey = "other"`

        - `Status ReputationUnhealthyChatStatus`

          The conversation's current health — the same value `GET /v3/chats/{chatId}` reports for it.

          - `const ReputationUnhealthyChatStatusAtRisk ReputationUnhealthyChatStatus = "AT_RISK"`

          - `const ReputationUnhealthyChatStatusCritical ReputationUnhealthyChatStatus = "CRITICAL"`

          - `const ReputationUnhealthyChatStatusOptedOut ReputationUnhealthyChatStatus = "OPTED_OUT"`

    - `PrimaryDriver string`

      The `key` of the most important driver. Empty string when the line has nothing to act on — the report then carries a single reassurance action item. Its values are the `ReputationDriverKey` vocabulary — see that schema for what each means and what to do about it.

    - `Severity ReputationReportSeverity`

      Current reputation of this phone line.

      - `HEALTHY` — The line is in good standing. Send normally.
      - `AT_RISK` — Warning signs on the line: engagement is low across many of its conversations, or it's starting too many brand-new conversations in a single day — and a spike in send volume can add to either. Slow the line's send pace, avoid opening many new conversations at once, and review your messaging patterns.
      - `CRITICAL` — Strong signals that messages from this line aren't landing well. Pause outbound on the line until it recovers.

      Defaults to `HEALTHY` for lines that have not yet been scored.

      - `const ReputationReportSeverityHealthy ReputationReportSeverity = "HEALTHY"`

      - `const ReputationReportSeverityAtRisk ReputationReportSeverity = "AT_RISK"`

      - `const ReputationReportSeverityCritical ReputationReportSeverity = "CRITICAL"`

    - `SummaryMarkdown string`

      Deterministic markdown rendering of this report, suitable for feeding directly to automated systems and AI agents as investigation context. Rendered from the structured fields above, which remain the source of truth.

### Example

```go
package main

import (
  "context"
  "fmt"

  "github.com/linq-team/linq-go"
  "github.com/linq-team/linq-go/option"
)

func main() {
  client := linqgo.NewClient(
    option.WithAPIKey("My API Key"),
  )
  reputationAudit, err := client.PhoneNumbers.GetReputationAudit(
    context.TODO(),
    "auditId",
    linqgo.PhoneNumberGetReputationAuditParams{
      PhoneNumber: "phoneNumber",
    },
  )
  if err != nil {
    panic(err.Error())
  }
  fmt.Printf("%+v\n", reputationAudit.AuditID)
}
```

#### Response

```json
{
  "audit_id": "audit_id",
  "status": "pending",
  "error": "error",
  "generated_at": "2019-12-27T18:11:19.117Z",
  "phone": "phone",
  "report": {
    "action_items": [
      {
        "detail": "detail",
        "expected_impact": "high",
        "priority": 0,
        "title": "title"
      }
    ],
    "drivers": [
      {
        "key": "low_engagement",
        "metric": "metric",
        "summary": "summary"
      }
    ],
    "evidence": {
      "opt_out_chats": [
        {
          "chat_id": "chat_id",
          "messages_after_stop": 0
        }
      ],
      "unhealthy_chats": [
        {
          "chat_id": "chat_id",
          "driver_keys": [
            "low_engagement"
          ],
          "status": "AT_RISK"
        }
      ]
    },
    "primary_driver": "primary_driver",
    "severity": "HEALTHY",
    "summary_markdown": "## Line reputation summary\n\n**Status:** AT_RISK\n..."
  }
}
```
