Coach reps with OnePageCRM and Chorus evidence together

**Objective:** For the rep with email `{rep-email}`, over the last `{lookback-days}` days (default 30), compare what they logged in OnePageCRM with what Chorus recorded and how it was scored.

**Step 1: Identify the rep in both systems**

Onepagecrm server: call `onepagecrm-get-user` with `email` equal to `{rep-email}` and keep id and name. Chorus server: call `chorus-get-users` with `email` equal to `{rep-email}` and keep name and user-id. If either returns nothing, say so and continue with the other side only.

**Step 2: CRM activity (onepagecrm server)**

Call `onepagecrm-get-action` with `assignee-id` set to the CRM user id and `due-date.from` set to the window start. Call `onepagecrm-get-deal` with `owner-id` set to the same id. Count completed and overdue actions and list open deals.

**Step 3: Recorded conversations and scores (chorus server)**

Call `chorus-get-engagements` with `user-email` set to `{rep-email}` and `date.from` set to the window start. Call `chorus-get-scorecards` with `recipient-name` set to the Chorus name, and `chorus-get-scorecard-scores` with the same `recipient-name` for question-level results.

**Step 4: Report**

* Volume: conversations recorded against actions completed and deals owned.
* Accounts the rep has open deals on but no recorded conversation in the window.
* Scorecards: count, average score, and the two lowest-scoring questions with a short reviewer comment each.
* Two or three coaching observations tied to evidence from both systems.

**Rules & Guidelines:**

* `onepagecrm-get-user`, `onepagecrm-get-action` and `onepagecrm-get-deal` are on the onepagecrm server; `chorus-get-users`, `chorus-get-engagements`, `chorus-get-scorecards` and `chorus-get-scorecard-scores` are on the chorus server.
* Read-only. Never call `onepagecrm-update-action`, `onepagecrm-add-action` or any chorus add, update or export tool.
* This is a coaching aid, not a performance rating. Do not rank the rep against others.

**Validation:**

* If the email matches several users in either system, list them and stop.
* Low counts: state the sample size before drawing conclusions.