Coach reps on their real weak spots, backed by Chorus scorecards

**Objective:** Build a coaching review for `{rep-name}` covering the last `{lookback-days}` days (default 90): what they have been scored on, where the scores are weak, and what has already been said to them.

**Step 1: Resolve the person**

Call `chorus-get-users` with `name.contains` set to `{rep-name}`. If more than one active user matches, list name, email, title and role and stop, asking which. Keep the exact `name` and the `email`, since the scorecard tools match on the full name and the conversation tools match on the email address.

**Step 2: Read the scorecards**

Call `chorus-get-scorecards` with `recipient-name` set to the exact name from step 1 and `submitted.from` set to the start of the window. Note `initiative-name`, `reviewer-name`, `submitted` and `scorecard-id`. If there are none, say so plainly, since an unreviewed rep is the finding.

**Step 3: Drill into the scores**

For each scorecard, call `chorus-get-scorecard-scores` with `scorecard-id` set to that scorecard's `scorecard-id`. Collect `question`, `score` and `comment`. Then work out, across all the scorecards together, which questions score lowest on average and which ones move over time.

**Step 4: Read the coaching already given**

Call `chorus-get-moments` with `recipient-name` set to the exact name and `submitted.from` set to the start of the window. These are the coaching moments already recorded. Note `initiative-name`, `reviewer-name`, `description` and `submitted`.

**Step 5: Ground it in activity**

Call `chorus-get-engagements` with `user-email` set to the email from step 1 and `date.from` set to the start of the window, to see how many conversations the scores are drawn from and how long they run.

**Step 6: Report**

Produce:

* **Summary** - how many conversations, how many scorecards, average score, and the direction of travel.
* **Strengths** - the two or three highest scoring questions, with a reviewer comment quoted for each.
* **Development areas** - the two or three lowest scoring questions, with the comments quoted.
* **Coaching already given** - what has been raised before, and whether the scores since suggest it landed.
* **Recommended focus** - one area to work on next, and why.

**Rules & Guidelines:**

* This is read only. Do not call `chorus-add-moment`, `chorus-add-playlist-moment`, `chorus-create-playlist`, `chorus-update-playlist`, `chorus-update-playlist-moment`, `chorus-update-moment`, `chorus-export-recordings` or `chorus-export-scorecards`.
* Do not average across different scorecard types as though they were one scale. Group by `scorecard-type` first and say so if they differ.
* Do not infer intent from a low score. Quote the reviewer comment or say there was none.

**Validation:**

* If no user matches `{rep-name}`, say so and suggest searching by part of the name.
* If the user exists but has no scorecards in the window, report the conversation activity alone and flag the review gap.
* If fewer than three scorecards exist, say the sample is too small to read a trend from.