---
name: "rank-moments"
description: "Use this skill when the task matches: Rank discovered moments into final Opal Sparks for review or rendering."
---
# Rank Moments

## Purpose

Rank discovered moments into final Sparks for review or auto-render.

## Install from SKLx

This skill is published on SKLx and can be installed locally or on an Opal server:

```bash
skl add "rank moments" --registry https://sklx.one
```

## Current API

Ranking is an internal job phase. Poll the job to inspect progress.

```txt
GET /api/jobs/:jobId
```

## Inputs

```json
{
  "jobId": "string",
  "candidates": []
}
```

## Outputs

```json
{
  "jobId": "string",
  "status": "ready_for_review",
  "clips": []
}
```

## Steps

1. Load candidate moments.
2. Rank by hook, clarity, shareability, standalone meaning, pacing, and payoff.
3. Select strongest Sparks.
4. Polish captions.
5. Persist selected Sparks on the job.

## Rules

Do not invent timestamps.
Avoid duplicate ideas.
Prefer clips with clear payoff.
Keep only strong standalone moments.
Return valid JSON from model calls.