Docs/Skills and code/skill script-cohort-analysis

.claude/skills/script-cohort-analysis/references/packaging-prompt.md

Markdown37 lines1.4 KBSkill: script-cohort-analysis

Blind packaging rater instructions

You will see only a list of video titles from one YouTube channel, with labels P1, P2, … You haven't seen the videos, and you don't know how they performed. Don't try to recall whether you know these videos.

For each title, judge how strongly it would make a typical viewer of this kind of channel click when the video appears on their home feed, based on the title alone (no thumbnail).

Consider:

• Curiosity gap: a specific, unanswered question or a tension.
• Clarity: you instantly know what the video is about.
• Stakes and relevance: why it matters to the viewer personally (money, life decisions, identity).
• Breadth of appeal: how many people would care about the topic.
• Clutter and gimmicks: format tags, keyword stuffing, sensational words with no specifics.

Score each title 1–5:

• 1: weak or unclear
• 3: a typical decent title
• 5: exceptional and hard to ignore

Then give a full ranking, where 1 is the strongest.

Output a single JSON file:

{
  "titles": [
    { "pid": "P1", "score": 3, "rationale": "one sentence" }
  ],
  "ranking": ["P3", "P1", "..."],
  "notes": "patterns you noticed, one or two sentences"
}

After writing, run the validation command from your task file. Your final message is one line: the path and the validator result.

# Blind packaging rater instructions

You will see only a list of video titles from one YouTube channel, with labels P1, P2, …
You haven't seen the videos, and you don't know how they performed. Don't try to recall
whether you know these videos.

For each title, judge how strongly it would make a typical viewer of this kind of channel
click when the video appears on their home feed, based on the title alone (no thumbnail).

Consider:
- **Curiosity gap**: a specific, unanswered question or a tension.
- **Clarity**: you instantly know what the video is about.
- **Stakes and relevance**: why it matters to the viewer personally (money, life decisions, identity).
- **Breadth of appeal**: how many people would care about the topic.
- **Clutter and gimmicks**: format tags, keyword stuffing, sensational words with no specifics.

Score each title 1–5:
- **1**: weak or unclear
- **3**: a typical decent title
- **5**: exceptional and hard to ignore

Then give a full ranking, where 1 is the strongest.

Output a single JSON file:

```json
{
  "titles": [
    { "pid": "P1", "score": 3, "rationale": "one sentence" }
  ],
  "ranking": ["P3", "P1", "..."],
  "notes": "patterns you noticed, one or two sentences"
}
```

After writing, run the validation command from your task file. Your final message is one line: the path and the validator result.

Shown as it is in the repository at build time (2026-11-03); the home directory appears as <home>. If this page and the file differ, the file is right. Back to the catalogue.