agentleFS
cellinlabcellinlab/cell-skills/skills/hook-optimizer/SKILL.md

hook-optimizer

Skill29 starsChanged 7 months agoMITView on GitHub

Hook and opening optimization for Chinese short-form content. Use when Codex needs to diagnose why the first 3-5 seconds or first 2-3 sentences are weak, determine whether the real problem is the opening or the material underneath it, and generate multiple opening directions that match the actual content instead of empty clickbait.

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. About 0 are common, 17 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈17

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when the user already has a topic, body, or draft, but the opening is not doing enough work.

    line 11
  2. Distinctive:

    Check whether the content underneath the opening is strong enough.

    line 23
  3. Distinctive:

    If the content itself is too thin, say so before generating hooks.

    line 29
  4. Distinctive:

    prefer openings that the body can actually cash

    line 38
  5. Distinctive:

    If the user only gave a title-like topic with no body or material, say the input is too thin for serious opening work.

    line 51
  6. Distinctive:

    Look for usable fuel:

    line 55
  7. Distinctive:

    When the content has none of the above, say the hook will be limited because the body is under-fueled.

    line 63
  8. Distinctive:

    Read references/hook-principles.md when deciding whether the problem is the opening or the content itself.

    line 65
  9. Distinctive:

    If the user already has one, check:

    line 69
  10. Distinctive:

    Read references/hook-types.md when comparing directions.

    line 89
  11. Distinctive:

    keep them distinct rather than slightly reworded copies

    line 99
  12. Distinctive:

    If one direction clearly dominates, say so and explain why.

    line 101
  13. Distinctive:

    Default to assets/hook-pack-template.md.

    line 105
  14. Distinctive:

    keep the opening aligned with the actual content

    line 129
  15. Distinctive:

    Read for the base rules of what a strong opening must do and how to detect weak-content cases.

    line 136
  16. Distinctive:

    Read for the main hook families, fit conditions, and failure modes.

    line 138
  17. Distinctive:

    Use for the standard output shape.

    line 140

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Hook Optimizer

Overview

Use this skill when the user already has a topic, body, or draft, but the opening is not doing enough work.

The opening should help the audience answer three questions quickly:

  • what is this about
  • why should I continue
  • why should I trust or care about this voice

This skill diagnoses first and generates later.

Quick Start

  1. Check whether the content underneath the opening is strong enough.
  2. Extract the best available materials before rewriting the opening.
  3. Diagnose the current opening if one exists.
  4. Generate several opening directions, not one magic line.
  5. Pick the strongest direction based on fit, not on loudness.

If the content itself is too thin, say so before generating hooks.

Default Contract

Assume the following unless the user says otherwise:

  • write in Chinese
  • optimize the opening, not the whole piece
  • diagnose before generating
  • prefer openings that the body can actually cash
  • for spoken scripts, keep mouth-feel and breath rhythm in mind

Workflow

Phase 1: Read the Input State

Determine whether the user provided:

  • an existing opening and a full body
  • only the body
  • a topic plus a few key materials

If the user only gave a title-like topic with no body or material, say the input is too thin for serious opening work.

Phase 2: Check Whether the Content Can Carry a Hook

Look for usable fuel:

  • a result or concrete outcome
  • a contradiction or reversal
  • a strong scene or vivid moment
  • a number, comparison, or unusual detail
  • a direct judgment worth arguing with

When the content has none of the above, say the hook will be limited because the body is under-fueled.

Read references/hook-principles.md when deciding whether the problem is the opening or the content itself.

Phase 3: Diagnose the Current Opening

If the user already has one, check:

  • independence: can it work even if the title is unseen
  • grip: does it give the audience a reason to stay
  • curiosity or tension: does it hold something back
  • credibility: why believe or care about this speaker
  • flow: is it easy to say or read
  • payoff fit: can the body deliver what the opening promises

Phase 4: Choose the Right Hook Type

Select from the smallest set that truly fits:

  • result-led
  • contradiction-led
  • scene-led
  • question-led
  • judgment-led
  • suspense-led

Read references/hook-types.md when comparing directions.

Phase 5: Generate a Hook Pack

Produce 3-4 strategic directions.

For each direction:

  • explain why it fits this content
  • generate 2-4 opening options
  • keep them distinct rather than slightly reworded copies

If one direction clearly dominates, say so and explain why.

Output Format

Default to assets/hook-pack-template.md.

At minimum, include:

  • content viability note
  • current opening diagnosis
  • extracted materials
  • multiple hook directions
  • top recommendations

Hard Rules

Do not:

  • pretend a weak body can be rescued by a flashy opening
  • depend on the title to complete the meaning
  • reveal the full answer too early when tension is needed
  • generate hooks the body cannot pay off
  • return one "best line" without showing alternatives

Always:

  • diagnose before generating
  • extract the best available materials first
  • keep the opening aligned with the actual content
  • mention when the problem is upstream from the opening
  • preserve readability and speakability

Resource Map

  • references/hook-principles.md
    • Read for the base rules of what a strong opening must do and how to detect weak-content cases.
  • references/hook-types.md
    • Read for the main hook families, fit conditions, and failure modes.
  • assets/hook-pack-template.md
    • Use for the standard output shape.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.