开头优化

v0.1.0

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, det...

0· 100·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for cellinlab/cell-hook-optimizer.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "开头优化" (cellinlab/cell-hook-optimizer) from ClawHub.
Skill page: https://clawhub.ai/cellinlab/cell-hook-optimizer
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install cell-hook-optimizer

ClawHub CLI

Package manager switcher

npx clawhub@latest install cell-hook-optimizer
Security Scan
Capability signals
CryptoCan make purchases
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name, description, and included files (SKILL.md and reference docs) all focus on diagnosing and generating openings for Chinese short-form content. Nothing requested or referenced suggests unrelated capabilities.
Instruction Scope
Runtime instructions are limited to reading the user's provided content and the bundled reference files, diagnosing, and generating multiple hook options. The skill does not instruct the agent to read system files, environment variables, or contact external endpoints.
Install Mechanism
No install spec and no code files — the skill is instruction-only. This minimizes on-disk changes and arbitrary code execution risk.
Credentials
The skill declares no required environment variables, credentials, or config paths. Its needs are proportional to the stated textual-editing purpose.
Persistence & Privilege
always is false and there is no indication the skill modifies agent/system-wide configs or requests persistent privileges. Autonomous invocation is allowed (platform default) but not escalated by this skill.
Assessment
This is an instruction-only skill that operates on text you provide and the included reference files; it does not require credentials or install steps. Before enabling, consider: • Only provide non-sensitive content (the skill will inspect the text you give it). • Because autonomous invocation is allowed by default on the platform, the agent could run this skill when relevant — if you prefer manual control, restrict automatic skill usage in your agent settings. • If you want provenance, review the GitHub repo linked in the metadata to confirm origin and maintenance activity. Otherwise, the skill appears coherent and proportionate to its stated purpose.

Like a lobster shell, security has layers — review code before you run it.

latestvk975sycyawnnbh46k80zms2x8d84jbhy
100downloads
0stars
1versions
Updated 2w ago
v0.1.0
MIT-0

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

Comments

Loading comments...