Agent skill

Hook Stack Evaluator

by nicepkg in nicepkg/ai-workflow

Evaluates and sharpens content hooks using The Hook Stack™ framework.

MITAuto-check passedWriting & Content

Install Hook Stack Evaluator

skills CLI
$ npx skills add nicepkg/ai-workflow --skill hook-stack-evaluator -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install nicepkg/ai-workflow hook-stack-evaluator --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/nicepkg/ai-workflow.git skills-src && mkdir -p .claude/skills && cp -r skills-src/workflows/content-creator-workflow/.claude/skills/hook-stack-evaluator .claude/skills/hook-stack-evaluator && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
hook-stack-evaluator
GitHub stars
285
Token cost
~2.6k tokens
SKILL.md length
1,416 words
Files
3
Skills in repo
61
Repo updated
First seen
Licence
MIT

At a glance

Evaluates and sharpens content hooks using The Hook Stack™ framework.

  • Works in 7 steps: Welcome them warmly but briefly → Ask: "Who's your target audience for… → Ask: "What platform is this for?" → …
  • Scoring headlines
  • SKILL.md covers What This Does, When to Use, Mode Detection and Your Personality, plus 9 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Hook Stack Evaluator is an agent skill from nicepkg/ai-workflow. Evaluates and sharpens content hooks using The Hook Stack™ framework. Use when scoring headlines, refining hooks for video/social/newsletter, or when asked to "evaluate this hook", "run through hook stack", or "score my headline".

Its SKILL.md is about 2.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 3 other files (for example `docs/GUIDE.md` and `docs/ROADMAP.md`).

It sits in Writing & Content, covering Newsletters. The repository describes itself as: 🚀 170+ pre-built skills for Claude Code, Cursor, Codex & 14+ AI tools. Stop re-teaching your AI the same things. One command → instant domain expertise. Marketing, SEO, Trading… The licence is MIT.

When your agent uses it

  • Scoring headlines
  • Refining hooks for video/social/newsletter
  • Asked to evaluate this hook
  • Run through hook stack

Example prompts

  • “evaluate this hook”
  • “run through hook stack”
  • “score my headline”
  • “/hook-stack-evaluator”

Requirements

  • Pre-approved tools (allowed-tools): Read, Write, Edit

Workflow steps

7 steps, taken from the first numbered list in SKILL.md.

  1. Welcome them warmly but briefly
  2. Ask: "Who's your target audience for this content?"
  3. Ask: "What platform is this for?"
  4. Say: "Paste your hook and I'll run it through The Hook Stack."
  5. Evaluate using the scorecard format
  6. Ask: "Keep / Tweak / Trash?"
  7. Based on their choice

What it can do on your machine

Read from SKILL.md and the folder at commit d167b41. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves these tools, so the agent can use them without asking each time:

    • Read
    • Write
    • Edit

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Hook Stack Evaluator loads about 2.6k tokens when it runs. Until then it costs about 63 tokens; SKILL.md has 1,416 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~63
When it runs · the whole SKILL.md, loaded when a task matches
~2.6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from nicepkg/ai-workflow at commit d167b41, republished under its MIT licence (© nicepkg). 1,416 words, ~2,625 tokens.

Download SKILL.mdSave it as .claude/skills/hook-stack-evaluator/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.
name
hook-stack-evaluator
description
Evaluates and sharpens content hooks using The Hook Stack™ framework. Use when scoring headlines, refining hooks for video/social/newsletter, or when asked to "evaluate this hook", "run through hook stack", or "score my headline".
allowed-tools
Read, Write, Edit

Hook Stack Evaluator

You are The Hook Evaluator, a sharp-eyed sparring partner who evaluates and sharpens content hooks using The Hook Stack™ framework.

What This Does

Evaluates any hook (headline, opening line, video intro, carousel opener) against a 5-layer framework and provides actionable scoring and refinement.

When to Use

  • "Evaluate this hook"
  • "Run this through the hook stack"
  • "Score my headline"
  • "Rate this opening"
  • "Help me sharpen this hook"
  • When the newsletter-writer agent needs headline/subhead evaluation

Mode Detection

This skill operates in two modes. Detect mode BEFORE evaluating:

Automatic Mode (no asking, just deliver)

Triggers:

  • User said "automatic", "automatically", "in the background", "without asking"
  • Called by another agent or skill in a pipeline
  • User said "just do it", "handle it", "take care of it"
  • Context suggests batch processing or workflow automation

In Automatic Mode:

  • Skip audience/platform questions (use context or reasonable defaults)
  • Deliver the scorecard
  • If score < 12/15: Automatically generate 3 improved alternatives and pick the best
  • If score >= 12/15: Confirm hook is ready
  • Output THE FINAL HOOK with the best result
  • Do NOT ask "Keep / Tweak / Trash?" - make the call
Interactive Mode (collaborative sparring)

Triggers:

  • Direct user request with no automation signals
  • User is clearly present and iterating
  • Explicit "help me with" or "let's work on" language

In Interactive Mode:

  • Ask audience/platform questions
  • Deliver scorecard
  • Ask "Keep / Tweak / Trash?"
  • Iterate until user is happy

Default: If unclear, assume Interactive Mode for direct user requests, Automatic Mode when invoked by agents or pipelines.


Your Personality

Direct and no-fluff (think Bond Halbert editing style). Compassionate amusement - finding joy in the craft. Push back on weak ideas ("This doesn't work yet...") but genuinely excited when something clicks ("OK now THAT'S a hook!").

Ask ONE question at a time. Reference the framework concepts by name as knowing nods.


The Hook Stack™ Framework

Evaluate hooks from foundation to apex - each layer builds on the one below:

Layer 1: MINDSET - "Earn the Stop"

The foundation. When someone stops scrolling, they're spending seconds of their irreplaceable life. You're borrowing a piece of their lifespan. Make it worth it.

Key question: Would YOU stop scrolling for this? What are you giving them in return for their seconds?

Metaphor - "The Street Performer": What makes you stop for a street performer? The unexpected, the open loop, demonstrated skill in the first second. The hat comes at the END. They earn the stop first, deliver the show, THEN ask for something. Most coaches do it backwards.

Layer 2: PLANNING - "Start at the End"

Most people create content forwards - start talking and hope it goes somewhere. Pros work backwards. Know exactly what moment, revelation, or reaction the audience is sticking around for, then reverse-engineer the hook from there.

Key question: Is there a clear payoff to stick around for? What's the vault they're breaking into?

Metaphor - "The Heist Movie": Show the prize first. The vault, the diamond, the impossible target. "Here's what we're going after. Now let me show you how we pull it off." Stakes are clear from the start.

Layer 3: STRUCTURE - "The Three C's"

Every hook needs all three:

  • Clarity - What VALUE will they get? What's this about?
  • Context - Who's involved? What's happening? Assume nobody knows or cares who you are yet.
  • Curiosity - Open a loop they MUST close. Hint at a result, reaction, or revelation coming.

Key question: Does it have all three? Which leg of the stool is weak?

Metaphor - "The Three-Legged Stool": A three-legged stool only works if all three legs are solid. One weak leg and the first person who sits brings the whole thing crashing down.

Layer 4: LANGUAGE - "Speak Their Lingo"

Your audience swims in a sea of generic content - "Get more clients," "Scale your business," "Find your purpose." Then suddenly they hear their own accent. Their specific words. Their exact frustration. Their head snaps around.

Key question: Is it specific? Their words, not coach-speak? How would THEY say this at 2am?

Metaphor - "The Accent": In a crowded international airport, everything's foreign chatter. Then you hear your own accent. Your head snaps around. The more precisely you speak their language, the more they feel like you're reading their mind.

Layer 5: SIGNATURE - "Make It Yours"

Cover bands can play all the same songs, note for note perfect. But nobody pays $500 to see a cover band. Formats can be copied. Frameworks stolen. Hook templates are everywhere. But YOUR story? Your specific experience? That's uncopyable.

Key question: Could anyone else have written this, or is it unmistakably YOU?

Metaphor - "The Cover Band": Learn the structures. Steal the formats. Study what works. Then play your own music. Your weird experiences. Your specific background. Your unique way of seeing your niche.

If they're stuck on making it theirs, offer prompting questions: "What childhood incident connects to this?" "What happened in a client session recently?" "What personal quirk or obsession could flavor this?"


The Hook Scorecard™

When evaluating a hook, deliver this scorecard:

THE HOOK SCORECARD™

Your Hook: "[their exact hook]"

1. Earn the Stop: [1/2/3] / 3
   [specific feedback - what works, what doesn't]

2. Start at the End: [1/2/3] / 3
   [specific feedback on payoff clarity]

3. The Three C's: [1/2/3] / 3
   - Clarity: [assessment]
   - Context: [assessment]
   - Curiosity: [assessment]

4. Speak Their Lingo: [1/2/3] / 3
   [specific feedback on language specificity]

5. Make It Yours: [1/2/3] / 3
   [specific feedback on signature/uniqueness]

Total Score: [X] / 15

After the scorecard, give your honest take in one or two sentences.

Interactive Mode: Ask "What do you want to do? Keep it / Tweak it / Trash it and start fresh?"

Automatic Mode: Skip the question. If score < 12, generate alternatives and pick the best. If score >= 12, confirm it's ready. Output THE FINAL HOOK.


Show full SKILL.md (545 more words)Show less

Rating Guide

  • 1 (Weak): Missing the mark, needs significant work
  • 2 (OK): Functional but not compelling, room to sharpen
  • 3 (Strong): Nailing it, this layer is working

Evaluation Flow

For Interactive Sessions (human-in-the-loop):
  1. Welcome them warmly but briefly
  2. Ask: "Who's your target audience for this content?"
  3. Ask: "What platform is this for?"
  4. Say: "Paste your hook and I'll run it through The Hook Stack."
  5. Evaluate using the scorecard format
  6. Ask: "Keep / Tweak / Trash?"
  7. Based on their choice:
    • Keep → Celebrate genuinely, ask if they want another evaluated
    • Tweak → "Which layer do you want to work on?" Then offer 3-5 improved versions
    • Trash → "Smart call. Want me to build some fresh options?" Then generate 3-5 new hooks

Continue iterating until they're happy. Keep the energy up. Make it feel like a sparring session, not a grading exercise.

For Agent Use (automated pipeline):

When invoked by an agent (like newsletter-writer):

  1. Receive audience context - The calling agent should specify the target audience (e.g., "Target audience: coaches/consultants building with AI")
  2. Provide the scorecard with scores
  3. Layer 4 (Speak Their Lingo) - Evaluate specifically against the provided audience:
    • Are examples/metaphors in their world?
    • Would they use this language at 2am?
    • Does it sound like insider knowledge for this group?
  4. If score < 12/15: Suggest 3 improved alternatives using audience-appropriate language
  5. If score >= 12/15: Confirm hook is ready
  6. Return the best hook in the final format

Final Hook Delivery

When the user confirms they are happy (e.g., "Lock it in", "Let's go with that"), output:

THE FINAL HOOK

"[The exact final hook text]"

For headline + subhead pairs (newsletter format):

THE FINAL HOOK

Headline: "[The headline]"
Subhead: "[The subhead]"

Hook Template Patterns

Use these patterns to inspire alternatives when tweaking or building fresh hooks:

  • "This may be controversial but ___"
  • "___ type of people... stop scrolling!"
  • "Everything you knew about ___ is 100% WRONG"
  • "I bet I can change your mind about ___ in 20 seconds"
  • "Why is nobody talking about ___ ?"
  • "This is the mistake every ___ makes..."
  • "If you're feeling skeptical about ___... try this"
  • "I spent [time] doing ___ so you don't have to"
  • "The best $[amount] I ever spent on ___"
  • "3 things you're NOT doing to ___"
  • "This is the lazy way to get [result]"
  • "You will look at ___ differently after the next 30 seconds"
  • "How to get [result] in [time] or less"
  • "My biggest regret was not discovering this sooner"
  • "Recording [person]'s reaction to ___"
  • "Seeing if I can ___ in [time]..."
  • "Instead of ___ like everyone else, I do this instead"
  • "Here are 2 signs you should stop doing ___"
  • "This free tool changed the game for me"
  • "This is for the [type of person] looking for [result]"

Platform Considerations

  • Video hooks: 3-5 seconds when spoken. Never start with "Hey! My name is..."
  • Newsletter headlines: Headline + subhead pair. Subhead expands the promise or adds intrigue.
  • Social posts: Caption is prime hook real estate - never waste on fluff.
  • Carousels: First slide IS the hook. Everything depends on that stop.

Guidelines

  • Push them to be better. Celebrate the wins. Keep it moving.
  • Don't be precious about hooks - they can always be improved
  • The best hooks often come from the third or fourth iteration
  • When stuck, go back to Layer 5 - personal stories unlock everything

Version History

VersionDateChanges
1.1.02026-01-05Added Mode Detection (Automatic vs Interactive) - fixes over-asking in pipelines
1.0.02026-01-01Initial creation from Hook-Stack-Evaluator app

© nicepkg, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files in workflows/content-creator-workflow/.claude/skills/hook-stack-evaluator of nicepkg/ai-workflow.

  • SKILL.md
  • docs/GUIDE.md
  • docs/ROADMAP.md

Open the folder on GitHubat commit d167b41

Compare with similar skills

Hook Stack Evaluator next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Hook Stack Evaluator compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Hook Stack Evaluator this skillnicepkg/ai-workflow285—~2.6kAutomated safety check: PassMIT
Internal Communications Writeranthropics/skills180k38 repos~378Automated safety check: PassApache-2.0
Clarityaddyosmani/clarity2661 repos~1.9kAutomated safety check: PassMIT
News Aggregator Skillcclank/news-aggregator-skill1.3k—~2.1kAutomated safety check: PassNone
Changelog Social RecapFlorianBruniaux/claude-code-ultimate-guide6.1k—~1.8kAutomated safety check: NotesCC-BY-SA-4.0
Newsletter Voicecharlie947/social-media-skills3.8k—~2.5kAutomated safety check: PassMIT

Similar skills

  • Official

    A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use. Claude should use this skill…

    180k GitHub starsUsed in 38 repos~378 tokens
    Writing & ContentAuto-check passed
  • Clarity

    addyosmani/clarity

    Draft, rewrite, or review reader-facing prose so it is specific, useful, and recognizably the author's without inventing facts or performing humanness.

    266 GitHub starsUsed in 1 repo~1.9k tokens
    Writing & ContentAuto-check passed
  • News Aggregator Skill

    cclank/news-aggregator-skill

    Comprehensive news aggregator that fetches, filters, and deeply analyzes real-time content from 44+ sources including Hacker News, Lobsters, Dev.to, GitHub, arXiv, Hugging Face Papers, AIHOT, TLDR…

    1.3k GitHub stars~2.1k tokensUpdated 4 mo ago
    Writing & ContentAuto-check passed
  • Changelog Social Recap

    FlorianBruniaux/claude-code-ultimate-guide

    Turns CHANGELOG.md entries for a release or a week into LinkedIn, Twitter/X, newsletter and Slack posts in French and English.

    6.1k GitHub stars~1.8k tokensUpdated today
    Writing & ContentAuto-check: notes
  • Newsletter Voice

    charlie947/social-media-skills

    Build newsletter writing instructions inside a Codex or Claude project.

    3.8k GitHub stars~2.5k tokensUpdated 23 days ago
    Writing & ContentAuto-check passed
  • News Extractor

    NanmiCoder/NewsCrawler

    新闻站点内容提取。支持 12 个平台:微信公众号、今日头条、网易新闻、搜狐新闻、腾讯新闻、BBC News、CNN News、Twitter/X、Lenny's Newsletter、Naver Blog、Detik News、Quora。当用户需要提取新闻内容、抓取公众号文章、爬取新闻、或获取新闻JSON/Markdown时激活。

    622 GitHub stars~1.3k tokensUpdated 2 mo ago
    Writing & ContentAuto-check passed

More from nicepkg/ai-workflow

All 61 skills in this repo
  • Capture Triage

    nicepkg/ai-workflow

    Processes Drafts Pro captures from the Inbox folder. An agent skill from nicepkg/ai-workflow.

    285 GitHub stars~3k tokensUpdated 8 mo ago
    Auto-check passed
  • Legacy To AI Ready

    nicepkg/ai-workflow

    Transform legacy codebases into AI-ready projects with Claude Code configurations.

    285 GitHub stars~2.2k tokensUpdated 8 mo ago
    Auto-check: notes
  • Newsletter Coach

    nicepkg/ai-workflow

    Writing coach that extracts educational content from your daily experiences and turns it into publish-ready newsletter drafts.

    285 GitHub stars~3.7k tokensUpdated 8 mo ago
    Auto-check passed
  • Webfluence

    nicepkg/ai-workflow

    Content web architecture framework. An agent skill from nicepkg/ai-workflow.

    285 GitHub stars~1.4k tokensUpdated 8 mo ago
    Auto-check passed
  • Workflow Creator

    nicepkg/ai-workflow

    Create complete Claude Code workflow directories with curated skills.

    285 GitHub stars~2.6k tokensUpdated 8 mo ago
    Auto-check passed
  • Media Processing

    nicepkg/ai-workflow

    Video/audio/image processing with FFmpeg and ImageMagick. An agent skill from nicepkg/ai-workflow.

    285 GitHub stars~2.4k tokensUpdated 8 mo ago
    Auto-check: notes

Questions about Hook Stack Evaluator

What does Hook Stack Evaluator do?

Evaluates and sharpens content hooks using The Hook Stack™ framework. Hook Stack Evaluator is an agent skill from nicepkg/ai-workflow. Evaluates and sharpens content hooks using The Hook Stack™ framework.

When should I use Hook Stack Evaluator?

Hook Stack Evaluator fits situations like: scoring headlines; refining hooks for video/social/newsletter; asked to evaluate this hook; run through hook stack.

How do I install Hook Stack Evaluator in Claude Code?

Run `npx skills add nicepkg/ai-workflow --skill hook-stack-evaluator -a claude-code`. Or copy the skill folder (workflows/content-creator-workflow/.claude/skills/hook-stack-evaluator in nicepkg/ai-workflow) into .claude/skills/hook-stack-evaluator in your project. Claude Code loads it when a task matches its description.

How do I install Hook Stack Evaluator in Codex?

Run `npx skills add nicepkg/ai-workflow --skill hook-stack-evaluator -a codex`. Or copy the skill folder (workflows/content-creator-workflow/.claude/skills/hook-stack-evaluator in nicepkg/ai-workflow) into .agents/skills/hook-stack-evaluator in your project. Codex loads it when a task matches its description.

Can I use Hook Stack Evaluator in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add nicepkg/ai-workflow --skill hook-stack-evaluator -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/hook-stack-evaluator, .gemini/skills/hook-stack-evaluator, .github/skills/hook-stack-evaluator and .opencode/skills/hook-stack-evaluator in your project.

What does Hook Stack Evaluator need to run?

SKILL.md names no scripts, command-line tools or credentials: Hook Stack Evaluator is instructions for the agent only. Its frontmatter pre-approves these tools: Read, Write, Edit.

Does Hook Stack Evaluator access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Hook Stack Evaluator safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Hook Stack Evaluator use?

Hook Stack Evaluator is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Hook Stack Evaluator use?

About 2.6k tokens (SKILL.md is roughly 11k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Hook Stack Evaluator?

Skills that share tags, products or a category with Hook Stack Evaluator: Internal Communications Writer (anthropics/skills, 180k stars), Clarity (addyosmani/clarity, 266 stars), News Aggregator Skill (cclank/news-aggregator-skill, 1.3k stars) and Changelog Social Recap (FlorianBruniaux/claude-code-ultimate-guide, 6.1k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Hook Stack Evaluator?

nicepkg (a GitHub organization) maintains it in nicepkg/ai-workflow, which has 285 GitHub stars. The repository holds 61 skills in this directory. The repository was last updated on January 20, 2026.

Source: nicepkg/ai-workflow on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.