HookCheck - Policy Risk Check for Video Hooks

What this does

Most hook advice tells you to promise more. The platforms' own rules say the opposite: if the video does not deliver what the title and thumbnail promised, that is the exact thing they act on.

Paste your draft before you post. You get back:

1. A verdict — High risk, Fix before posting, Polish, or No findings in what was supplied.
2. Policy findings — the exact words you wrote, what the problem is, the platform's rule quoted with its source and the date it was checked, the specific outcome that rule states, and a fix.
3. Craft findings — where the opening is weak rather than risky.
4. Rewrites — two or three that keep the curiosity and stay true to your video.
5. What it could not check, and what to paste to unlock it.
6. What does not apply to your platform, and why.

What it checks against

• YouTube — the Spam Policy clause "Malicious clickbait", the Thumbnails policy, the advertiser-friendly guidelines (which apply to "all portions of your content ... thumbnail, title, description, and tags"), and the Partner Program review.
• TikTok — Community Guidelines, 2026 August version, effective 2026-09-24.
• Meta — the clickbait and engagement-bait distribution guidelines.
• LinkedIn — the Professional Community Policies.

What makes it different

• It judges; it does not generate. Hook generators give you more options. This tells you which words are a problem and why.
• Every policy finding is quoted, with its source and the date it was checked. Where a platform has published no rule, it says "unverified" instead of inventing one. TikTok, for example, has no published rule naming cover images and does not use the word "clickbait" — the skill says so rather than making something up.
• It names the actual consequence. Not "you'll get demonetized" — limited ad earnings, no ad earnings, reduced distribution, feed ineligibility, a strike, or a suspended channel are different things, and it tells you which one applies.
• It will tell you nothing is wrong. A clean draft gets a short, honest report. The skill is instructed not to manufacture findings.
• It checks only what applies to you. Post to TikTok only, and the YouTube rules are set aside rather than quoted at you.

What is in the package

• The skill (SKILL.md)
• A policy reference with every quote, its source URL, and the date checked — including the four places where a platform publishes no rule, so nothing gets invented
• A craft reference: concrete-scene openings, banned template phrases, bait patterns, and rewrite tables
• Three worked examples, including one written in Japanese that is almost clean
• README, license and disclaimer

Who it is for

• Creators whose monetization matters to them, especially around a Partner Program review
• Brand and company accounts that cannot use bait
• People posting in finance, health or education, where wording is examined harder
• Agencies and freelancers posting on a client's channel, where a strike is not theirs to risk

Limits

• It does not predict views or retention. No tool can do that from text.
• It never says a draft is safe, compliant or approved.
• It reads text, not video or images.
• The platforms state their own example lists are not complete. A clean result means no listed pattern was found, not that your video is approved.
• Rules change. Every report carries the date its rules were checked (2026-10-02).
• Not legal advice, and not affiliated with YouTube, Google, TikTok, Meta or LinkedIn.