Reverse-engineer the hook from a viral X tweet or thread URL

X hook extractor

by sergebulaev · MIT license · ★ 95 on GitHub · updated Sep 17

Reverse-engineer the hook from a viral X (Twitter) tweet or thread URL. Identifies which of the 10 canonical 2026 X formulas it uses (one-liner contrarian, data-point, build-in-public, quote-tweet, mini-list, relatable cold-open,…

---
name: x-hook-extractor
description: "Reverse-engineer the hook from a viral X (Twitter) tweet or thread URL. Identifies which of the 10 canonical 2026 X formulas it uses (one-liner contrarian, data-point, build-in-public, quote-tweet, mini-list, relatable cold-open, listicle-thread, story thread, curiosity-gap, how-I teardown), explains why it worked, and returns a blank template mapped to your topic with its primary goal. Use to learn from a tweet you admire. Not for writing your own (use x-post-writer or x-thread-builder)."
---

# X Hook Extractor

Paste a viral tweet or thread URL. Get back: which hook formula it uses, the
exact structure, why it worked, and a blank template you can fill with your own
voice.

## When to use

- User finds a viral tweet or thread they want to study
- User wants to replicate a specific creator's pattern
- Before `x-post-writer` or `x-thread-builder`, to seed a draft with a proven shape

## Input

Overview

Paste a viral tweet or thread URL. Get back: which hook formula it uses, the exact structure, why it worked, and a blank template you can fill with your own voice.

What you get

Formula identified(X1-X10 from ../../references/hook-formulas.md) with a
Containersingle tweet vs thread, and why that container fit the idea
Structural breakdown
The hook line (and for a thread, how tweet 1 opens the loop)
Body architecture (per-tweet roles for a thread)
The close (what earns the repost or bookmark)

Manual install (npx, ZIP)

Claude Codenpx skills add sergebulaev/x-skills --skill x-hook-extractor -g -a claude-code

Claude Code (manual)

mkdir -p ~/.claude/skills && curl -sL https://codeload.github.com/sergebulaev/x-skills/tar.gz/061109fdcb0b \
  | tar -xz -C ~/.claude/skills --strip-components=4 x-skills-061109fdcb0b/.codex-marketplace/x-skills/skills/x-hook-extractor

Windows (PowerShell)

New-Item -ItemType Directory -Force "$HOME\.claude\skills" | Out-Null; Invoke-WebRequest "https://m.agentalley.io/zip/33cb612ed98d90e41f2f/x-hook-extractor.zip" -OutFile "$env:TEMP\x-hook-extractor.zip"; Expand-Archive "$env:TEMP\x-hook-extractor.zip" -DestinationPath "$HOME\.claude\skills" -Force

Claude.ai / desktopDownload the ZIP and upload it under Settings → Capabilities → Skills.

Codex, Cursor, other agentsPoint them at SKILL.md. The other files are plain text.

Connects with Agent Alley MCP · agentalley.io/mcp

Files included

SKILL.md3.8 KB
references/2 files · 6.2 KB

Read the files here

sergebulaev/061109f3 files

SKILL.md3.8 KB

Shared references1SKILL.md reads these from elsewhere in the repo. Not in the ZIP.

hook-formulas.md.codex-marketplace/x-skills/references/15.4 KB

Get them on GitHub ↗

SKILL.md

Show the full text84 lines

x-hook-extractor/SKILL.md84 lines · 3.8 KB

PreviewCode

Outline

RawView on GitHub

X Hook Extractor

Paste a viral tweet or thread URL. Get back: which hook formula it uses, the exact structure, why it worked, and a blank template you can fill with your own voice.

When to use

User finds a viral tweet or thread they want to study
User wants to replicate a specific creator's pattern
Before x-post-writer or x-thread-builder, to seed a draft with a proven shape

Input

An X tweet or thread URL (x.com or twitter.com, /status/<id>). For a thread, the URL of the first tweet is best.

Output

Formula identified (X1-X10 from ../../references/hook-formulas.md) with a confidence score
Container: single tweet vs thread, and why that container fit the idea
Structural breakdown:
- The hook line (and for a thread, how tweet 1 opens the loop)
- Body architecture (per-tweet roles for a thread)
- The close (what earns the repost or bookmark)
- Reaction-triggering devices (numbers, named entities, the open loop)
Primary goal the original chased (replies / reposts / likes / bookmarks)
Why it worked psychologically and algorithmically
Blank template with {slot} markers matched to the original, ready for the user's topic
Cautions: anything in the original that would fail a 2026 audit (more than one em dash in a tweet, an AI-vocab cluster, 3+ hashtags, link in tweet 1)

Steps

Parse the URL. lib.url_parser.parse_x_url(url) returns handle, tweet_id, url_type.
Get the text. This bundle has no built-in tweet reader, so ask the user to paste the tweet or the full thread text. (If they later wire an Apify tweet actor, read it automatically.)
Detect the container. One self-contained tweet, or a multi-tweet thread.
Classify against the 11 formulas using features:
- Single tweet: a flat contrarian claim (X1)? one hard number (X2)? a personal metric/confession (X3)? a quote tweet adding a layer (X4)? a one-line-per- item list (X5)? a relatable shared moment (X6)?
- Thread: a numbered teaching promise (X7)? a story starting at the tension (X8)? a surprising result with the mechanism withheld (X9)? a first-person "how I" teardown (X10)?
Score confidence. If two formulas fit, return the top 2 with fit scores.
Extract structure. Label each part by its role. For a thread, map tweet 1 (the loop), the front-loaded payoff, the body beats, and the closer.
Name the primary goal the original optimized for.
Generate a blank template with {slot} markers matched to the original shape and the user's topic.
Audit the source. Flag any AI tells in the original so the user does not copy them.

Example

See references/examples.md for worked teardowns.

Formulas reference

See ../../references/hook-formulas.md for the 11 canonical X formulas with full skeletons and goal tags.

Files

SKILL.md - this file
references/classification-rules.md - feature extraction + scoring heuristics
references/examples.md - worked teardowns (single tweet and thread)

x-post-writer - use the extracted single-tweet template to draft your own
x-thread-builder - use the extracted thread template
x-humanizer --mode audit - audit your draft before shipping

1	---
2	name: x-hook-extractor
3	description: "Reverse-engineer the hook from a viral X (Twitter) tweet or thread URL. Identifies which of the 10 canonical 2026 X formulas it uses (one-liner contrarian, data-point, build-in-public, quote-tweet, mini-list, relatable cold-open, listicle-thread, story thread, curiosity-gap, how-I teardown), explains why it worked, and returns a blank template mapped to your topic with its primary goal. Use to learn from a tweet you admire. Not for writing your own (use x-post-writer or x-thread-builder)."
4	---
5
6	# X Hook Extractor
7
8	Paste a viral tweet or thread URL. Get back: which hook formula it uses, the
9	exact structure, why it worked, and a blank template you can fill with your own
10	voice.
11
12	## When to use
13
14	- User finds a viral tweet or thread they want to study
15	- User wants to replicate a specific creator's pattern
16	- Before `x-post-writer` or `x-thread-builder`, to seed a draft with a proven shape
17
18	## Input
19
20	An X tweet or thread URL (x.com or twitter.com, `/status/<id>`). For a thread,
21	the URL of the first tweet is best.
22
23	## Output
24
25	- Formula identified (X1-X10 from `../../references/hook-formulas.md`) with a
26	confidence score
27	- Container: single tweet vs thread, and why that container fit the idea
28	- Structural breakdown:
29	- The hook line (and for a thread, how tweet 1 opens the loop)
30	- Body architecture (per-tweet roles for a thread)
31	- The close (what earns the repost or bookmark)
32	- Reaction-triggering devices (numbers, named entities, the open loop)
33	- Primary goal the original chased (replies / reposts / likes / bookmarks)
34	- Why it worked psychologically and algorithmically
35	- Blank template with `{slot}` markers matched to the original, ready for the
36	user's topic
37	- Cautions: anything in the original that would fail a 2026 audit (more
38	than one em dash in a tweet, an AI-vocab cluster, 3+ hashtags, link in tweet 1)
39
40	## Steps
41
42	1. Parse the URL. `lib.url_parser.parse_x_url(url)` returns `handle`,
43	`tweet_id`, `url_type`.
44	2. Get the text. This bundle has no built-in tweet reader, so ask the user to
45	paste the tweet or the full thread text. (If they later wire an Apify tweet
46	actor, read it automatically.)
47	3. Detect the container. One self-contained tweet, or a multi-tweet thread.
48	4. Classify against the 11 formulas using features:
49	- Single tweet: a flat contrarian claim (X1)? one hard number (X2)? a personal
50	metric/confession (X3)? a quote tweet adding a layer (X4)? a one-line-per-
51	item list (X5)? a relatable shared moment (X6)?
52	- Thread: a numbered teaching promise (X7)? a story starting at the tension
53	(X8)? a surprising result with the mechanism withheld (X9)? a first-person
54	"how I" teardown (X10)?
55	5. Score confidence. If two formulas fit, return the top 2 with fit scores.
56	6. Extract structure. Label each part by its role. For a thread, map tweet 1
57	(the loop), the front-loaded payoff, the body beats, and the closer.
58	7. Name the primary goal the original optimized for.
59	8. Generate a blank template with `{slot}` markers matched to the original
60	shape and the user's topic.
61	9. Audit the source. Flag any AI tells in the original so the user does not
62	copy them.
63
64	## Example
65
66	See `references/examples.md` for worked teardowns.
67
68	## Formulas reference
69
70	See `../../references/hook-formulas.md` for the 11 canonical X formulas with full
71	skeletons and goal tags.
72
73	## Files
74
75	- `SKILL.md` - this file
76	- `references/classification-rules.md` - feature extraction + scoring heuristics
77	- `references/examples.md` - worked teardowns (single tweet and thread)
78
79	## Related skills
80
81	- `x-post-writer` - use the extracted single-tweet template to draft your own
82	- `x-thread-builder` - use the extracted thread template
83	- `x-humanizer --mode audit` - audit your draft before shipping
84