Skip to content
Back to search

x-agent-skill

Eleven free Claude skills that run an X account, with a virality engine behind every idea: posts off 21 hook formulas, threads, replies, profile score, weekly plan, analytics audit and a humanizer. Free, MIT.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/jakeschincariol--x-agent-skill/badge.svg)](/listings/jakeschincariol--x-agent-skill)

HTML

<a href="/listings/jakeschincariol--x-agent-skill"><img src="/api/listings/jakeschincariol--x-agent-skill/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Paste this into Claude:

https://github.com/Jakeschincariol/x-agent-skill

Install this skill, then confirm /x-post works.

Or do it yourself, in Claude Code:

git clone https://github.com/Jakeschincariol/x-agent-skill.git
cp -r x-agent-skill/skills/x-* ~/.claude/skills/

Or as a plugin:

/plugin marketplace add Jakeschincariol/x-agent-skill
/plugin install x-agent

Project-local instead of global: copy the same folders into your repo's `.claude/skills/`. No Claude Code at all? Paste any single `SKILL.md` at the top of a chat and it runs as a mode. You lose the Python tools, which is most of the point of `/x-post`, `/x-viral` and `/x-human`, but the rest works.

Then spend ten minutes on [`templates/voice.md`](templates/voice.md). Copy it to `~/.claude/x/voice.md` and fill it in, or paste five of your own posts into Claude and say "write my voice.md from these". Every skill reads that file. Skip it and everything comes out sounding like everyone else on the timeline.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 15 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
  3. "name": "x-agent-skill",
  4. "description": "The X agent skill pack for Claude",
  5. "owner": { "name": "Jake Schincariol", "url": "https://opusjake.ai" },
  6. "plugins": [
  7. {
  8. "name": "x-agent",
  9. "description": "Eleven Claude skills that run an X account, including a virality engine, a hook scorer, X's weighted character counter, a thread builder and a humanizer.",
  10. "source": "./",
  11. "category": "productivity",
  12. "homepage": "https://github.com/Jakeschincariol/x-agent-skill"
  13. }
  14. ]
  15. }
README.md · README · 312 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # The X agent skill
  2. Eleven Claude skills that run an X account. Free, MIT, no signup, no API key,
  3. nothing to connect.
  4. They write your posts, comment on other people's threads, handle your replies,
  5. fix your profile and plan your week. And one of them is the virality skill: it
  6. finds the posts that are actually outperforming in your niche, breaks each one
  7. down into hook, structure and format, and recreates the winners in your voice,
  8. for your audience.
  9. Under the skills are nine tools that actually run. One counts a post the way X
  10. counts it (links are 23, emoji are 2). One scores your first line before you
  11. post it. One cuts a draft into a thread that fits. One reads X's own published
  12. ranking code and tells you what it rewards this week. One reads your analytics
  13. export and tells you what worked, with the sample size on every claim.
  14. **Nothing is posted until you approve it.** These skills write. You post.
  15. ## Install
  16. Paste this into Claude:
  17. ```
  18. https://github.com/Jakeschincariol/x-agent-skill
  19. Install this skill, then confirm /x-post works.
  20. ```
  21. Or do it yourself, in Claude Code:
  22. ```bash
  23. git clone https://github.com/Jakeschincariol/x-agent-skill.git
  24. cp -r x-agent-skill/skills/x-* ~/.claude/skills/
  25. ```
  26. Or as a plugin:
  27. ```
  28. /plugin marketplace add Jakeschincariol/x-agent-skill
  29. /plugin install x-agent
  30. ```
  31. Project-local instead of global: copy the same folders into your repo's
  32. `.claude/skills/`. No Claude Code at all? Paste any single `SKILL.md` at the
  33. top of a chat and it runs as a mode. You lose the Python tools, which is most
  34. of the point of `/x-post`, `/x-viral` and `/x-human`, but the rest works.
  35. Then spend ten minutes on [`templates/voice.md`](templates/voice.md). Copy it
  36. to `~/.claude/x/voice.md` and fill it in, or paste five of your own posts into
  37. Claude and say "write my voice.md from these". Every skill reads that file.
  38. Skip it and everything comes out sounding like everyone else on the timeline.
  39. ## The eleven
  40. | command | | what it does |
  41. | --- | --- | --- |
  42. | `/x-post` | Post Writer | One idea into a post. Three hooks from [21 formulas](skills/x-post/hooks.json), scored, one draft, counted the way X counts, humanized. |
  43. | `/x-comment` | Comment Drafter | Replies on other people's posts and threads. Nine types, picked by what the post is. Reply or quote, decided first. |
  44. | `/x-profile` | Profile Optimizer | Scores your profile against a [12-part rubric](skills/x-profile/rubric.json) out of 100, then rewrites bio, pinned post and header in fix-first order. |
  45. | `/x-reply` | Reply Handler | The replies under your own posts, sorted into lead / substance / peer / question / support / noise, then answered in that order. |
  46. | `/x-plan` | Content Planner | The week: what to post, which format, when, and the 10 accounts to engage with. |
  47. | `/x-viral` | Virality Engine | Finds what is outperforming in your niche, ranks it by multiple over each account's own median, breaks down hook, structure and format, recreates the top three for you. |
  48. | `/x-thread` | Thread Builder | Thread, long post or single post, decided first. Then post 1, one idea per post, cut to fit, and a last post people act on. |
  49. | `/x-human` | Humanizer | Strips em dashes, slop and invisible characters, then scores the draft on a five-check panel. |
  50. | `/x-repurpose` | Repurposer | One video, podcast, newsletter or LinkedIn post into a week of X posts and a thread. |
  51. | `/x-dm` | DM Handler | The first message, two follow-ups, and triage for your inbox and message requests. |
  52. | `/x-audit` | Account Audit | Reads your analytics export. Ranks by engagement, replies and shares per impression, not by impressions. |
  53. ## The nine tools
  54. Every one runs on a clean Python 3 with no dependencies. Seven never touch the
  55. network. `pull.py` calls X's documented oEmbed API and `weights.py` reads a
  56. public GitHub file; neither needs an account or a key.
  57. ```bash
  58. python3 skills/x-post/count.py post.txt # X's weighted count + six lints
  59. python3 skills/x-post/hookscore.py hooks.txt # score and rank first lines, name the formula
  60. python3 skills/x-thread/thread.py draft.txt --number # cut a draft into a thread that fits
  61. python3 skills/x-human/humanize.py draft.txt --report # strip the fingerprint
  62. python3 skills/x-human/detect.py before.txt after.txt # five-check panel, before and after
  63. python3 skills/x-viral/pull.py urls.txt > swipe.tsv # post URLs you picked -> text, format, date
  64. python3 skills/x-viral/swipe.py swipe.tsv # outliers, broken down, as a swipe file
  65. python3 skills/x-audit/audit.py analytics.csv # what worked, with n on every claim
  66. python3 skills/x-plan/weights.py # what X's ranker rewards, live from X's code
  67. ```
  68. ### Counting the way X counts
  69. X does not count characters. It counts weighted characters: every link is 23
  70. however long it is, every emoji is 2 (a family emoji built from seven code
  71. points is still 2), Chinese, Japanese and Korean are 2, and "…" is 2, which is
  72. more than the three dots it replaces. `count.py` implements those rules, then
  73. draws the post the way the timeline shows it:
  74. ```
  75. WHAT FITS
  76. +------------------------------------------------------+
  77. | Most people use AI to post more. I used it to post |
  78. | less. |
  79. | |
  80. | In March I cut from 3 posts a day to 1. Followers |
  81. | went from 2,100 to 6,800 by August. |
  82. | |
  83. | The system: one idea a day, drafted with Claude, cut |
  84. | in half by me, and I'm there for the first hour of |
  85. | replies. |
  86. | |
  87. | Showing up beats post |
  88. +------------------------------------ cut here, 4 over +
  89. FAIL LENGTH 284 / 280, 4 over. X will not let you post it.
  90. PASS FIRST LINE 56 characters, reads in one glance
  91. PASS LINKS none
  92. PASS HASHTAGS 0
  93. PASS MENTIONS 0
  94. PASS ENDING no reflex bait
  95. ```
  96. With `--premium` the limit is 25,000 and the box shows where the timeline cuts
  97. a long post to "Show more", at about 280. It also flags an @mention past that
  98. point, because X does not notify it.
  99. ### The hook
  100. On X the first line is not the start of the post. It is most of the post.
  101. `hookscore.py` scores it on five properties (length, specificity, stakes,
  102. whether the payoff is up front, who it is aimed at), flags the dealbreakers,
  103. and names which of the 21 formulas it uses:
  104. ```
  105. HOOK RANKING
  106. ==============================================================================
  107. -> 100.0 STRONG Nobody tells you your first 1,000 followers come from rep...
  108. #6 Nobody Tells You · weakest: LENGTH (100)
  109. 78.4 STRONG I posted every day for 90 days. Here's what actually move...
  110. #2 Number Reveal · weakest: FRONTLOAD (70)
  111. 52.2 OK The hottest new marketing hire is a skill file.
  112. #7 The Hottest New · weakest: SPECIFICITY (15)
  113. 36.2 WEAK Hey guys, here are some thoughts on growing on X
  114. #21 The One-Liner · weakest: STAKES (20)
  115. dealbreaker: Greeting. Nobody opened X to be greeted.
  116. 4.8 WEAK A thread 🧵
  117. #16 Thread Promise · weakest: FRONTLOAD (0)
  118. dealbreaker: Opens on the word "thread". That is packaging. Put the
  119. payoff in line 1 and let X's own UI say it is a thread.
  120. ```
  121. Every formula in [`hooks.json`](skills/x-post/hooks.json) carries a template,
  122. a filled example, what it is for, how it usually gets ruined, and a regex, so
  123. the same 21 names describe a hook you are writing and a hook you found in
  124. someone else's feed. That is what lets `/x-viral` say "the top third of your
  125. niche is The Hottest New and Field Report" instead of "posts that did well".
  126. ### The virality engine
  127. ```bash
  128. python3 skills/x-viral/pull.py urls.txt > swipe.tsv
  129. # fill in the views column from the posts, they are public on every post
  130. python3 skills/x-viral/swipe.py swipe.tsv --out ~/.claude/x/swipe.md
  131. ```
  132. Raw numbers are not evidence. An account with 900,000 followers doing 400,000
  133. views had an ordinary Tuesday; an account with 6,000 doing the same found
  134. something copyable. So `swipe.py` ranks on the **outlier multiple**, each
  135. post's views over its own account's median, computed from the 10 to 15 posts
  136. you collected per account. Accounts with too few posts are skipped instead of
  137. ranked on noise. Then each post gets broken down:
  138. - **hook**: which of the 21 formulas, and its score
  139. - **structure**: one line, one paragraph, short lines, a list, a long post, a thread
  140. - **format**: text, media, link, long
  141. It compares the top third with the bottom third on all three, lists the hooks
  142. it could not classify (read those, they are where a new formula is hiding) and
  143. writes a swipe file with the reusable shape of each winner. `/x-viral` then
  144. recreates the top three in your voice: the found post's structure, none of its
  145. words.
  146. ### What X rewards, from X
  147. X open-sourced its current recommendation system in January 2026 and has
  148. published the production ranking weights in it since August. They move: two
  149. of them changed in the week this pack was written. So `weights.py` reads the
  150. file live and prints it with the date of its last change:
  151. ```
  152. X FOR YOU RANKER · action weights · live from GitHub, last changed 2026-09-29 (commit a707cc2)
  153. ==============================================================================
  154. REWARDED weight vs a like
  155. share by copying the link 20 40.0x
  156. + added to a reply between mutuals 15 30.0x
  157. reply 5 10.0x
  158. share by DM 5 10.0x
  159. quote 5 10.0x
  160. follow the author 4 8.0x
  161. share (any) 2 4.0x
  162. repost 1 2.0x
  163. like 0.5 1.0x
  164. ...
  165. PUNISHED
  166. report -234
  167. mute the author -58.8
  168. "not interested" -47.52
  169. block the author -31.2
  170. ```
  171. Read it the way X's own comment in that file says to: each weight multiplies
  172. the *predicted probability* of an action for a given viewer, not a count of
  173. actions, so "a reply is worth ten likes" is the wrong conclusion. The right
  174. one is the direction. Write things people reply to, quote and send to one
  175. friend. Do not write the things people mute.
  176. ### The humanizer
  177. `humanize.py` strips invisible characters, fixes typography (em dash to comma,
  178. curly to straight) and replaces 176 stock words and phrases from
  179. [`slop.json`](skills/x-human/slop.json), including an X block: "if you found
  180. this thread helpful", "RT if you agree", "let's dive in". It keeps the
  181. invisible characters that emoji need, so 👨‍💻 stays one emoji. Structural tells
  182. ("A thread 🧵" as the hook, "it's not just X, it's Y", a pile of hashtags) are
  183. flagged, not rewritten, because changing the shape of a sentence needs
  184. judgement.
  185. `detect.py` scores five checks: burstiness, specificity, slop density,
  186. fingerprint, voice. A post of one or two sentences cannot show burstiness or a
  187. specificity density, so those sit out as n/a and the output says it scored on
  188. 3 of 5 rather than pretending.
  189. A deliberately bad post, the automatic passes, then a rewrite by hand:
  190. ```
  191. HUMAN SCORE #####................... 19.6 FLAGGED the original
  192. HUMAN SCORE ##########.............. 42.9 FLAGGED after humanize.py alone
  193. HUMAN SCORE #####################... 85.7 PASS after rewriting the flagged lines
  194. ```
  195. The script gets you the first 23 points. The last 43 are the part it
  196. deliberately leaves to you.
  197. ## The fine print, which is the honest part
  198. **These skills do not post, reply, like, follow or DM.** X's API does allow
  199. posting with your own developer app and OAuth, unlike LinkedIn's. As of
  200. September 2026 it is pay-per-use: $0.015 a post, $0.20 for a post with a link,
  201. credits bought upfront ([pricing](https://docs.x.com/x-api/getting-started/pricing)).
  202. It is still not built here, on purpose. X's
  203. [automation rules](https://help.x.com/en/rules-and-policies/x-automation)
  204. prohibit automated likes, bulk follows and unsolicited automated replies, and
  205. AI reply bots need X's written approval, and replies and comments are most of
  206. what this pack writes. So every skill ends the same way: a copy-ready block,
  207. and you paste it.
  208. **`/x-viral` reads, it does not scrape.** X's Terms prohibit crawling or
  209. scraping "in any form, for any purpose" without written consent. You pick the
  210. posts; `pull.py` only fetches URLs you give it, through the documented oEmbed
  211. API, one per second, capped at 100 a run. It never logs in, never uses your
  212. session and never fetches counts. You read those off the screen.
  213. **`hookscore.py` is a lint, not a predictor.** The same five-property panel,
  214. tested on 74 real short-form hooks in the Instagram version of this pack,
  215. separated real hooks from deliberately bad ones well (AUC 0.83) and a
  216. creator's own hits from their own misses barely at all (AUC 0.56, a coin flip
  217. is 0.50). It has not been re-calibrated on X. It catches labels, greetings and
  218. hooks with nothing concrete in them. It will not tell you which of two decent
  219. hooks will travel. `/x-audit` on your own posts will.
  220. **The link question is left to your data.** X says it stopped down-ranking
  221. posts with links more than a year ago (Musk and Nikita Bier, July 2026). An
  222. independent analysis in January 2026 still measured fewer views on them. The
  223. pack does not pick a side. `/x-audit` splits your posts by format and tells
  224. you which is true for your account.
  225. **The five detection checks are local heuristics**, not GPTZero, Originality,
  226. Copyleaks, Winston or Turnitin. They do not call those services and cannot
  227. promise their verdicts. Nobody can honestly sell you "undetectable".
  228. **Nothing here fabricates.** No invented metrics, clients or outcomes go under
  229. your name. If a draft needs a number you have not given, it comes back with
  230. `{{your number}}` in it and a flag, every time.
  231. ## Files
  232. ```
  233. skills/x-post/hooks.json 21 hook formulas: template, example, use, trap, match regex
  234. skills/x-post/hookscore.py five-property hook panel + formula classifier
  235. skills/x-post/count.py X's weighted character count + six lints
  236. skills/x-thread/thread.py draft -> thread that fits, five thread lints
  237. skills/x-viral/pull.py post URLs -> swipe sheet, via X's oEmbed API
  238. skills/x-viral/swipe.py outlier ranking + hook / structure / format breakdown
  239. skills/x-audit/audit.py analytics export -> what worked, with n
  240. skills/x-plan/weights.py X's published ranking weights, live
  241. skills/x-human/slop.json 176 terms, 18 invisible classes, 18 structural tells
  242. skills/x-human/humanize.py the three cleaning passes
  243. skills/x-human/detect.py the five-check panel
  244. skills/x-profile/rubric.json the 100-point profile score
  245. templates/voice.md your voice profile. Fill this in first.
  246. ```
  247. Your own files live in `~/.claude/x/`: `voice.md`, `plan.md`, `log.md` and
  248. `swipe.md`. The skills read each other's.
  249. ## Credit
  250. Made by Jake Schincariol, [opusjake.ai](https://opusjake.ai). Siblings:
  251. [LinkedIn](https://github.com/Jakeschincariol/linkedin-agent-skill),
  252. [Instagram](https://github.com/Jakeschincariol/instagram-agent-skill),
  253. [YouTube](https://github.com/Jakeschincariol/youtube-agent-skill).
  254. ## License
  255. MIT. Take it, change it, ship it.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.