Claude Skill

x-post

Publish a vetted post to an X (Twitter) account through the owner's real logged-in Chrome tab, rate-guarded and draft-first. Counts length the way X does, treating every link as 23 characters, and refuses over-limit text instead of silently truncating. Triggers: "/x-post", "tweet

LLM Mart · 0 points · 0 views 0 listing impressions 0 install-command copies
Download tonydzi-second-brain-starter-kit-skills_x-post-7bc2efa.zip · 2 KB

Install

skills CLI npx skills add https://github.com/tonydzi/second-brain-starter-kit/tree/master/skills/x-post
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install tonydzi-second-brain-starter-kit@llmmart
Git git clone https://github.com/tonydzi/second-brain-starter-kit.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole tonydzi/second-brain-starter-kit collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

/x-post — post to X from the owner's account (Chrome, draft-first)

Why. EN teasers (from /episode) kept piling up as drafts — there was no poster for X. Same safe pattern as /fb-post: a live logged-in tab, never headless.

0. Safety catch (mandatory)

python "$USERPROFILE/.claude/scripts/_shared/social_guard.py" check x --text "<final text>"

BLOCKED (exit 3) → STOP and report (the 6/day limit, or a duplicate). Do not work around it.

1. The text (the owner's voice, top model)

Take a ready draft (an /episode EN teaser or an intention-lane text) or write it with the top model. A single tweet is ≤280 characters — count BEFORE opening the browser; longer → a thread (each tweet ≤280, chained by reply) or offer the operator a shorter version.

2. Tier-2 gate (draft-first)

Show the final text (+ the thread split, if it is a thread) → wait for an explicit +. Nothing is published without it.

3. The browser (Claude-in-Chrome, a live tab)

Browser work stays local on this machine; never drag the window to the foreground elsewhere. ⛔ IP gate (2026-07-16): posting to X and other ban-sensitive social platforms happens ONLY from the hub HUB-1 (a stable IP). Do NOT post from another machine — send the task to the hub as text. Canon: the "IP-sensitive actions from the hub only" rule.

  1. list_connected_browsers → no extension → block and tell the operator (do not fall back to Playwright).
  2. Open the x.com tab. Verify the logged-in handle (avatar / profile menu) against the registry. A different account → STOP and ask. Not logged in → block (we never touch login/2FA on X — checkpoint risk).
  3. Composer: find "post composer / What's happening". Enter the text (form_input). Do not press Post until the gate in §2 is satisfied.
  4. After the + — press Post. For a thread: use the "+" button in the composer after the first tweet, or reply to your own tweet.
  5. A screenshot of the published post + the tweet URL (click the timestamp → address bar) is the proof.

4. Record it (AFTER a successful publication)

python "$USERPROFILE/.claude/scripts/_shared/social_guard.py" record x --text "<text>"

Report: the link + the screenshot + "x today N/6".

Stop switches

  • Any checkpoint / captcha / "unusual activity" from X → STOP immediately, report, zero retries.
  • Never republish identical text; never post from someone else's account.
  • Links in the text must be live and ours (link-safety).

Related

/fb-post (the pattern this follows) · /tg-post · /episode (tiers: EN teaser → X) · the gate scripts\_shared\social_guard.py · the registry 00-System\Channels-Registry.md.


Like this skill? It is part of second-brain-starter-kit: the second brain we built for ourselves and run every day at Palo Alto AI Research Lab. Install the whole set with npx skills add tonydzi/second-brain-starter-kit. Everything is open source and free, so take what you need.

Flagships worth a look on their own: secondop-panel (a second opinion from a panel of external models), claude-memory-tidy (stop your agent's memory from rotting), telegram-mcp-kit (your own Telegram over MCP in about 15 minutes).

Author: Anton Dziatkovskii, Palo Alto AI Research Lab. Telegram @tonydzi - WhatsApp +1 341 222 9178 - X @Tony_Stef_

Engineers: want to test-drive this setup? Message me. I hand out free starter seeds to engineers who test and report back, and custom skill requests are welcome.

Files (second-brain-starter-kit)
  • SKILL.md 4.1 KB
    ---
    name: x-post
    description: >-
      Publish a vetted post to an X (Twitter) account through the owner's real logged-in Chrome tab,
      rate-guarded and draft-first. Counts length the way X does, treating every link as 23
      characters, and refuses over-limit text instead of silently truncating. Triggers: "/x-post",
      "tweet this", "post to X".
    license: MIT
    ---
    
    # /x-post — post to X from the owner's account (Chrome, draft-first)
    
    **Why.** EN teasers (from /episode) kept piling up as drafts — there was no poster for X. Same safe pattern as /fb-post: a live logged-in tab, never headless.
    
    ## 0. Safety catch (mandatory)
    ```bash
    python "$USERPROFILE/.claude/scripts/_shared/social_guard.py" check x --text "<final text>"
    ```
    `BLOCKED` (exit 3) → STOP and report (the 6/day limit, or a duplicate). Do not work around it.
    
    ## 1. The text (the owner's voice, top model)
    Take a ready draft (an /episode EN teaser or an intention-lane text) or write it with the top model. A single tweet is ≤280 characters — count BEFORE opening the browser; longer → a thread (each tweet ≤280, chained by reply) or offer the operator a shorter version.
    
    ## 2. Tier-2 gate (draft-first)
    Show the final text (+ the thread split, if it is a thread) → wait for an explicit `+`. Nothing is published without it.
    
    ## 3. The browser (Claude-in-Chrome, a live tab)
    > Browser work stays local on this machine; never drag the window to the foreground elsewhere.
    > ⛔ IP gate (2026-07-16): posting to X and other ban-sensitive social platforms happens ONLY from the hub `HUB-1` (a stable IP). Do NOT post from another machine — send the task to the hub as text. Canon: the "IP-sensitive actions from the hub only" rule.
    1. `list_connected_browsers` → no extension → block and tell the operator (do not fall back to Playwright).
    2. Open the `x.com` tab. **Verify the logged-in handle** (avatar / profile menu) against the registry. A different account → STOP and ask. Not logged in → block (we never touch login/2FA on X — checkpoint risk).
    3. Composer: `find` "post composer / What's happening". Enter the text (`form_input`). **Do not press Post** until the gate in §2 is satisfied.
    4. After the `+` — press Post. For a thread: use the "+" button in the composer after the first tweet, or reply to your own tweet.
    5. A screenshot of the published post + the tweet URL (click the timestamp → address bar) is the proof.
    
    ## 4. Record it (AFTER a successful publication)
    ```bash
    python "$USERPROFILE/.claude/scripts/_shared/social_guard.py" record x --text "<text>"
    ```
    Report: the link + the screenshot + "x today N/6".
    
    ## Stop switches
    - Any checkpoint / captcha / "unusual activity" from X → STOP immediately, report, zero retries.
    - Never republish identical text; never post from someone else's account.
    - Links in the text must be live and ours (link-safety).
    
    ## Related
    `/fb-post` (the pattern this follows) · `/tg-post` · `/episode` (tiers: EN teaser → X) · the gate `scripts\_shared\social_guard.py` · the registry `00-System\Channels-Registry.md`.
    
    <!--kit-footer-->
    
    ---
    
    **Like this skill?** It is part of [second-brain-starter-kit](https://github.com/tonydzi/second-brain-starter-kit): the second brain we built for ourselves and run every day at Palo Alto AI Research Lab. Install the whole set with `npx skills add tonydzi/second-brain-starter-kit`. Everything is open source and free, so take what you need.
    
    Flagships worth a look on their own: [secondop-panel](https://github.com/tonydzi/secondop-panel) (a second opinion from a panel of external models), [claude-memory-tidy](https://github.com/tonydzi/claude-memory-tidy) (stop your agent's memory from rotting), [telegram-mcp-kit](https://github.com/tonydzi/telegram-mcp-kit) (your own Telegram over MCP in about 15 minutes).
    
    Author: **Anton Dziatkovskii**, Palo Alto AI Research Lab. Telegram [@tonydzi](https://t.me/tonydzi) - WhatsApp [+1 341 222 9178](https://wa.me/13412229178) - X [@Tony_Stef_](https://x.com/Tony_Stef_)
    
    **Engineers: want to test-drive this setup?** Message me. I hand out free starter seeds to engineers who test and report back, and custom skill requests are welcome.
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.