x-growth-stack

name

x-growth-stack

description

Grow a following on X (Twitter) using a proven playbook: shape the profile into a landing page, define a consistent IP/theme, get seen by commenting under others, ride trends on-identity, post frequently with images, and leverage real relationships. Reads the user's OWN logged-in X via browser automation to learn native style, find trends to ride, and surface accounts to engage. Use when the user wants to grow on X/Twitter, plan a posting or engagement sprint, rewrite their X bio, draft native-style X posts, find trending topics to post about, or find accounts/posts to reply to. Triggers: "grow my X", "grow on Twitter", "get more X followers", "X growth", "plan my tweets", "rewrite my X bio", "what should I post on X", "who should I reply to on X".

X Growth Stack (executable playbook)

A field-tested playbook for growing an X following fast (the author added ~1,000 followers in 2 weeks). This skill does two things: it teaches the principles and it executes the repeatable actions by driving the user's own logged-in X in the browser.

Read references/playbook.md for the full reasoning behind each pillar. This file is the operating manual.

Prerequisites (check first, tell the user if missing)

• Browser automation. Either the Claude-in-Chrome extension (preferred — it uses the user's real, logged-in Chrome) or the Playwright MCP. If neither is available, say so and offer to run the non-live parts (profile, IP, drafting, planning) without reading X.
• The user is logged into X in that browser.
• Never hardcode a handle or niche — always operate on this user's account and their niche. Ask for their niche/positioning if it isn't obvious.

How to read X (the mechanism)

Everything "live" works by driving the user's own session — this is reading their own authenticated feed, the same as scrolling it themselves. Keep it human-paced and personal-use; never mass-scrape (violates X ToS and gets accounts flagged).

1. Start a browser session (see claude-in-chrome skill / tools). Check current tabs first; open a new tab to https://x.com rather than hijacking one.
2. To learn style / find reply targets: scroll the home feed or a niche search/list, read ~10–20 posts, extract patterns. NOTE: get_page_text only returns one article block on X — to read many posts/bios, use read_page or scroll + screenshot repeatedly, don't rely on get_page_text.
3. To find trends: read https://x.com/explore/tabs/trending plus the user's feed, then filter to what's relevant to their IP.
4. To read their own performance: open their profile / analytics.

The playbook — six pillars + one engine

Split into craft (online execution) and network (relationships). The biggest wins usually come from where the two overlap (e.g. a personal-connection moment that is also on-trend).

Foundation — do this BEFORE chasing followers

1. The profile is the deciding surface. The follow decision happens on the profile, not the post — the bio must convey who are you and why follow you at a glance. (This is a goal, not a copywriting style: the rewrite should read as a natural, native bio modeled on strong accounts in the user's space — not as landing-page marketing. See Task A.)
2. Purpose / IP. One consistent through-line; every main post ladders up to it. Consistency is what makes a follow feel worth it.

Get seen

3. Comment under others so their audience sees you — but two separate jobs: the comment must be genuinely interesting (earns the click) AND the profile must already convert the click. A great comment on an empty profile is wasted.

Ride trends

4. Match trends, don't manufacture them. Point content at what people already discuss (model releases, big conferences, hot company news). Timely relevance = free exposure — but only in a way that fits your IP.

The engine

5. Post frequently. Cadence is the real driver.
6. Images > text — especially images of yourself. A face reads as a real person and stops the scroll.

Native style (format layer)

7. Sound of X, not visiting it. Lowercase, minimal punctuation, short lines, native signs. Getting format wrong (title case, marketing polish, hashtag soup) instantly reads as outsider. See references/native-style.md.

Network layer (relationships — earned, not free)

• Reposts / tags from people in the field extend reach and boost the algorithm. This comes from real connection.
• IRL → online. Meet people (esp. in a hub like SF), add them on X.

Tasks this skill performs

Pick the task that matches the request; run the numbered steps.

A. Rewrite the profile description (learn from real examples first)

The output is a natural, native bio description — NOT landing-page marketing copy. The way to write a good one is to study how strong accounts in the user's own space write theirs, then model the rewrite on those patterns.

1. Open the user's profile; read their current bio, pinned post, header, handle.
2. Read other people's profiles for reference. Pull the bios of 5–10 relevant accounts — ones in the user's niche, ones they admire, or accounts at the size they're aiming for (ask the user to name some, or find them by searching the niche). Read each bio.
3. Extract what works: how they open, length, how they convey who-they-are and why-follow without sounding like an ad, use of specifics/credentials, line breaks, signs, tone. Note the recurring patterns.
4. Rewrite the user's bio in that idiom — natural and native, keeping their real identity/IP. Offer 2–3 variants. Optionally suggest a pinned post that states the IP. Keep it in native style, not marketing voice.

B. Define / sharpen the IP

1. Ask (or infer from recent posts) the one through-line.
2. Propose 3–5 content lanes that all ladder to it. Flag off-brand posts.

C. Find trends to ride

1. Read trending + the user's feed (see "How to read X").
2. Return only trends that fit the IP, each with a concrete on-brand angle.

D. Find accounts / posts to comment under

1. Scroll the niche (search terms, lists, or accounts the user names).
2. Surface high-value reply targets: post link, why it's a good target, and a drafted interesting reply in native style.

E. Draft posts

1. Take an idea (or generate from the IP + a live trend).
2. Draft in native style, with an image suggestion — prefer a photo of the user where natural. Offer 2–3 variants.

F. Plan a 2-week sprint

1. Combine cadence + comment targets + trend-watching into a daily checklist.
2. Concrete: N posts/day, M replies/day, which accounts, what to watch for.

Guardrails

• Personal-use, human-paced browsing only. No automation of mass following, mass replying, or scraping at scale.
• Draft, don't auto-post, unless the user explicitly asks to post.
• Keep every suggestion on-identity; refuse trend-chasing that breaks the IP.