Home / All Skills / Workflow & Core Tools / x-mastery-mentor
English (EN)简体中文 (ZH)日本語 (JA)
Workflow & Core Tools · alchaincyf/nuwa-skill

x-mastery-mentor

A $10K/hr-level X/Twitter operations mentor. Built upon the methodologies of six top creators—Nicolas Cole, Dickie Bush, Sahil Bloom, Justin Welsh, Dan Koe, and Alex Hormozi—combined with in-depth analysis of the X open-source algorithm and specialized strategies for the AI/tech niche, this skill distills 6 core mental models, 10 decision heuristics, and a comprehensive topic-ideation, writing, and growth playbook. General methodology serves as the foundation, with specialized focus on the AI/tech vertical. Trigger when users mention "X operations", "Twitter", "how to write a tweet", "how to grow followers", "X strategy", "Twitter topic ideation", "tweet", "thread", or "X algorithm". It should also trigger if the user simply says "how do I write this tweet", "help me brainstorm X content", "Twitter growth", "post a tweet", "write a tweet", "X account", or "grow on X".

L2 Local · Restricted workspaceV1 Commit-PinnedWrite FileDual-pass Audit
Author alchaincyfGitHub alchaincyf/nuwa-skill ↗License MIT（仓库根 LICENSE）Commit fe03746870
Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

Declares trigger words (X operations/Twitter/write a tweet/grow followers/X algorithm/tweet, etc.) in the frontmatter description to govern activation, and explicitly hardcodes capability boundaries in the body using "What I can help with / What I cannot help with" to prevent overstepping into other platforms such as Douyin or Xiaohongshu.

examples/x-mastery-mentor/SKILL.md
**我能帮你的**：选题策略、推文写作、Thread结构、增长引擎、算法利用、AI赛道内容打法、变现路径、账号诊断
What it does: Explicitly writes activation conditions and capability boundaries into the frontmatter and body, clearly stating that this skill only focuses on X when the user asks about out-of-scope platforms.

The question routing table categorizes user queries into 11 types, mapping them to 5 execution scenarios (A-E) or direct answers, and designates on-demand loaded reference files for each category to avoid loading all knowledge at once.

examples/x-mastery-mentor/SKILL.md
| 怎么写推文/Thread | → 场景A | `writing-workshop.md` + `algorithm-niche.md` |
What it does: Decouples the main file from references using a "question type → execution scenario → knowledge file" routing table to control context costs.

Scenario A (Tweet/Thread Writing) executes in four steps: confirm type & objective → generate 3 Hooks → flesh out the body using a 1/3/1 rhythm → review against a quality checklist; sets a checkpoint after generating Hooks for user selection.

examples/x-mastery-mentor/SKILL.md
每个标注用了哪个公式（好奇缺口/可信度锚点/Value Equation）
What it does: Deconstructs "tweet writing" into an inspectable, fixed pipeline, requiring each Hook to state the formula used rather than relying on intuition.

Scenario D (Growth/Strategy) first identifies the follower tier (0-1K / 1K-10K / 10K-100K) and Premium status, then diagnoses bottlenecks, and finally delivers a weekly plan focused on "actions, not principles".

examples/x-mastery-mentor/SKILL.md
粉丝量？（决定路由到0-1K/1K-10K/10K-100K）
What it does: Enforces tier confirmation before offering recommendations to avoid misaligning strategies across different follower scales.

Scenario E (Account Diagnosis) uses a 3-tier fallback path to collect the latest ~100 tweets (computer-use as primary → claude-in-chrome as backup → manual user input as fallback), persists data to user-data/, and generates an Economist-style HTML report.

examples/x-mastery-mentor/SKILL.md
目标100条，每滚动一屏约10条，需滚动约10次
What it does: Formats data collection as a fault-tolerant "tool + fallback" workflow, specifying minimum sample sizes and time spans.

Sets a hard STOP/CHECKPOINT gate before each critical scenario (three questions each for Scenarios A/D/E); answering "No" to any rolls back to the corresponding step, preventing unvetted, gut-feeling outputs.

examples/x-mastery-mentor/SKILL.md
1. **Hook 用了哪个公式**（好奇缺口 / 可信度锚点 / Value Equation）？说不出 = 凭感觉写
What it does: Front-loads quality control as self-check questions rather than patching issues post-generation.

A 9-rule Failure Mode Fallback Tree provides primary and alternate paths based on trigger signals, such as immediately switching to method 2/3 when a scraping tool fails instead of stubbornly retrying the same path.

examples/x-mastery-mentor/SKILL.md
| 6 | 工具失败(computer-use 登录墙 / Chrome 扩展未连) | 立即切方式 2 → 方式 3,不要硬重试同一条路 | 退到「用户手动提供数据」,标注样本受限 |
What it does: Pairs every failure signal with a recovery path to minimize the probability of getting stuck.

2Core Capabilities

01Tweet/Thread writing and Hook engineering (including 1/3/1 rhythm and 8-12 tweet Thread structures)
02Topic ideation (4A Matrix / Endless Idea Generator, 100+ ideas in 30 minutes)
03Content review and tiered diagnosis for drafts (Algorithm tier → Hook tier → Content tier → CTA tier)
04Tiered growth strategies and actionable weekly plans (0-1K / 1K-10K / 10K-100K)
05X algorithm rule quick-lookup (engagement weights, negative signals, posting parameters, TweepCred)
06Account data collection and 6-dimensional diagnostic reports (including Economist-style HTML templates)
07AI/tech niche account categorization, content performance matrix, and global positioning advice
08Monetization pathways and comparison across six creator schools of thought (Cole/Bush/Sahil/Welsh/Koe/Hormozi)

3External Dependencies

KindDependency / Endpoint
cli宿主 computer-use / claude-in-chrome 浏览器工具（场景E采集推文用；由宿主提供，非 skill 自带脚本）
networkx.com 用户主页 / analytics.x.com（采集目标站点，经宿主浏览器工具访问，非 skill 自身端点）

4Recommended Governance & Evidence

Suggested Controls for Safe Operation
For Agent Runtime
Restrict working directory
For Administrators
Retain operation logs
Evidence & Triggers (File Paths & Excerpts)
  • 本地工作目录: File I/O and artifact output: user-data/{username}/tweets_{date}.json and tweets_{date}.md—Scenario E Step 3 specifies saving collected data
Audit Notes: Grading basis (evaluating only this skill directory): The contents of examples/x-mastery-mentor consist entirely of Markdown files (SKILL.md + FIDELITY.md + references/), with no scripts, no executable assets, and no reading of environment variables or credentials. However, SKILL.md explicitly requires writing JSON/Markdown/HTML files under user-data/{username}/ at runtime (Scenario E) and silently reading strategy.md, falling under the "local file I/O" surface, thus classified as blue. Non-directory aspects to note (do not alter classification): Tweet collection in Scenario E relies on the host's computer-use / claude-in-chrome browser tools to navigate to https://x.com/{username}, which belongs to host capabilities; the report template imports ECharts.js via CDN, which is an external artifact reference; neither is executed via native skill code. No hazardous flags, no guardrail bypasses, and no arbitrary code execution.

5Second-Pass Verification

  • [ok] 能力边界与 frontmatter 触发词一致 — SKILL.md frontmatter 的 description 列触发器（X运营/推特/写推文/涨粉/X算法/tweet 等），正文「我能帮你的／我不能帮你的」与之呼应，并明确「超出skill范围时明确说」（如抖音/小红书），三处口径一致。
  • [ok] 问题路由表条目完整且与场景编号对应 — 路由表共 11 行，覆盖场景A-E 与「直接回答」五类；每行「按需加载」列出的 reference 文件都存在于 §1 清单（writing-workshop / algorithm-niche / quality-analytics / growth-monetization / mental-models-heuristics）。
  • [ok] 五场景流程、检查点与 STOP 三问描述属实 — 场景A 四步（确认类型→3个Hook→完善正文→质量检查）、场景D 三阶段、场景E 五步均可在内文逐条找到；STOP 段落确有「场景A/D/E · 必答 3 问」三组，与 claim 相符。
  • [ok] references 文件与 §1 清单一一对应 — §1 列出的 FIDELITY.md、SKILL.md、5 份操作层 reference 与 6 份 research 报告，正好对应 Reference 索引表的分层（操作层按需加载 / 调研层追溯来源），无缺失或多列。
  • [ok] §3 风险扫描 274 条 URL 的性质归类 — 抽样命中（x.com/Nicolascole77、ship30for30.com、artandbiz.substack.com、x.com/{username}）除场景E 的访问目标外，均为调研正文里的来源链接引用；无运行时 API 端点、无 webhook。凭证类唯一命中是同一条 substack URL，属误报。
  • [ok] 目录内无脚本 / 无凭证读取 / 无写盘代码 — §1 全部 13 个文件类型均为 text，§3「执行外部程序/写盘/网络请求/危险旗标/沙箱」五类均 0 命中；SKILL.md 里的 user-data/ 写入是给宿主的自然语言指令，非本目录代码。

6Verdict & Scenarios

  • Structured operations manual: Five scenarios (tweet writing / topic ideation / review / growth / diagnosis) + question routing table + checkpoints break ambiguous "how to do X" into an executable pipeline with quantifiable metrics (short tweets 120-130 chars, ER > 2% is good, post weekdays 9 AM-2 PM).
  • Layered and traceable knowledge: The main file only handles routing, 5 operational reference files load on demand, and 6 research reports are read only when tracing sources, with research dates and sources explicitly marked for algorithm data.
  • Anti-hallucination and boundary awareness: Distinguishes advice into "community consensus / speculation / experimentation", explicitly refuses out-of-scope queries (e.g., Douyin/Xiaohongshu), and includes 6 honesty boundaries in the body to reduce misleading advice.
  • Fault-tolerant design: 9 failure mode fallback trees + Scenario E's 3-tier data collection fallback path (computer-use → claude-in-chrome → manual user input), requiring explicit labeling when sample sizes are < 50.
    • Best for: Ideal for operators of X/Twitter accounts who want to systematize tweet writing, topic ideation, draft reviews, follower growth strategies, and account diagnostics—especially indie hackers and global creators in the AI/tech niche. Also serves as an exemplary blueprint for multi-scenario operations skills (complete with routing table, on-demand references, checkpoints, fallback trees, and anti-pattern blacklists).
    Not for: Not suitable for operations inquiries regarding non-X platforms like Douyin, Xiaohongshu, or WeChat Official Accounts; not suitable for users demanding "guaranteed follower growth", ghostwriting, or predicting future algorithm changes; not suitable for environments unable to run host browser tools or unwilling to persist files in user-data/—Scenario E will degrade or fail, leaving manual user data input as the only fallback.
    Install Agent-Ready
    ① Verified Mirror 2026-10-10
    Option A · Direct DownloadDownload x-mastery-mentor.tar.gz
    sha256: 8e5f288e5d0c82be…
    Option B · Copy Guide for Agent
    install.json
    Agents automatically download and verify mirrors from this JSON guide.
    ② Upstream GitHub
    GitHub Repo ↗
    Pinned to commit fe03746870.