本文へ移動
cccskills
無料GitHub で公開

darkseoking-mindset

Use when facing SEO algorithm questions, content strategy decisions, Threads growth planning, GEO optimization, AI tool selection for SEO, or evaluating SEO vendors. Triggers on any algorithm or content marketing decision — even without mentioning darkseoking. Every insight is backed by patent numbers or tested data.

インストール方法を見る

含まれるファイル(9)

  • SKILL.md4.0 KB
  • references/ai-tooling.md2.4 KB
  • references/algorithm-analysis.md4.2 KB
  • references/experiment-framework.md2.9 KB
  • references/geo-ai-citation.md2.8 KB
  • references/operational-philosophy.md5.2 KB
  • references/seo-content-strategy.md4.7 KB
  • references/threads-growth.md4.4 KB
  • references/vendor-evaluation.md3.3 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

darkseoking-mindset

A battle-tested analysis framework distilled from a practitioner with 10+ years of SEO experience who spent real money on controlled experiments. The value: nothing here is theory — every point is backed by patent filings or test data. You don't need to imitate anyone; just use these mental models to break down problems.


Eight Core Mental Models

1. Reverse-engineer from patents, don't guess

Algorithm logic is written in patent filings and litigation documents. Patents must describe real mechanisms (legal requirement), making them more reliable than any KOL's speculation. Litigation documents are even stronger — executives testify under oath.

2. Pay to test, let data decide

Design controlled experiments to verify strategies. Isolate variables, run sufficient duration, judge by data — not feelings. "Seems effective" and "proven effective" are different things.

3. Question "industry consensus" first

Many "everyone does this" practices are wrong when traced to underlying logic. More perfect SEO looks less natural; flooding extensions after a viral post gets penalized; high follower counts can drag you down — all counterintuitive conclusions from first-principles reasoning + real testing.

4. Find cross-platform common logic

Google, Meta, and Bing algorithms share ~80% underlying logic. Master one platform's mechanics deeply, then transfer to others quickly. Don't treat each platform as an island.

5. Find the boundary conditions

Every system has rules, and rules always have edge cases. Understanding boundaries isn't about cheating — it's about knowing when standard approaches don't apply and when shortcuts exist.

6. Minimize cost, maximize experiments

Never spend $100 testing one hypothesis when you can spend $100 testing ten. Free tiers, account rotation, open-source alternatives, cheap models for grunt work — the goal isn't being cheap, it's running more experiments per dollar. The person who tests 50 ideas at $2 each beats the person who tests 5 ideas at $20 each.

7. Toolify knowledge to kill middlemen

When you understand a process well enough, turn it into a tool and give it away. This does two things: eliminates low-end vendors who profit from information asymmetry, and forces the market to value real expertise over basic operations. The "Anti-Low-End SEO" series is this principle in action — every free tool released makes one more paid service obsolete.

8. Account positioning is irreversible

Your Creator Embedding is shaped by what goes viral, not what you intend to post. Once algorithm-topic followers flood in, your original niche content underperforms because the new audience doesn't engage with it. darkseoking experienced this firsthand: algorithm patent posts attracted 14k followers who don't engage with SEO content. The lesson: never chase off-topic virality on your main account. The damage compounds and cannot be undone quickly.


Scene Index

SituationRead this
Need to understand how a platform's algorithm worksreferences/algorithm-analysis.md
Planning website SEO content strategyreferences/seo-content-strategy.md
Growing a Threads account, optimizing postsreferences/threads-growth.md
Want content cited by ChatGPT/AIreferences/geo-ai-citation.md
Evaluating SEO vendors or writersreferences/vendor-evaluation.md
Choosing AI tools for SEO writingreferences/ai-tooling.md
Unsure if a strategy actually worksreferences/experiment-framework.md
Thinking about cost structure, scaling as one-person team, or toolifying expertisereferences/operational-philosophy.md

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Agent Skill design knowledge base — mechanisms, philosophy, patterns, pitfalls. Use when: designing new skills, reviewing skill quality, or deciding whether something should be a skill.

日本語の概要は準備中です。原文の説明を表示しています。

cablate/ai-toolkit162026年4月6日 更新

Claude Code 專案配置審計。觸發:review/優化 CLAUDE.md、skills、settings、定期清洗累積內容、新專案上線前檢查。

日本語の概要は準備中です。原文の説明を表示しています。

cablate/ai-toolkit162026年4月6日 更新

Agent 配置設計指南 — 基於 Claude Code 6 個 built-in agent 的逆向分析。Use when: 設計新 agent、優化現有 agent prompt、決定工具/模型配置、撰寫 dispatch prompt。

日本語の概要は準備中です。原文の説明を表示しています。

cablate/ai-toolkit162026年4月6日 更新

AI Agent 成本工程 — 基於 Claude Code 的成本追蹤、prompt cache 最佳化、token 預算控制逆向分析。Use when: 優化 token 消耗、設計成本控制機制、分析 cache 效率、選擇模型配置。

日本語の概要は準備中です。原文の説明を表示しています。

cablate/ai-toolkit162026年4月6日 更新

Harness Engineering 設計模式 — 基於 Claude Code 原始碼逆向分析的 12 條可遷移原則。Use when: 設計 agent 系統架構、實作 tool orchestration、設計 context 管理策略、建構 agent loop。

日本語の概要は準備中です。原文の説明を表示しています。

cablate/ai-toolkit162026年4月6日 更新

System Prompt 工程 — 基於 Claude Code 914 行系統提示詞的逆向分析。Use when: 撰寫 system prompt、設計 prompt 動態組裝、最佳化 prompt cache 效率、撰寫安全指令。

日本語の概要は準備中です。原文の説明を表示しています。

cablate/ai-toolkit162026年4月6日 更新

cablate のスキルをすべて見る

このスキルの問題を報告する