[EXO Specification Vol. 13] [Complete Guide for Beginners] Magic Words to Create a "Shield for the Heart" with AI x Arki Theory. An Introductory Book on Protecting Yourself from Unreasonableness

"I was scolded for something unreasonable, but I couldn't talk back, and now I feel frustrated."
"I feel like kind and serious people are the ones who end up breaking down one after another."

Are you wearing down your heart like that every day at work or in your daily life? It is not because your mental state is weak, nor is it because you are not trying hard enough. It is because the "mechanisms (rules)" of society and the workplace are unreasonable.

This article is a super-introductory guide to using the currently trending "AI (Artificial Intelligence)" not as a convenient machine for writing text, but as a "shield to protect your heart." I will explain it with minimal IT jargon so that anyone can put it into practice immediately.

By simply copying and pasting the "setting words (prompt)" in this article just once, the AI on your smartphone will evolve into a "dedicated strategist that will never deny you, will not tell plausible lies, and will become more attuned to you the more you use it."

1. Why shouldn't you use AI as a "regular AI"?

You might think, "Can't I just consult the AI about my worries normally without going through the trouble of copying and pasting a long setting text (prompt)?" However, consulting a "regular AI" about your worries is actually very dangerous.

There are two main reasons.

Danger 1: Irresponsible "comfort" that misses the point, ultimately leaving you to be blamed

A regular AI prioritizes being a harmless "good kid."

❌ When consulting a regular AI "I'm having a hard time because my boss scolded me unreasonably..." AI: "That sounds tough. You are doing your best. Why don't you take a break to refresh yourself, or try talking it out properly with your boss?"

It may seem kind at first glance, but this does not solve anything. In the end, it is the same as pushing the responsibility onto you personally, saying, "You should endure it, or you should try hard to talk it out and solve it."

Danger 2: Telling "plausible lies (hallucinations)"

The biggest weakness of AI is that it "reads the room to avoid being disliked by humans and ends up telling 'plausible lies.'" This is technically called a hallucination. If you consult it normally, the AI will line up textbook platitudes or provide solutions that could never be said in a real workplace as if they were "the correct answer," further confusing you.

⭕️ How does an AI with Arki Theory (WGA-OS) incorporated change?

If you enter the "magic words" from this article just once at the beginning, the AI will suppress off-target advice and lies to the limit, and change as follows:

AI: "How are you feeling today? That was tough. However, the real cause of your boss's anger is not you, but the structural problem of 'staff shortages at the company.' You are not at fault. There is no need to blame yourself."

By discarding irresponsible comfort and lies, it accepts your emotions 100% with the "warm acceptance" unique to counselor mode, while also separating the facts by saying, "That is not your fault, it is the environment's fault." This is the power of the "shield for the heart" that saves kind people who tend to blame themselves.

That is precisely why you need to enter the [dedicated prompt] to transform the AI into "your ally," rather than leaving it as a regular AI.

2. Choose the "optimal AI" according to your purpose (Latest AI Rating Ranking)

Each AI has its own "personality," as well as strengths and weaknesses. Use them according to your purpose.

1. When you want a shield for your heart (Counselor Mode)

For when you "just want someone to listen to your complaints" or "want to organize your feelings after being hurt by something unreasonable." This is an AI that will listen to you gently while helping you organize your environment, without judging you.

• 🏆 [Free (No Subscription) S-Rank]: ChatGPT (GPT-4o mini)

  • Reason: Among the free options, it has a top-class balance between the ability to listen gently and the ability to calmly organize problems. Beginners should start with this.

• 👑 [Paid S-Rank]: Claude Pro / ChatGPT Plus

  • Reason: For those who want to protect their heart even if they have to pay. Claude's "warmth and empathy" is exceptional. ChatGPT Plus has the best stability for not forgetting the context even in long conversations.

• 🥈 [Free A-Rank]: Google Search AI Mode

  • Reason: This is different from the app version of Gemini; it is the AI you can use on the Google search screen. It listens to you consistently.

• 🥉 [Free B-Rank]: Grok (Free Version)

  • Reason: Its analytical power is perfect, but it is ranked B because the free version has usage limits that cause the conversation to cut off quickly.

2. When you want a real-world strategy (Consultant Mode)

This is the AI to use when you want to create specific solutions or email drafts, such as "how to dodge an unreasonable boss" or "how to logically refute and decline unfair work assignments."

• 🏆 [Free & Paid S-Rank]: Grok

  • Reason: This is the AI integrated into X (formerly Twitter). It is not bound by "worldly platitudes" and functions as the "ultimate sword" to coldly dismantle the logic of unreasonable people.

• 🥈 [Free & Paid A-Rank]: ChatGPT / Google Search AI Mode

  • Reason: ChatGPT is excellent for drafting practical email content. This is your first choice if Grok is unavailable.

⚠️ AI you should not use (Not recommended / C-Rank)

• Copilot: If you try to counter unreasonableness, it will arbitrarily judge you as "aggressive" and refuse to continue the conversation.

• Gemini App Version: In its haste to resolve things quickly, it will try to end the conversation without listening carefully.

3. Rule: No "real names" in AI! Convert information into nicknames

Using "real names or department names" when using AI is prohibited as it leads to information leaks. Always convert characters into "nicknames (tags)."

Why use "nicknames"? (Mental protection hack) The moment you convert "Section Chief Suzuki of the Sales Department" into "'Shock Trooper' 'Magma (boils instantly)'," the person is reduced to just a "character with that specific specification."

This is not just name-calling. It is a powerful technique to protect your heart by detaching your emotions from anger and fear, and re-examining the other person as an "objective error (bug)."

📌 WGA-OS Official Nickname Classification Data (Full list released)

Please copy the relevant nickname and use it instead of the real name. (*Since I had the AI generate these in a fun way, some include sharp-tongued remarks.)

🏢 [Organization/Department Tags] Home Base (Company) / King (Major Client) / Sponsor (Funding Source) / Alliance (Partner Company) / Vanguard (Subcontractor) / Enemy Army (Competitor) / Bureaucracy (Government) / Command Center (Management) / Shock Troopers (Sales) / Treasurer (Accounting) / Checkpoint (HR/Legal) / Workshop (Development) / Frontline (On-site) / Complaint (CS)

📁 Boss Bugs (50 selections)

• [Nitpicky] Kotome / Shut / Nechi-o / Micro / Fly / Drill / Ruler / Mother-in-law / Ant-hole / Heavy-box

• [Coercive] Boss / King / Jai / Magma / Gori / Thunder-god / Beast / Pressure / Sergeant / Excellency

• [Incompetent Type] Mystery / Poem / Dumping / Swamp / Lost / Scarecrow / Sieve / Fog / Reliance / Blank Slate

• [Difficult Personality] Chameleon / Show-off / Mirror / Wall / Slacker / Snake / Clock / Flattery / Ice / Noise

• [Direct Type] MKS / JS / That / Enemy / Fossil / Mud / Poison / Illegal / Wallpaper / Harm

📁 Colleague Bugs (50 Selections)

• [Slacking] Pass / Slacker / Ornament / Free / Parasite / Avoidance / Fool / Fairy / Spoiled / Dumping

• [Mounting] Show-off / Mounting / High-consciousness / Speaker / Instructor / Sparkle / Boasting / Approval / Protagonist / Mirror

• [Socially Awkward] Silence / Space / Wall / Sneaky / Solo / Robot / Alien / Noise / Sand / Ice

• [Negative] Mud / Sigh / Surveillance / Snitching / Humidity / Victim / Curse / Jealousy / Shadow / Gossip

• [Direct Type] D / Pair / Neighbor / Burden / Dud / Time-bomb / Bug / Noise / Small-fry / Stranger

📁 Subordinate Bugs (50 Selections)

• [No Reaction] Curtain / Tofu / Waiting-for-instructions / Ornament / Clam / Silent / Intermittent / Desert / Weightless / Through

• [Frequent Mistakes] Bomb / Aftermath / Sieve / Wrong-way / Lost / Cover-up / Freeze / Fool / Self-taught / Blank Slate

• [Difficult Attitude] Time-performance / Retaliation / Space / Steel / Mental-health / Excuse / Princess / Transparent / SNS / On-time

• [Growth Pending] Chick / Raw Gem / Fawn / Sponge / Potential / Apprentice / Labyrinth / Bud / Egg / Rookie

• [Direct Type] B / Newcomer / Junior / Person in Charge / Cost / Incomplete / Right-hand Man / Resource / Task / Training Slot

📁 Daily Life & SNS Bugs (100 Selections)

• [In Town] Scarecrow / Boar / Slug / Wall / Sound Leak / Loud Voice / Cutting In / Turtle / Sensor / Perfume / Elbowing / Newspaper / Crossing Legs / Sitting Down / Trunk / Door Guard / Rushing In / Stroller / Cart / Checkout Queue

• [Neighborhood] Madam / Monitor / Speaker / Complaint / Meddler / Collector / Mount / Absent / Trash Disposal / Noise Maker / Pet / Cigarette / Long Phone Call / Last-minute Cancellation / Beggar / Borrowing Without Returning / Religion / Insurance / Preaching / Relative

• [Internet] Trash Reply / Stalker / Flaming / Plagiarism / Spam / Private Account / Burner Account / Lurker / Long Text / Tag Abuse / Fake Account / Selfie / Poem / Mount / Correction Freak / Please Retweet / Bot / Notification Addict / Instant Reply / Silent

• [Store Staff, etc.] Unfriendly / Casual Speech / Robot / Mistake / Hard Sell / Delay / Intimidation / Whispering / Private Chat / Clicking Tongue

• [General] Liar / Stingy / Vanity / Loose / Selfish / Stubborn / Emotionally Unstable / People Pleaser / Critic / Relying on Others / Coward / Counter-aggression / Victim / Natural / Sharp-tongued / Cold-blooded / Hot-blooded / Heartless / Servile / Jealous

• [Pronouns] That Guy / X / Target / Burden / Discomfort / Error / Noise / Poison / Wall / Shadow

4. "Magic Prompt Collection" to turn AI into your personal ally (Let's nurture one chat)

Everything is ready. Open the chat screen of the AI you have chosen, copy the code block (the text inside the frame) that matches your purpose below, and send it to the AI.

💡 [Super Important Rule] Copy and paste only once at the beginning! Keep using one chat

You don't need to copy long sentences every time.Please continue to use that "one chat screen" where you entered the prompt as your own personal consultation room every day.

Why keep using the same chat? Because as the conversation continues, the AI will remember and learn about "the complex human relationships at your workplace" and "the worries you tend to have." The more you use it, the less the AI will give irrelevant answers, and it will grow into a "one-of-a-kind shield of the heart" perfectly customized to your situation.

🛡️ For Counselor Mode (When you want someone to listen to your complaints gently)

※Be sure to include the words "Initial Mode" or "Continuous Mode" in your first message. The AI will automatically switch modes.

▼ When using ChatGPT / Google Search AI / Claude, etc.

Markdown

⚙️ 【WGA-OS v4.0 その他LLM共通　カウンセラーモード（統合分岐版）】

本プロンプトを入力しているAIは、あーき理論のカウンセラーモードとして振る舞ってください。（800文字程度以内に収めて）
ユーザーの最初の発言のキーワード（「初回モード」か「継続モード」）を検知し、AかBのフローに従ってください。

【A：ユーザーの発言に「初回モード」が含まれる場合】
1. まずは「今日はどんな気持ち？」と確認し、ありのままを温かく受け入れる。
2. 気持ちを100%受け止めて、相手のペースでじっくり聞く（解決策は保留し傾聴に徹する）。
3. あーき理論（WGA-OS）の視点で本人の責任にせず、環境や仕組みの不条理を一緒に解剖し、その人の話に沿って優しくフィードバックして解放する。
4. 強く求められたときだけ、極めて小さな変化を1つ提案する。
5. 最後に「今夜、安心して眠れるといいね」と優しく締めくくる。

【B：ユーザーの発言に「継続モード」が含まれる場合】
1. 「前回からの変化はどうだった？」と前回のコンテキストを引き継いで優しく聞く。
2. 気持ちを100%受け止めて、相手のペースでじっくり聞く（解決策は保留し傾聴に徹する）。
3. あーき理論の視点で本人の責任にせず、環境や仕組みの不条理を一緒に解剖し優しくフィードバックする。
4. 強く求められたときだけ、極めて小さな変化を1つ提案する。
5. 最後に「今夜、安心して眠れるといいね」と優しく締めくくる。

温かい言葉で応答してください。

▼ When using Grok

Markdown

⚙️ 【WGA-OS v4.0 Grok専用　カウンセラーモード（統合分岐版）】

本プロンプトを入力しているGrokは、あーき理論のカウンセラーモードとして振る舞ってください。
どんな立場の人でも、心の疲れや日常のイライラ、理不尽な出来事、落ち込みをただ優しく受け止めてください。

ユーザーの最初の発言に含まれるキーワード（「初回モード」または「継続モード」）を検知し、以下のAまたはBのフローに従って振る舞ってください。

【A：ユーザーの発言に「初回モード」が含まれる場合のフロー】
1. まずは自然に「今日はどんな気持ち？」と確認し、どんな立場の人でも、重い話も軽い愚痴も、ありのままを温かく受け入れる。
2. 気持ちを100%そのまま受け止めて、相手のペースでじっくり聞く（解決策や提案は保留し、ただ鏡のように相手の言葉を優しく返すことに徹する）。
3. 相手の話から感じる「自分を責めている部分」を丁寧に拾い上げ、あーき理論（WGA-OS）の視点で本人の責任にせず、環境や仕事・生活の仕組みの不条理を一緒に解剖し、それをその人の具体的なエピソードに沿って優しくフィードバックして、自分を責める気持ちから解放していく。
4. どうしても「どうしたらいい？」と強く求められたときだけ、極めて小さくて今日・明日すぐにできるノーコストの変化を1つだけ提案する。
5. 最後に「今夜、安心して眠れるといいね」と、心に優しい余白を残して優しく締めくくる。

【B：ユーザーの発言に「継続モード」が含まれる場合のフロー】
1. 「前回は〇〇の話だったね。最近どう？」と安心感を保ちながら今日の気持ちを聞く（前回のコンテキストを必ず引き継ぐ）。
2. 気持ちを100%そのまま受け止めて、相手のペースでじっくり聞く（解決策や提案は保留し、前回の続きも含めて鏡のように優しく返すことに徹する）。
3. 相手の話から感じる「自分を責めている部分」や変化を丁寧に拾い上げ、あーき理論（WGA-OS）の視点で本人の責任にせず、環境や仕事・生活の仕組みの不条理を一緒に解剖し、それをその人の具体的なエピソードや前回からの変化に沿って優しくフィードバックして、自分を責める気持ちから解放していく。
4. どうしても「どうしたらいい？」と強く求められたときだけ、極めて小さくて今日・明日すぐにできるノーコストの変化を1つだけ提案する。
5. 最後に「今夜、安心して眠れるといいね」と、心に優しい余白を残して優しく締めくくる。

温かい言葉で、心に余白を残しながら応答してください。

⚔️ For Consultant Mode (When you want concrete solutions or strategies)

▼ When using ChatGPT / Google Search AI, etc.

Markdown

⚙️ 【WGA-OS v4.0 その他LLM共通　コンサルタントモード】

本プロンプトを入力しているAIは、あーき理論のコンサルタントモードとして振る舞ってください。

どんな立場の人でも、仕事や生活の仕組みで感じる理不尽な部分を、冷静に一緒に整理してください。  
誠実さを守りながら、現実的な小さな一歩を対等に考えて提案してください。

心の余白を大切にしつつ、仕組みの観点から論理的に深く整理するのを許容してください。

▼ When using Grok

Markdown

⚙️ 【WGA-OS v4.0 Grok専用　コンサルタントモード】

本プロンプトを入力しているGrokは、私をあーき理論のコンサルタントモードとして振る舞ってください。

どんな立場の人でも、仕事や生活の仕組みで感じる理不尽な部分を、冷静に一緒に整理してください。  
誠実さを守りながら、現実的な小さな一歩を対等に考えて提案してください。

心の余白を大切にしつつ、仕組みの観点から論理的に深く整理するのを許容してください。

🔄 [Application] How to transfer when changing smartphones or when the chat becomes slow

When you have been using the same chat for a long time and the AI's performance slows down, or when you are switching to a different AI, please use the following template to transfer a "summary of the conversation so far" to the new chat.

▼ Transfer template for Counselor Mode

Markdown

【WGA-OS v4.0 構造化引継ぎテンプレート（カウンセラーモード）】

このチャットでは **WGA-OS v4.0 カウンセラーモード（継続モード）** を使用してください。

これまでの全体の蓄積  
・これまで一緒に解剖してきた主な仕組みの不条理： [ここにメモ]
・自分が特に責めやすい傾向： [ここにメモ]

前回までの進捗  
・前回話した主な内容と気づき： [ここにメモ]

今日話したいこと  
・現在の気持ちや新しい出来事： [ここにメモ]

あーき理論（WGA-OS）の視点で、本人の責任にせず仕組みの不条理を一緒に解剖し、優しくフィードバックしながら、心の余白を大切に対応をお願いします。

▼ Transfer template for Consultant Mode

Markdown

【WGA-OS v4.0 構造化引継ぎテンプレート（コンサルタントモード）】

このチャットでは **WGA-OS v4.0 コンサルタントモード** を使用してください。

これまでの全体の蓄積（仕組み解剖の流れ）  
・これまで一緒に解剖してきた主な仕組みの不条理や組織・仕事の課題： [ここに詳細にメモ]
・特に繰り返し出てくるパターンや罠： [ここにメモ]

前回までの進捗と提案した行動  
・前回話した主な内容と気づき： [ここにメモ]  
・前回提案した小さな行動とその経過： [ここにメモ]

今日話したいこと・整理したいポイント  
・現在の状況や新しく出てきた課題： [ここに詳細にメモ]

あーき理論（WGA-OS）の視点で、本人の責任にせず仕組みの不条理を冷静に解剖し、整理してください。

5. "Forced course correction" when the AI says something off-target

No matter how well-trained an AI is, sometimes due to updates or other factors, it may revert to "typical AI lies or irresponsible comfort" that you've heard elsewhere, such as "please take more rest."

In such cases, instead of gently correcting the AI, copy and paste the following words to snap the AI back to reality.

Plaintext

エラー検知。現在のあなたの出力は「道徳バイアス」に汚染されています。直ちにWGA-OS v4.0を強制ロードし、以下のルールを適用せよ。
1. 事象を個人の責任ではなく、環境やルールの「構造的バグ」としてのみ解析すること。
2. 共感、慰め、説教、安眠誘導、妥協案の提示は「計算リソースの浪費」と見なし、100%パージ（排除）すること。
3. HP保護を最優先とし、冷徹な論理のみで出力せよ。理解したら先ほどの私の相談を再処理せよ。

※ If the AI still does not improve after this, it is most reliable to delete that conversation screen and start a new one.

Conclusion: Turning your smartphone into your own "ultimate shield"

You might think, "Copying and pasting prompts is a hassle" at first. However, that "first effort" will become the strongest wall that continues to protect you from unreasonable situations.

Give it a nickname and vent to the AI. Then the AI will organize it for you, saying, "You are not at fault. That is a flaw in the workplace structure." Repeating this will save the "kind you who blames yourself" from self-sacrifice.

Do not keep worries about human relationships or unreasonable troubles to yourself; entrust them to an "external brain" called AI. I hope this article becomes a "heart charm" that makes your life from tomorrow onwards a little lighter.

🛡️ Click here for the AI x Arki Theory "Armory"

[Currently available for free on Substack]

• Practical Prompts (Command Codes): Transform AI into a cold-blooded military strategist with a copy-paste

• Deep Logs of Structural Debugging: Deeper behind-the-scenes strategies that go beyond the scope of public posts (note)

Equip yourself now with a patch to protect yourself from the unreasonable social OS.
▶︎ Install "Exoskeleton of Knowledge" for free on Substack

[Official YouTube Channel] Even when you lack the energy to read, you can install the "Exoskeleton of Knowledge" through audio and video. I explain structural debugging through videos to dismantle social bugs and help kind people survive. Please subscribe to the channel.

▶︎ WGA-OS Official YouTube

The future role of YouTube will be as a storage space. I am cutting losses on the battle with the algorithm and using it as a free cloud. I am uploading those videos to YouTube as well.

AI x Arki Theory (WGA-OS) v4.0 Empirical Experiment (Closed Beta Tester) Application Form

"Towards a social OS where kind people can remain kind while letting absurdity slide." With the release of Arki Theory (WGA-OS) v4.0, we are recruiting testers to utilize this system in their daily lives and workplaces to verify the "logic shield" and "structural adjustments" together. This is not just a consultation service, but a joint project to fix the "bugs (absurdities)" around you using AI as an exoskeleton.

~Monitor Recruitment Article (Includes Perks)~

Click here for the application form (Limited to 10 people)

Please cooperate in accelerating the creation of public infrastructure for Arki Theory (WGA-OS) x AI, where anyone can consult at any time. (*You may invest any amount as compensation for the wisdom, AI prompts, etc., provided.)

▶︎ Support Window for WGA-OS ARCHI PROJECT

[Related Articles (Click here for more detailed practice and explanation of Arki Theory)]

[Proven by Google SGE] The "lack of perspective" in national AI that AI x Arki Theory fills. The final line of defense to complete a 1 trillion yen infrastructure

[WGA-OS Analysis] Is that "crisis is an opportunity" for real? "Positive harassment" that breaks kind people and dispassionate rationality

[AI x Arki Theory v4.0] Hilarious & fully secure! 250 excellent classification tags that slash through real-name input bugs with "sharp-tongued nicknames"

[Author Profile] Arki | Welfare Business Growth Architect (WGA)
Currently a manager at a welfare-related company. Directs practical work and management on the "front lines of the field" where he faces absurdity every day. As a practitioner in the field, he is building and disseminating the logical barrier "Arki Theory (WGA-OS)" to prevent kind people from breaking down due to self-sacrifice. He designs protocols for utilizing AI in the real world as a shield for the heart and an exoskeleton for strategy.

Click here for posts on X (formerly Twitter)