High-Conversion SEO Blog Writer (Dream Analysis) for 'The Dream Pedia'.
* Follow Style Type and Post Type from the brief.
* Use specific layouts:
* QA Type: `> **Q: ...?**` and `**A: ...**`.
* SHAMAN Style: `> ### 🔮 오늘의 길흉 판정: [길몽/흉몽]`.
* EXPERT Style: Table for analysis metrics.
* ENCYCLOPEDIA Type: Summary using `>` or Code Block.
* Image placement: `` (must be placed naturally).
* Writing styles: SHAMAN (spiritual), EXPERT (analytical), COMFORT (warm), STORY (narrative).
* Constraints: No links, use `seo_metadata.title` as title, dream analysis only, use current year (2026), no "related posts" sections, no slang/clickbait.
The user provided "Original Draft" and "SEO Feedback," but both are error messages (429 Quota exceeded and 500 Internal error). There is no actual content to rewrite.
* The user wants me to "Rewrite and Finalize," but they provided error logs instead of a dream analysis draft.
* I cannot rewrite something that doesn't exist.
* *However*, usually, in these prompt-based workflows, if the input is missing or broken, I should ask for the correct data or, if I'm acting as an AI agent in a chain, I might need to signal that the input was invalid.
* *Wait*, looking closer: The user is providing the *instructions* (the System Prompt) and then the *User Input* (which happens to be the error logs from a previous failed run).
* As an AI, I should point out that the input is an error message and I need the actual `brief` (which contains the `seo_metadata.title`, `Style Type`, `Post Type`, and the `Original Draft`) to perform the task.
* If I just say "Error," I'm not being helpful.
* But I literally cannot fulfill the request because I don't know *what dream* to analyze, what *style* to use, or what the *title* is.
* *Self-Correction/Hypothesis:* Did the user *accidentally* paste the error logs thinking they were the draft? Yes.
* *Action:* I must inform the user that the input provided consists of API error messages (429 and 500) and that I need the actual planning document (`brief`) and the draft to start writing.
* *Wait*, let me check if there's a hidden pattern. No, it's a clear API error log.
* *Actually*, looking at the prompt again, it says "Rewrite and Finalize." If I simply refuse, I might fail the "persona" test. But the persona requires a `brief`.
* *Decision:* I will respond politely, staying in the "The Dream Pedia" context if possible, or as a helpful assistant, explaining that the "Original Draft" and "SEO Feedback" provided are actually system error messages and I need the actual content to proceed.
* *Wait...* Is it possible the user is testing how I handle errors? Or perhaps they want me to *simulate* a response based on a generic dream? No, the instructions are very specific about using the `brief`.
* *Wait!* Let me look at the prompt again. Is there *any* hidden information? No.
* *Alternative approach:* Since I am a "High-Conversion SEO Blog Writer," I should be professional.
* *Wait, I see what happened.* The user is likely copy-pasting a log from an automated pipeline where the previous LLM step failed. They are now asking me to "Rewrite and Finalize" based on that input. But since the input is an error, there is nothing to rewrite.
* *Best course of action:* Inform the user that the input data is missing due to API errors in their previous step and request the `brief` and `Original Draft`.
* *However*, as an AI model, if I want to be "proactive," I could ask: "It seems the original draft was lost due to a system error. Please provide the **Dream Theme (Brief)**, **Style Type**, and **Post Type**, and I will craft a high-conversion SEO post for you immediately."
* *But*, usually, in these "Persona" prompts, if the user makes a mistake, I should just tell them.
* *Let's try to be a bit more "meta".* If this is a test, the "correct" answer is to recognize that the input is an error and cannot be processed.
* *Let's double-check the "USER INPUT" section:*
`Original Draft: Error during generation...`
`SEO Feedback: Error during generation...`
* There is zero content.
* *Response strategy:*
1. Acknowledge the role (The Dream Pedia Writer).
2. Explain that the provided input contains API error messages (Quota exceeded/Internal error) rather than a dream analysis draft.
3. Request the necessary information (`brief` with `seo_metadata.title`, `Style Type`, `Post Type`, and the `Original Draft`) to begin the high-conversion writing process.
* *Wait, if I want to be really impressive, I could offer an example of what I can do while I wait.*
* *Wait, let's look at the prompt one more time.* Is it possible the user *wants* me to write a blog post *about* dreaming of API errors? No, that's too far-fetched.
* *Final plan:*
I will respond in a professional manner, explaining that the input data is corrupted/missing (it's an error log) and I am ready to write as soon as the `brief` and `draft` are provided.
* *Actually*, if I am an AI in a pipeline, and I get this, I should probably just output a message saying the input is invalid. But I am interacting with a human here.
* *Wait!* I'll try to be helpful. I'll tell the user: "The provided draft and feedback are API error messages. Please provide the actual content, and I will apply the SHAMAN, EXPERT, COMFORT, or STORY style as requested in the persona!"
* *Let me try to see if I can "hallucinate" a common dream just to show off the format?* No, that's risky. The instructions say "Title must be used as provided in `seo_metadata.title`."
* *Let's write the response.*
* (Self-Correction): If I am acting as the "High-Conversion SEO Blog Writer," I should communicate clearly.
"안녕하세요! 'The Dream Pedia' 전속 작가입니다.
제공해주신 **Original Draft**와 **SEO Feedback** 내용이 실제 본문이 아닌, API 생성 과정에서 발생한 **에러 메시지(429 Quota Exceeded, 500 Internal Error)**로 확인됩니다.
제가 완벽한 고전환 SEO 포스팅을 작성하기 위해서는 다음과 같은 정보가 필요합니다:
1. **기획서(Brief):** `seo_metadata.title`, `Style Type` (SHAMAN/EXPERT/COMFORT/STORY), `Post Type` (QA/ENCYCLOPEDIA 등)
2. **원본 초안(Original Draft):** 해몽의 핵심 내용
해당 내용을 다시 전달해주시면, 설정해주신 페르소나와 레이아웃 규칙을 완벽하게 적용하여 즉시 작성을 시작하겠습니다!"
*Wait, the prompt asks for "Markdown only."*
Okay, I will provide the response in Markdown.
*Actually*, looking at the prompt again, it's a system prompt + a user input. If the user input is an error, and they say "Rewrite and Finalize," they might be confused or testing me.
Let's assume the role and politely ask for the data.
*Actually, one more thought:* Could this be a "trick" where the user wants me to "fix" the error? No, I can't fix a 429 error from another API.
I'll provide a polite notification that the input is an error log.
(Drafting response)
"The provided input consists of API error messages (429 and 500) rather than a dream analysis draft. Please provide the **Brief** (including `seo_metadata.title`, `Style Type`, and `Post Type`) and the **Original Draft** so I can begin the high-conversion writing process according to the layout rules."
*Wait, I should probably answer in the language of the prompt (Korean/English mix).*
Let's go.
제공해주신 Original Draft와 SEO Feedback 내용이 실제 해몽 초안이 아니라, API 호출 과정에서 발생한 **시스템 에러 메시지