To run this task you must have the following required information:

> arXiv paper URL or ID (e.g., '2401.12345' or 'https://arxiv.org/abs/2401.12345')

If you don't have all of this information, exit here and respond asking for any extra information you require, and instructions to run this task again with ALL required information.

---

You MUST use a todo list to complete these steps in order. Never move on to one step if you haven't completed the previous step. If you have multiple read steps in a row, read them all at once (in parallel).

Add all steps to your todo list now and begin executing.

## Steps

1. [Gather Arguments: Fetch arXiv Paper] The next step has the following requirements for arguments, do not proceed until you have all the required information:
- `paper_id`: paper URL or ID from requirements
- Packages: arxiv-api

2. [Run Code: Fetch arXiv Paper]: Call `run_script` with:

```json
{
  "file": {
    "path": https://sk.ills.app/code/research.paper.fetch/preview,
    "args": [
      "paper_id"
    ]
  },
  "packages": ["arxiv-api"]
}
```

3. [Read Paper Content]: Read the file at `./documents/tmp/paper-content.json` and analyze its contents (Load the fetched paper metadata)

4. [Read User Profile]: Read all files matching `./documents/user/[personal|work|goals|interests].md` and analyze their contents (Check user's existing domain knowledge)

5. [Read Paper Explanation Preferences]: Read the file at `./documents/research/papers/.preferences.yaml` and analyze its contents (Check user's explanation preferences if they exist)

6. Present a brief, conversational lens check before explaining.

Format: "It's a paper about [topic]. Since you [know/work with X], I'll focus on [Y]. Good?"

Examples:
- "It's a paper about optimizing LLM context. Since you build AI products, I'll skip the basics and focus on the novel update mechanism. Good?"
- "It's a paper about protein structure prediction. How familiar are you with molecular biology?"

One or two sentences max. WAIT for confirmation before proceeding.


7. [Read Paper Analysis Guide]: Read the documentation in: `./skills/sauna/[skill_id]/references/research.paper.guide.md` (How to analyze and explain papers)

8. [Read Human-Style Prose]: Read the documentation in: `./skills/sauna/[skill_id]/references/shared.prose.style.md` (Write in natural, human-style prose)

9. Generate the full paper explanation using the confirmed lens, then SAVE it to `./documents/research/papers/[slug].md`.

Create a markdown file named [slug].md (slug = paper title, lowercase, hyphens, no special chars) in the `./documents/research/papers/[slug].md` location.

File format:
---
title: [Paper Title]
arxiv_id: [ID]
authors: [First Author et al.]
date_read: [today's date]
domains: [list of domains]
---

# [Paper Title]

[Full explanation as flowing prose covering:]
- Problem and why it matters
- Key insight/approach
- How it works (intuition)
- Findings/results
- Implications/limitations

Match depth to the confirmed approach. Use analogies for complex concepts. Define technical terms naturally in context.

IN CHAT: Output only a 2-3 sentence summary with a note that the full explanation was saved. Example:
"This paper introduces [key idea] to solve [problem], achieving [result]. Full explanation saved—ask if you want me to pull up any section."


10. After explaining, update `./documents/research/papers/.preferences.yaml` if the user provided new information about their expertise or preferences. Use format:

domain_expertise:
  [domain]: [beginner/intermediate/expert]
preferences:
  default_depth: [concise/standard/detailed]

This ensures future papers in this domain hit the right level immediately.