One command completes: read paper → generate interpretation → mint card. Supports multiple papers in parallel.
Mode
Force NATIVE mode. This workflow is a pure skill pipeline (ljg-paper → ljg-card) and does not require the Algorithm's seven-step process. Directly call the skills according to the execution steps below; do not go through OBSERVE/THINK/PLAN/BUILD/EXECUTE/VERIFY/LEARN.
Arguments
| Arguments | Description |
|---|---|
| No arguments | Paper links/files already provided in the conversation |
-l |
Use the long-image mode for the card template (default -c comic) |
-i |
Use infographic mode for the card template |
Execution
1. Collect the paper list
Extract all paper sources from the user message (arXiv URL, PDF path, paper title, etc.).
2. Process each paper in parallel
For each paper, launch an Agent subagent; each subagent executes two steps in order:
Step A — Read paper (ljg-paper):
Call the Skill tool to execute ljg-paper, passing in the source of that paper. Wait for completion and obtain the path of the generated org file.
Step B — Mint card (ljg-card):
Read the org file generated in Step A, call the Skill tool to execute ljg-card(default -c, or the template argument specified by the user), using the org file contents as input. Wait for completion and obtain the PNG file path.
3. Summarize and report
After all papers are processed, summarize and output:
════ 論文流完成 ═══════════════════════
📄 {論文標題1}
📝 解讀: {org 檔案路徑}
🖼️ 卡片: {PNG 檔案路徑}
📄 {論文標題2}
📝 解讀: {org 檔案路徑}
🖼️ 卡片: {PNG 檔案路徑}
...
Key constraints
- The two steps for each paper must be sequential (paper first, then card), but multiple papers run in parallel
- The respective quality standards, red lines, and taste guidelines of ljg-paper and ljg-card remain unchanged
- The card content comes from the generated org file, not the original paper