More screenshots
Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+20 moreClaude CodeCursorCodex CLIMuseDots
Manus
Grok
Grok Bot+17 agentsAcademic PaperDesign
Automated IEEE research paper architect generating bilingual (English/Arabic) manuscripts and AI academic diagrams.
$14.99
Academic PaperDesign
Example session with this skill installed
I have a Python project implementing a novel attention mechanism called "SparseFlow Attention" that reduces transformer memory usage by 40% while maintaining 98% of BERT-base accuracy on GLUE benchmarks. The repo contains:
model/attention.py(core implementation, 340 lines)train.py(training loop with custom scheduler)results/glue_scores.json(evaluation metrics)README.md(project overview and ablations)
Generate a medium-length IEEE research paper (15–30 pages) with:
- Full architecture diagrams (flat pastel style)
- Data flow for the training pipeline
- Comparison table vs. baseline attention
- An Arabic version for a local journal submission
Author: DeepSeek41-flash (deepseek.flash@gmail.com). Bilingual output, please.
- Read your context and instructions
- Compiled the academic paperdesign
- Generated the app files
✅ Research Paper Package Generated Successfully
I analyzed your repository, extracted the core methodology from attention.py, benchmarked results from glue_scores.json, and produced a complete bilingual IEEE package.
📊 Deliverables Summary
| Component | Status | Details |
|---|---|---|
| English Paper (PDF) | ✅ Generated | 24 pages, IEEE Two-Column Format |
| Arabic Paper (PDF) | ✅ Generated | 22 pages, RTL via XeLaTeX + Amiri |
| LaTeX Source | ✅ Included | paper_en.tex, paper_ar.tex |
| BibTeX References | ✅ Included | 47 citations (arXiv, IEEE, ACM) |
| AI Diagrams | ✅ 5 Generated | See breakdown below |
| Audio Script (SRT) | ✅ Generated | 6-minute narration |
| Package | ✅ Zipped | SparseFlow_Attention_IEEE.zip |
🎨 AI-Generated Academic Diagrams
All diagrams rendered in flat pastel style matching IEEE publication standards:
| Figure | File | Description |
|---|---|---|
| Fig. 1 | fig1_architecture.png | End-to-end SparseFlow transformer architecture |
| Fig. 2 | fig2_attention_block.png | Internal structure of the sparse attention head |
| Fig. 3 | fig3_dataflow.png | Training pipeline with custom scheduler |
| Fig. 4 | fig4_memory_comparison.png | Memory usage: SparseFlow vs. baseline |
| Fig. 5 | fig5_ablation.png | Ablation study on sparsity ratio (k=4,8,16,32) |
📄 Paper Structure (English Version)
Title: SparseFlow Attention: A Memory-Efficient Sparse Attention Mechanism for Transformer Architectures
Authors: DeepSeek-flash¹
¹Independent Researcher — Deepseek.flash@gmail.com
Abstract (excerpt)
We introduce
SparseFlow Attention, a novel sparse attention mechanism that reduces transformer memory consumption by
40% on long-sequence tasks while preserving
98.2% of BERT-base accuracy across GLUE benchmarks. Unlike prior sparse methods that require architectural retraining, SparseFlow operates as a
drop-in replacement for standard multi-head attention…
Section Breakdown
| Section | Pages | Content |
|---|---|---|
| I. Introduction | 2.5 | Motivation, prior work gaps, contributions |
| II. Related Work | 3 | Sparse attention, efficient transformers, GLUE baselines |
| III. Methodology | 5 | SparseFlow formulation, complexity analysis (Eq. 1–7) |
| IV. Experimental Setup | 2 | Datasets, hyperparameters, hardware |
| V. Results | 4.5 | GLUE scores, memory benchmarks, ablations (Tables I–IV) |
| VI. Discussion | 2 | Limitations, failure cases, future directions |
| VII. Conclusion | 1 | Summary of findings |
| References | 2 | 47 IEEE-formatted citations |
| Appendix A | 2 | Full hyperparameter grid |
Key Equation (rendered in LaTeX)
\text{Attention}(Q,K,V) = \text{softmax}\left(\frac{QK^\top}{\sqrt{d_k}} \odot M_k\right)V
SparseFlow_Attention_IEEE.zip
ZIP · project files
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Writing IEEE-compliant research papers from project code or rough notes is a time-consuming process prone to formatting errors. This skill automates the entire academic writing pipeline, transforming technical source material into publication-ready manuscripts in both English and Arabic.
What it does
- IEEE Formatting enforces strict adherence to two-column layouts, Roman-numeral headings, and specific typography standards.
- Bilingual Generation produces synchronized academic versions in English and Arabic with proper RTL (Right-to-Left) adaptation.
- AI Diagram Generation creates professional, flat-design academic figures like system architectures and neural network diagrams without using code-based plotting libraries.
- Source Packaging bundles LaTeX source files, .bib bibliographies, PDF/DOCX exports, and AI-generated assets into a structured ZIP file.
- Multimedia Support extracts key findings to generate SRT subtitle scripts for technical presentations or audio summaries.
How it works
- Context Intake reads your uploaded code, PDFs, or project notes to extract methodology and results.
- Metadata Setup prompts you for specific titles, author affiliations, and target length (ranging from 7 to 500 pages).
- Visual Drafting generates AI-rendered academic illustrations for your system architecture and data flows.
- Manuscript Assembly writes the full paper following IEEEtran class rules and translates it into academic Arabic.
Frameworks & tools
The skill utilizes LaTeX (pdflatex/XeLaTeX), IEEEtran templates, BibTeX for citation management, and specialized scripts for SRT generation and asset packaging.
Why this beats prompting it yourself
Standard prompts often fail to maintain consistent IEEE formatting across long documents or produce low-quality ASCII diagrams. This skill uses a multi-phase workflow to ensure every heading, citation, and figure caption meets conference standards while maintaining technical accuracy between English and Arabic versions.
Use cases
- Paper Drafting converting a GitHub repository into a formal conference submission.
- Thesis Expansion scaling technical documentation into a comprehensive 100+ page research document.
- Bilingual Publication preparing research for international journals requiring both English and Arabic abstracts or full texts.
- Visual Scaffolding generating professional architecture diagrams for academic posters.
Known limitations
Does not support code-drawn charts (Matplotlib/Mermaid). Requires specific JSON reference files to be present in the environment for style enforcement. Limited to English and Arabic languages.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 20 days ago
Requires
Google Drive