- Home
- Skills
- Huangwb8
- Chineseresearchlatex
- Nsfc Schematic
nsfc-schematic_skill
- Python
1.3k
GitHub Stars
4
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill huangwb8/chineseresearchlatex --skill nsfc-schematic- CHANGELOG.md34.4 KB
- config.yaml9.4 KB
- README.md10.1 KB
- SKILL.md23.0 KB
Overview
This skill generates editable and publication-ready schematic/mechanism diagrams for NSFC proposals and research documents. It produces a draw.io source (.drawio) plus vector/raster renderings (.svg, .png, optional .pdf) and supports multi-round optimization with traceable artifacts. Use it when you need to convert research mechanisms, algorithm architectures, or module relationships into embeddable, editable figures for proposals or papers.
How this skill works
The tool accepts a structured spec, a LaTeX proposal file, or a proposal directory and parses module/node/connectivity information to auto-complete layout and generate draw.io XML. It runs prechecks, renders SVG/PNG/PDF, evaluates visual and structural quality across multiple dimensions, and records per-round evidence so the best iteration can be exported. Outputs are organized into isolated run directories and a hidden workspace folder containing specs, optimization reports, and evaluation artifacts.
When to use it
- You need an editable schematic (.drawio) for embedding into a proposal or paper.
- You want to convert research mechanism descriptions or algorithm/module relations into a clear figure.
- You require multi-round visual optimization with objective evaluations (readability, overlap, routing).
- You plan to review and adjust a generated plan/spec before final rendering (planning mode).
- You need reproducible artifacts and audit trail for figure iterations and decisions.
Best practices
- Provide a structured spec_file when possible for the most controllable output.
- Use planning mode (context or plan script) to review node names, connections, and layout before generation.
- Keep personal data out of node labels; include only research-relevant terms.
- Set reasonable rounds for optimization (e.g., 3–5) to balance quality and turnaround.
- Check evaluation reports in the hidden workspace to understand and fix flagged issues (font overflow, overlaps).
Example use cases
- Turn a methods section or LaTeX proposal into an editable schematic for grant submission.
- Generate a modular algorithm architecture diagram for inclusion in a conference paper.
- Produce a mechanism diagram showing interactions between system components for reviewers.
- Iteratively refine a figure across several rounds to satisfy readability and layout constraints.
- Export vector graphics (.svg) from the draw.io source for high-quality document embedding.
FAQ
Provide at least one of: a structured spec_file, a single proposal file (main.tex), or a proposal directory; specs give the most reliable control.
Will the skill include personal data from my proposal?
By default the tool treats proposals as sensitive and only preserves research-relevant terms; avoid embedding unrelated personal identifiers in your files.
What happens if draw.io CLI is missing?
The pipeline falls back to an internal renderer with a strong notice; you can require CLI availability in config to force a hard failure.