What problem does it solve?
Writing an NSFC (National Natural Science Foundation of China) grant proposal in LaTeX involves risks like fabricated or mismatched citations, inconsistent abbreviations, stiff prose, unbalanced section lengths, and broken argumentation. This Skill audits a proposal without modifying any source files and produces a standardized, evidence-backed QC report graded P0/P1/P2.
Core Features & Use Cases
- Citation Verification: Deterministically checks bibkey existence and BibTeX field completeness, then fetches metadata from Crossref/arXiv/Unpaywall with URL accessibility and title matching to flag fake or miscited references.
- Multi-Threaded Independent QC: Runs several parallel threads over the same checklist (style, citations, length, structure, logic, abbreviations, terminology) and aggregates deduplicated findings into a fixed report template.
- Deterministic Prechecks: Scans for straight-quote typography issues, render-order-aware abbreviation first-use/definition conflicts, terminology inconsistency, and per-file length distribution.
- Use Case: Before submitting an NSFC Young Scientists Fund proposal, run the Skill on the project directory to receive a report listing missing bibkeys as P0, conflicting abbreviation definitions as P1, and optional style optimizations as P2, plus a minimal fix roadmap.
Quick Start
Ask the AI to use nsfc-qc to run a read-only quality control check on your NSFC proposal directory with 5 threads and output a standardized P0/P1/P2 QC report.