What problem does it solve? Writing grant applications, thesis proposals, and research plans often mixes unverified claims with real evidence, drifts from template requirements, and leaves logical gaps between research questions, objectives, and methods. This Skill enforces a staged workflow that separates confirmed facts from AI suggestions, keeps the document aligned with the official template, and actively revises drafts after simulated review. ## Core Features & Use Cases - Information Layering: Maintains three distinct tiers—basic task parameters, verifiable facts, and AI suggestions—so unconfirmed content never enters the final text as fact; missing data is marked with 【待补充：...】 placeholders instead of being fabricated. - Template Matching: Ships prebuilt templates for NSFC general and youth programs and the National Social Science Fund (including the anonymous review leaflet), and prioritizes user-provided institutional templates. - Staged Checkpoints & Review: Confirms the task brief, fact ledger, literature positioning, logic framework, and draft at each stage, then runs a simulated review covering statistics, citations, duplication, and template compliance, actively fixing the draft before delivery. - Use Case: A researcher with published papers and preliminary data asks for an NSFC youth fund application; the Skill builds a fact ledger from their materials, supplements verifiable recent literature, designs the question-objective-content-evidence chain, and produces PROPOSAL.md plus references.bib with placeholders for anything not yet provided. ## Quick Start Help me write an NSFC youth fund application based on my published papers and preliminary experiment data, following the official template.
