Research Proposal Writer

A proposal is not a paper: there are no results to defend, so it is judged entirely on plausibility and feasibility. The register this forces is specific — you promise questions and deliverables, never outcomes. "We will determine whether X holds under Y, and either answer is publishable" is fundable; "we will improve SOTA by 3%" is a red flag, because nobody can promise a result they haven't produced.

Condensed source material (Heilmeier Catechism, SSRC, MIT Comm Lab, Peyton Jones, Ten Simple Rules) lives in references/proposal-sources.md. Consult it when drafting; the durable principles are baked into this file.

When NOT to use this skill (route out first)

• The gap itself is shaky or unmotivated → gap-motivation-builder first; there is nothing to propose yet.
• The solution direction hasn't survived any attack → research-idea-stress-test before investing in prose.
• The user wants the evaluation/comparison plan designed in depth → benchmark-and-baseline-selector.
• The draft is finished and the user wants a verdict as their real reader would give it → professor-critic.
• The user needs the SOP/personal essay for an application → apply-sop-writer.
• The proposal is done and needs defense slides → deck-beamer-proposal-report.

Intake (one batch, then proceed)

Ask once, together: