| name | storage-systems-paper-writer |
| description | Write storage, database, and distributed-systems papers, design notes, evaluation narratives, and rebuttals from real implementation and evidence. Use when turning a system or prototype into a workshop or conference paper while keeping claims bounded to what the artifact and experiments actually support. |
Storage Systems Paper Writer
Overview
Use this skill to write evidence-bounded systems papers.
The writing should clarify the system boundary and supported claim, not inflate novelty.
Workflow
- State the paper type and strongest claim.
- Build the outline around that claim.
- Bind each section to specific evidence.
- Write the system and evaluation sections before polishing the introduction.
- Keep non-claims explicit.
- Use related work to sharpen the boundary, not to spray citations.
Default Paper Structure
Default sections:
- problem and why the current boundary is bad
- core observation
- system model and boundary
- design and protocol
- implementation status
- evaluation
- limitations and non-claims
- related work
Use references/section-contracts.md for section-by-section contracts.
Writing Rules
- Do not claim a new consensus protocol unless you actually changed consensus.
- Do not package engineering cleanup as novelty.
- Do not claim generality beyond the tested boundary.
- Do not promise production readiness unless the operational evidence exists.
Evidence Binding
Every major paragraph should answer one of:
- what exact problem exists
- what boundary changed
- what mechanism enforces the property
- what evidence supports the claim
- what remains outside the claim
Output Modes
Use this skill for:
- paper outline
- abstract and introduction
- design section rewrite
- evaluation narrative
- related-work positioning
- rebuttal drafting
- long-form design note that may later become a paper
Reference Loading
Load only what you need:
references/section-contracts.md for paper-section writing contracts
references/venue-targeting.md for top venue fit and framing differences across databases, systems, and protocol work