| name | writing-qec-papers |
| description | Provides guidance for structuring, evidence-checking, and drafting QEC manuscripts so that benchmark, formalization, and systems claims stay aligned with explicit artifacts and reproducible workflows. |
| version | 0.1.0 |
| author | QEC Research Skills |
| license | MIT |
| tags | ["QEC","Writing","Papers","Reproducibility"] |
| dependencies | ["qec-research"] |
Writing QEC Papers
What This Skill Is For
Use this skill when a task is about turning QEC work into a manuscript:
- frame the contribution,
- map claims to evidence,
- build a paper outline,
- keep methods, artifacts, and limitations explicit.
When To Use It Versus Alternatives
Use this skill for general manuscript planning and drafting across QEC topics such as:
- decoder benchmarks,
- threshold or memory experiments,
- formalization results,
- software and tooling papers,
- architecture or workflow papers.
Use tailoring-qec-papers-for-physical-review-and-quantum when the task is specifically about venue fit, positioning, and adapting the narrative for Physical Review journals or Quantum.
Do not use this skill as a substitute for checking current journal author instructions. Treat it as a workflow guide, not a frozen policy manual.
Required Artifacts And Assumptions
Expected inputs:
- a candidate contribution statement,
- the main artifacts or results that support it,
- the intended audience or venue family if known.
Expected outputs:
- a manuscript plan,
- a manuscript outline,
- a claim-to-evidence matrix,
- a citation bundle,
- a list of missing figures, tables, experiments, or proof details.
Assumptions to make explicit:
- which code family, noise model, and decoder assumptions define the story,
- whether the paper is theory-heavy, benchmark-heavy, software-heavy, or mixed,
- whether reproducibility depends on registry artifacts, local scripts, or external toolchains.
Standard Workflow Checklist
- Read
research-state.yaml and confirm the current project scope and backend assumptions.
- Classify the paper's main contribution type before drafting prose.
- Write one sentence for the main claim and one sentence for the strongest supporting evidence.
- Fill the claim-evidence matrix in
templates/claim-evidence-matrix.template.md before expanding the abstract or introduction.
- Fill the manuscript plan in
templates/manuscript-plan.template.yaml so title, claims, venues, reproducibility anchors, and blockers are recorded in one place.
- Start a source inventory with
templates/citation-bundle.template.yaml before the related-work section turns into a memory exercise.
- Build the outline around evidence-bearing sections instead of generic paper headings alone.
- Keep figures, tables, proofs, and artifact references tied to named results, not aspirational future work.
- If a target venue is known, run the venue-positioning skill before finalizing the abstract and introduction.
Validation Loop
Before closing work:
- confirm every headline claim has an explicit supporting artifact, theorem, or experiment,
- confirm QEC-specific assumptions are stated in the methods or setup plan,
- confirm the draft distinguishes results, limitations, and open problems,
- confirm any missing benchmark or proof work is surfaced as a blocker, not hidden in prose.
Common Issues And Fixes
Contribution inflation
Problem: the draft sounds broader than the evidence actually supports.
Fix: rewrite the title, abstract, and contribution bullets to match the strongest validated result only.
Benchmark without research question
Problem: a paper becomes a pile of plots without a clean scientific or systems question.
Fix: state what decision the benchmark resolves and why existing baselines were insufficient.
Hidden setup assumptions
Problem: the paper omits the code family, schedule, decoder compatibility assumptions, or confidence methodology.
Fix: move those assumptions into a visible setup section and tie figures back to them explicitly.
References
