ds-write

Drafts evidence-grounded research papers and reports with LaTeX venue templates and citation integrity checks.

1.1k|116|Updated Feb 26, 2026
One-click install
npx skills add https://github.com/OpenLAIR/dr-claw --skill ds-write
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ds-write
Source: https://github.com/OpenLAIR/dr-claw/tree/main/skills/ds-write
Command: npx skills add https://github.com/OpenLAIR/dr-claw --skill ds-write

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and assets (resource) components.

What problem does it solve?

It turns accumulated experimental evidence into a faithful research paper, report, or summary without inventing unsupported claims, enforcing outline-first drafting, claim-to-evidence mapping, and submission-readiness gates.

Core Features & Use Cases

  • Evidence-led drafting: Assembles baselines, run artifacts, and analysis results into a claim-evidence map before any prose is written, and routes back to experiments when gaps appear.
  • Paper contract management: Maintains a selected outline, paper experiment matrix, evidence ledger, and section-level files so drafting never outruns the actual evidence.
  • Venue templates and review pipeline: Ships LaTeX templates for ICLR, ICML, NeurIPS, ACL, AAAI, COLM, OSDI, NSDI, ASPLOS, and SOSP, plus harsh self-review, proofing, and submission-checklist stages.
  • Use Case: After an experiment campaign finishes, use it to select an outline, bind every claim to artifacts, draft sections in the ICLR 2026 template, run a five-pass revision, and produce a compiled paper bundle.

Quick Start

Use the ds-write skill to draft a research paper from my accepted experiment results using the ICLR 2026 template.

Frequently Asked Questions about ds-write

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a research paper from experiment results with AI?

Provide an accepted baseline and at least one meaningful result; the skill assembles the evidence, builds a claim-evidence map and outline, then drafts sections inside a venue LaTeX template. If evidence is missing, it reports the gap or routes back to experiments instead of inventing content.

What LaTeX conference templates are included for paper writing?

Bundled templates cover ICLR 2026, ICML 2026, NeurIPS 2025, COLM 2025, AAAI 2026, ACL, ASPLOS 2027, NSDI 2027, OSDI 2026, and SOSP 2026. General ML work defaults to the ICLR 2026 template unless a stronger venue target is specified.

When should I not use this paper writing skill?

Avoid it when the quest lacks a credible evidence base, when baseline establishment or ideation is still the main work, or when the need is follow-up analysis rather than narrative consolidation. It refuses to polish prose over missing evidence.

How does the skill prevent unsupported claims in a draft?

Every claim must link to durable artifacts through a claim-evidence map and evidence ledger, and a paper experiment matrix gates the experiments section. Weak results trigger explicit claim downgrades rather than inflated prose.

Does the skill handle figures and citations for papers?

Yes. It enforces muted paper-figure palettes, vector-first export, a render-inspect-revise pass per figure, and citation integrity checks with verified references, typically targeting 30 to 50 verified sources for a full paper.