How to use the Systems Paper Writing agent skill

What an agent skill is and what this one does

An agent skill is a folder of instructions and reference files that an AI agent loads when a task matches the skill's description. This one is named systems-paper-writing and is built from SKILL.md plus five reference files covering section blueprints, writing patterns, venue checklists, reviewer guidelines and systems conferences.

Once installed, the agent can allocate page budgets to each section, suggest paragraph-by-paragraph outlines, and check a draft against venue-specific checklists and reviewer expectations. The skill gives structure, not technical content: the research results and domain expertise still have to come from you.

Install the skill from the ZIP file

Download the ZIP from this page and unzip it into one of two locations. For a personal skill available in all your projects, unzip it into ~/.claude/skills/. For a skill that applies only to one project, unzip it into that project's .claude/skills/ folder.

If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.

• Personal: ~/.claude/skills/
• Single project: .claude/skills/ inside the project folder
• Claude apps: Settings → Capabilities → Skills

Trigger the skill with natural requests

You do not need a special command. The skill's description covers requests about systems conference paper writing, so ask in plain language and name the venue or section you care about. Mentioning the venue helps, because the skill has venue-specific checklists for OSDI, SOSP, ASPLOS, NSDI and EuroSys.

If you ask about a machine learning venue such as NeurIPS, ICML or ICLR, the skill itself points you to a different skill, ml-paper-writing.

• Plan a 12-page OSDI paper on my storage system with a page budget per section.
• Give me a paragraph-by-paragraph outline for the evaluation section of my NSDI submission.
• Check this draft against the SOSP checklist and tell me which expected sections are missing.
• Review my introduction: does it state the target environment and application and the gap in prior systems?

What the skill covers and what it leaves out

The page plan allocates roughly 0.25 page to the abstract, 1.5 to 2 pages to the introduction, 3 to 4 pages to design, and 3 to 4 pages to evaluation, with the rest split across background, implementation, related work and conclusion. It also recommends stating each experimental conclusion three times: as a hypothesis, as a closing conclusion, and in the figure caption.

The skill explicitly delegates citation verification and LaTeX template setup to ml-paper-writing. It does not verify your references or produce a working template.

• Included: section page budgets, paragraph blueprints, writing patterns, pre-submission checklist, reviewer criteria, conference deadlines
• Not included: citation verification, LaTeX template setup, the research results themselves

Safety notes: scripts and network access

According to the skill's listing, it does not run scripts and does not use the network. It needs no listed requirements. Its contents are text and reference files, which you can open and read before installing.

Because the agent's output can be generic prose, review anything it writes with a domain expert before it goes into a submission.

• Runs scripts: no
• Uses network: no
• Requires: nothing listed

Tips for getting useful output

Give the agent your draft, your target venue and your page limit, and say which section you want checked. Ask for one section at a time, since the blueprints are per section. Treat the outline as a starting structure and replace the placeholders with your own numbers and results.

• Name the venue and page limit (submission vs camera-ready) in your request.
• Ask for the checklist review for one stage at a time.
• Keep your technical claims and data in your own notes; the skill does not supply them.