Embodied AI Paper Writer
A portable agent skill (SKILL.md + reference playbooks) for writing top-conference embodied-AI papers, distilled from 63 papers (CoRL/RSS/ICRA/IROS/Science Robotics…
Install
npx skills add OpenGHz/embodied-ai-paper-writerInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →Embodied AI Paper Writer
English | 简体中文
A portable agent skill (SKILL.md + reference playbooks) that coaches the writing craft of embodied-AI papers — distilled from 63 top-conference papers across CoRL, RSS, ICRA, IROS, and Science Robotics (2022–2026). Works with any LLM agent that can load markdown context.
What this is
A drop-in agent skill that turns any modern LLM agent — Claude Code, Cursor, Continue, Cline, Aider, the Anthropic / OpenAI SDKs, or anything else that can load a system prompt — into a corpus-tuned coach for writing robotics / embodied-AI papers. It teaches:
- Title patterns, abstract moves, intro arcs.
- Method / Related Work organization.
- Experiments setup framing, results paragraph rhythm, ablation narration.
- Figure roles, caption templates, table conventions.
- Conclusion / Limitations / Future Work / Appendix.
- Section openers, pivots, connectors, the contribution-restatement spiral.
- A phrasebook of openers, hedges, and anti-patterns.
It does not decide which experiments to run, validate technical claims, suggest research directions, translate, or run your LaTeX build. See SKILL.md for the full boundary list.
Why a skill, not a textbook
Most writing advice is generic ("be clear, cite related work"). This skill is quantified and corpus-tuned:
- "Abstract = 120–250 words, 5 moves (Frame → Gap → Contribution → Method → Results)."
- "Method tense = present, system-as-subject."
- "Plot caption = mean ± StdErr + sample size, always."
- "Contribution noun phrase repeats 5–7× verbatim across the paper."
- "F1 teaser caption = 3–6 sentences; F3 hardware caption = 1 sentence."
Each rule traces back to observed patterns in the 63-paper corpus.
Quick start
1. Install
npx skills add OpenGHz/embodied-ai-paper-writer
Follow the prompts to choose your agents and installation scope.
2. Invoke it
Once loaded, the agent will engage on phrases like:
- "Help me write the abstract for my CoRL submission."
- "Caption this figure — it's a 3-panel success-rate plot."
- "My Limitations section sounds defensive — fix it."
- "Review my paper's arc."
- 「帮我润色一下这段 Intro」
- 「rebuttal 怎么写」
See the full trigger list in SKILL.md (frontmatter description).
Project layout
embodied-ai-paper-writer/
├── SKILL.md # Operating manual (entry point)
├── references/
│ ├── titles.md # Title patterns and architectures
│ ├── abstract-intro-playbook.md # Abstract + Intro construction
│ ├── method-relatedwork-playbook.md # Method + Related Work
│ ├── experiments-results-playbook.md # Experiments, Results, Ablations
│ ├── figures-tables-playbook.md # F1–F8 figure roles, caption templates
│ ├── language-phrasebank.md # Rhetorical phrasebook (A–K)
│ ├── flow-transitions.md # 6-move paper arc, pivots, connectors
│ ├── closing-appendix-playbook.md # Conclusion / Limitations / Appendix
│ └── research/ # Raw extraction (read for traceability)
│ ├── _paper_roster.md # 63-paper corpus index
│ └── 00–08 + topical extracts
├── LICENSE
├── CONTRIBUTING.md
├── CODE_OF_CONDUCT.md
├── SECURITY.md
├── SUPPORT.md
├── CHANGELOG.md
└── CITATION.cff
The skill has two layers:
Similar plugs
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local…
Why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.
Production-grade engineering skills for AI coding agents.
The 1st Collaborative AI Design Agent Workspace. Local-first and open source. Create prototypes, websites, apps, dashboards, presentations, images, and videos with your team…