{"id":49921124,"url":"https://github.com/aniketkrs/research-paper","last_synced_at":"2026-05-23T02:01:00.104Z","repository":{"id":358310075,"uuid":"1240887511","full_name":"aniketkrs/research-paper","owner":"aniketkrs","description":"Three complementary agent skills for academic research: WRITES papers, FINDS papers on a topic, READS any paper as a visual experience (mind maps, flowcharts, plain-English). Runtime-neutral, works with 50+ agents. Install: npx skills add aniketkrs/research-paper","archived":false,"fork":false,"pushed_at":"2026-05-16T19:46:36.000Z","size":608,"stargazers_count":3,"open_issues_count":1,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-22T07:59:54.907Z","etag":null,"topics":["academic-writing","agent-skills","anthropic","apa","arxiv","citations","claude-code","claude-skill","harvard","ieee","literature-review","multi-agent","research-paper","thesis","whitepaper"],"latest_commit_sha":null,"homepage":"https://github.com/aniketkrs/research-paper","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/aniketkrs.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-05-16T17:37:30.000Z","updated_at":"2026-05-20T13:38:46.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/aniketkrs/research-paper","commit_stats":null,"previous_names":["aniketkrs/research-paper"],"tags_count":7,"template":false,"template_full_name":null,"purl":"pkg:github/aniketkrs/research-paper","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aniketkrs%2Fresearch-paper","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aniketkrs%2Fresearch-paper/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aniketkrs%2Fresearch-paper/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aniketkrs%2Fresearch-paper/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/aniketkrs","download_url":"https://codeload.github.com/aniketkrs/research-paper/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aniketkrs%2Fresearch-paper/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33379721,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-23T01:21:08.577Z","status":"online","status_checked_at":"2026-05-23T02:00:05.530Z","response_time":53,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["academic-writing","agent-skills","anthropic","apa","arxiv","citations","claude-code","claude-skill","harvard","ieee","literature-review","multi-agent","research-paper","thesis","whitepaper"],"created_at":"2026-05-16T20:02:37.125Z","updated_at":"2026-05-23T02:01:00.068Z","avatar_url":"https://github.com/aniketkrs.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# research-paper\n\n\u003e Three agent skills for academic research, in one repo. **Find** real papers,\n\u003e **read** any paper as a visual experience, and **write** new papers grounded\n\u003e in real sources. Runtime-neutral — works with 50+ AI coding agents.\n\n```bash\nnpx skills add aniketkrs/research-paper\n```\n\nAfter install, in any compatible agent session:\n\n```\n/research \"graph neural networks for fraud detection\" --style ieee\n/find-paper \"retrieval-augmented generation\" --years last-3\n/read-paper https://arxiv.org/abs/1706.03762\n```\n\n\u003e **Date freshness:** every paper run starts by checking today's actual\n\u003e date (via `date -u +%Y-%m-%d` or runtime context). Year-range flags\n\u003e like `--years last-3` are computed from today, not from the model's\n\u003e training cutoff. The skill never silently uses stale data — see\n\u003e [`instructions/freshness.md`](skills/research-paper/instructions/freshness.md)\n\u003e in any of the three skills.\n\n[![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)\n[![Skills: 3](https://img.shields.io/badge/skills-3-green.svg)](#the-three-skills)\n[![Tests: 116/116](https://img.shields.io/badge/tests-116%2F116-brightgreen.svg)](tests/)\n\n---\n\n## What this is\n\nThree small, focused skills that work together — or independently — to\nmake academic-paper work fast, rigorous, and not boring.\n\n| | Skill | One-line job |\n|---|---|---|\n| ✍️ | **`research-paper`** | **Writes** new papers — research papers, lit reviews, theses, whitepapers, surveys, policy briefs |\n| 🔎 | **`get-research-paper`** | **Finds** real existing papers — searches arXiv, Scholar, PubMed, Semantic Scholar |\n| 📖 | **`read-research-paper`** | **Reads** any paper (URL / arXiv / DOI / PDF / DOCX / PPTX / image) and renders it visually |\n\nAll three are runtime-neutral. They install with one command and work\nwith **Claude Code, OpenCode, Cursor, Cline, Codex, Aider, Amp,\nAntigravity, AiderDesk, Augment, IBM Bob,** and 50+ other agent\nruntimes via [`npx skills`](https://www.npmjs.com/package/skills).\n\n---\n\n## Install\n\n### One command — all three skills, no prompts\n\n```bash\nnpx skills add aniketkrs/research-paper --yes --skill '*'\n```\n\nThe `--yes --skill '*'` flags install all three skills in one shot\nwithout prompting. The installer auto-detects every agent runtime on\nyour machine and places the skills in the universal `.agents/skills/`\ndirectory.\n\n### Verify\n\n```bash\nnpx skills list\n```\n\nExpected: **`research-paper`, `get-research-paper`, `read-research-paper`** all listed.\n\n### Other install options\n\n```bash\n# Install only one skill\nnpx skills add aniketkrs/research-paper --yes --skill research-paper\n\n# Pin to a version\nnpx skills add aniketkrs/research-paper#v2.4.0 --yes --skill '*'\n\n# User-scope (global) instead of project-scope\nnpx skills add aniketkrs/research-paper --yes --skill '*' --global\n\n# Direct from GitHub (no skills CLI; this installer also installs all skills by default)\nnpx -y github:aniketkrs/research-paper install\n```\n\nFor per-platform manual install (Claude Desktop, claude.ai web,\nAnthropic API/SDK, etc.), see\n**[INSTALLATION.md](INSTALLATION.md)**.\n\n---\n\n## The three skills\n\n### ✍️ research-paper — writes new papers\n\nProduces publication-ready research papers, literature reviews,\ntheses, whitepapers, surveys, and policy briefs.\n\n**Trigger it:**\n```\n/research \"topic\" --style ieee --depth comprehensive\n/literature-review \"topic\" --systematic --sources 50\n/whitepaper \"topic\" --audience technical\n/thesis \"Chapter 3: Methodology\" --style harvard\n/policy \"topic\" --depth standard\n```\n\n…or in plain English: *\"write a research paper on graph neural networks\nfor fraud detection\"*.\n\n**Produces:**\n- 10 paper formats: arXiv, IEEE, ACM, Nature, Harvard, lit-review,\n  thesis chapter, whitepaper, survey, policy brief.\n- 7 citation styles: Harvard, APA, IEEE, MLA, Chicago, Nature,\n  arXiv-numeric. Switchable with one flag.\n- Real visualizations: charts, tables, heatmaps, flowcharts, PRISMA\n  diagrams, forest plots.\n- Methodology section with sample-size justification and validity\n  threats.\n- Statistical validation with effect sizes, CIs, multiple-comparison\n  correction.\n- Three-persona simulated peer review (methodologist + domain expert\n  + reader).\n- Plain-English summary alongside the technical content.\n\n**What it won't do:** invent citations, DOIs, or coauthors.\nUnverifiable items are flagged `[UNVERIFIED]` and surfaced in\n`Known-gaps.md`.\n\n### 🔎 get-research-paper — finds papers on a topic\n\nSearches arXiv, Google Scholar, PubMed, Semantic Scholar, DBLP,\nACM DL, IEEE Xplore, and OpenReview. Returns a curated reading list\nwith verified DOIs, key findings, and ready-to-cite metadata.\n\n**Trigger it:**\n```\n/get-research-paper \"topic\" --n 25 --years 2020-2024 --depth deep\n/find-paper \"topic\"\n/papers-on \"topic\"\n/scholar \"topic\"\n```\n\n…or: *\"find research papers on retrieval-augmented generation\"*.\n\n**Produces:**\n- A ranked reading list (`reading-list.md`) with quality scores.\n- A canonical `bibliography.yaml` ready to feed the writer skill.\n- A 1–3 paragraph field briefing.\n- Source-quality scoring (authority + rigor + recency).\n- Diversity heuristics: per-author cap, per-venue cap, ≥1 review,\n  ≥1 foundational paper.\n- DOI verification + retraction screening (when web tools are\n  available).\n\n### 📖 read-research-paper — reads any paper, makes it not boring\n\nTake any paper input — arXiv URL, arXiv ID, DOI, PDF, DOCX, PPTX,\nimage, plain text — and render it as a visual reading experience.\n\n**Trigger it:**\n```\n/read-research-paper https://arxiv.org/abs/1706.03762\n/read-paper ./paper.pdf\n/explain-paper https://doi.org/10.1145/3589334\n/visualize-paper ./slides.pptx\n/tldr-paper ./scan.png\n```\n\n…or: *\"read this research paper [URL]\"*, *\"explain this paper\"*,\n*\"make this paper visual\"*.\n\n**Produces a multi-layer Markdown rendering** with:\n- One-page infographic at the top (mind map + headline numbers).\n- TL;DR (5–8 sentences).\n- Plain-English summary (5–10 sentences).\n- Section-by-section walk-through with plain-English **alongside**\n  the technical content (not replacing it).\n- Method flowchart (Mermaid).\n- Key-findings infographic (matplotlib when available, Markdown\n  otherwise).\n- Comparison table to baselines.\n- Related-work timeline.\n- \"Why this matters\" footer.\n- Verification trail.\n\n**Three-tier \"don't bluff\" cascade:** local cache → live fetch →\nbundled corpus → model knowledge with `[UNVERIFIED]` flags. Source\ntier is always declared in the output footer.\n\n---\n\n## Supported file formats\n\n### Read (input)\n\nThe `read-research-paper` skill handles **any** of these as input:\n\n| Format | Extensions | Always available? |\n|---|---|---|\n| Markdown, plain text, JSON | `.md`, `.markdown`, `.txt`, `.json` | ✅ |\n| LaTeX | `.tex`, `.latex` | ✅ |\n| HTML | `.html`, `.htm` | ✅ (with regex fallback) |\n| CSV / TSV | `.csv`, `.tsv` | ✅ (basic) |\n| RTF | `.rtf` | ✅ (with regex fallback) |\n| **PDF** | `.pdf` | install `pdfplumber` or `pypdf` |\n| **DOCX** | `.docx` | install `python-docx` |\n| **PPTX** | `.pptx` | install `python-pptx` |\n| **XLSX** | `.xlsx`, `.xls` | install `pandas` + `openpyxl` |\n| **EPUB** | `.epub` | install `ebooklib` |\n| **Images** (OCR) | `.png`, `.jpg`, `.tiff`, `.bmp` | install `pytesseract` + Tesseract |\n\nOne command to enable everything:\n\n```bash\npip install pdfplumber python-docx python-pptx pandas openpyxl \\\n            beautifulsoup4 striprtf ebooklib pytesseract Pillow\n```\n\n### Write (output)\n\nThe `research-paper` skill produces Markdown by default. To convert\nto other formats:\n\n| Format | Extension | Renderer |\n|---|---|---|\n| Markdown | `.md` | native (always) |\n| HTML | `.html` | Pandoc |\n| DOCX | `.docx` | Pandoc |\n| LaTeX | `.tex` | Pandoc |\n| **PDF** | `.pdf` | Pandoc + LaTeX engine |\n| RTF | `.rtf` | Pandoc |\n| EPUB | `.epub` | Pandoc |\n| ODT | `.odt` | Pandoc |\n| PPTX | `.pptx` | Pandoc |\n\nInstall Pandoc (and optionally LaTeX for PDFs):\n\n```bash\n# macOS\nbrew install pandoc                      # base\nbrew install --cask mactex-no-gui         # for PDF output\n\n# Linux\napt install pandoc                        # base\napt install texlive                        # for PDF output\n\n# Windows\nchoco install pandoc                       # or scoop install pandoc\nchoco install miktex                        # for PDF output\n```\n\nThen:\n\n```bash\n/research \"topic\" --output paper.pdf\n/research \"topic\" --output paper.docx\n```\n\n**Self-test** to see what's available on your machine:\n\n```bash\npython skills/read-research-paper/toolchains/read_any_file.py --self-test\npython skills/research-paper/toolchains/convert_output.py --self-test\n```\n\nFull reference: [`skills/read-research-paper/sources/file-formats.md`](skills/read-research-paper/sources/file-formats.md).\n\n---\n\n## End-to-end workflow\n\nThe three skills chain cleanly:\n\n```\n┌─────────────────────────────────────────────────────────────┐\n│   /find-paper \"topic\"                                        │\n│   └─→ reading-list.md + bibliography.yaml                    │\n└─────────────────────────────┬───────────────────────────────┘\n                              │\n                              ▼\n┌─────────────────────────────────────────────────────────────┐\n│   /read-paper \u003cURL\u003e      (visualize ANY single paper)         │\n│   └─→ paper-visual.md  (mind map + flowchart + plain English) │\n└─────────────────────────────┬───────────────────────────────┘\n                              │\n                              ▼\n┌─────────────────────────────────────────────────────────────┐\n│   /research \"topic\" --bibliography ./topic/bibliography.yaml  │\n│   └─→ paper-final.md → convert_output → paper.pdf / .docx    │\n└─────────────────────────────────────────────────────────────┘\n```\n\nOr use any one standalone — they don't require each other.\n\n---\n\n## Quick start\n\n### Find papers on a topic\n\n```\n/find-paper \"transformer architectures for time-series forecasting\" --n 10\n```\n\nThe skill returns a curated reading list with verified DOIs, key\nfindings per paper, and a `bibliography.yaml` ready for the writer\nskill.\n\n### Read a single paper visually\n\n```\n/read-paper https://arxiv.org/abs/1706.03762\n```\n\nThe skill returns a multi-layer Markdown rendering with mind map,\nflowchart, infographic, and plain-English alongside the technical\ncontent. Cached locally so re-asking is instant.\n\n### Write a paper using a curated bibliography\n\n```\n/research \"predicting customer churn with graph neural networks\" \\\n    --style ieee \\\n    --depth comprehensive \\\n    --bibliography ./gnn-fraud-detection/bibliography.yaml\n```\n\nThe skill walks: plan → literature review → methodology → analysis →\nvisualization → drafting → citations → validation → review → ship.\n\n### Convert the result to PDF / DOCX\n\n```bash\npython skills/research-paper/toolchains/convert_output.py \\\n    --input ./gnn-fraud-detection/paper-final.md \\\n    --to pdf \\\n    --out paper.pdf\n```\n\n---\n\n## What \"done\" means\n\nA paper is delivered only after passing every quality gate:\n\n- ≥ 1500 words (configurable).\n- ≥ 8 references with verified DOIs.\n- ≥ 1 figure or table.\n- All required sections present (abstract, intro, methodology,\n  results, discussion, limitations, conclusion, references).\n- Plain-English summary present.\n- Reproducibility statement present.\n- Future Work section present.\n- Academic-quality rubric mean ≥ 4 / 5.\n- All three reviewer personas score ≥ 3.0.\n- All `[CITATION NEEDED]` and `[UNVERIFIED]` flags resolved or\n  surfaced in `Known-gaps.md`.\n\nFailures surface in `Known-gaps.md`. Never silent.\n\n---\n\n## Compatibility\n\n`npx skills add` auto-detects and installs into 50+ agent runtimes:\n\n| Universal | Symlink-supported |\n|---|---|\n| Amp, Antigravity, Cline, Codex, Cursor, +10 more | AiderDesk, Augment, IBM Bob, Claude Code, OpenCode, +35 more |\n\nRun `npx skills add aniketkrs/research-paper --list` to see the full\nlist of agents detected on your machine.\n\n---\n\n## Repo structure\n\n```\nresearch-paper/\n├── README.md, LICENSE, CHANGELOG.md, INSTALLATION.md, package.json\n├── bin/install.js               ← direct npx installer\n├── tests/test-runner.js         ← 116/116 tests pass\n├── docs/                        ← architecture, design decisions, FAQ\n└── skills/\n    ├── research-paper/          ← writes papers (106 files)\n    ├── get-research-paper/      ← finds papers (20 files)\n    └── read-research-paper/     ← reads any paper (26 files)\n```\n\nEach skill is self-contained — manifest, instructions, prompts,\ntemplates, schemas, toolchains, examples, sources, tests.\n\n---\n\n## Optional Python toolchain\n\nThe skills work with **just** filesystem read/write. They emit\nMarkdown tables and Mermaid diagrams as the always-available default.\n\nTo enable real charts, statistical validation, multi-format file I/O,\nand Pandoc output:\n\n```bash\n# Charts + statistical validation (research-paper)\npip install pandas numpy scipy statsmodels matplotlib seaborn pyyaml\n\n# Multi-format file reading (read-research-paper)\npip install pdfplumber python-docx python-pptx openpyxl \\\n            beautifulsoup4 striprtf ebooklib pytesseract Pillow\n\n# Multi-format output (research-paper) - install Pandoc + LaTeX system-wide\n# macOS:    brew install pandoc \u0026\u0026 brew install --cask mactex-no-gui\n# Linux:    apt install pandoc texlive\n# Windows:  choco install pandoc miktex\n```\n\nWhen any tool is missing, the skill detects it and falls back —\nnever silent failure, always a clear \"install X to enable\" message.\n\n---\n\n## Extending\n\nThe skills are modular. Each one is a folder with a clear extension\nboundary.\n\n- **Add a new venue** (e.g., LNCS): drop a template in\n  `templates/\u003cvenue\u003e.md` + register in `manifest.json`.\n- **Add a new citation style**: extend\n  `citation_engine/citation-styles.md` and\n  `toolchains/format_bibliography.py`.\n- **Add a new chart type**: extend\n  `visualization_engine/decision-engine.md` and\n  `toolchains/generate_charts.py`.\n- **Add a new file format**: extend\n  `read-research-paper/toolchains/read_any_file.py` and\n  `read-research-paper/sources/file-formats.md`.\n\nFull extension guide: [`docs/extending.md`](docs/extending.md).\n\n---\n\n## Tests\n\nLightweight test suite in `tests/test-runner.js`:\n\n```bash\nnode tests/test-runner.js\n```\n\nCovers:\n- File structure (every required file present).\n- SKILL.md frontmatter (valid YAML).\n- manifest.json (valid JSON, required fields).\n- JSON Schemas (paper-schema, citation-schema, visual-paper schema, etc.).\n- Python toolchain self-tests (graceful degradation).\n- Citation pipeline smoke test against fixtures.\n\n**116/116 tests pass.**\n\n---\n\n## Troubleshooting\n\n| Symptom | Likely cause | Fix |\n|---|---|---|\n| Skill never activates | Restart your agent session after install | |\n| Charts come out as Markdown only | Python plotting deps missing | `pip install pandas matplotlib seaborn` |\n| Can't read a `.pdf` | PDF library missing | `pip install pdfplumber` |\n| Can't read a `.docx` | DOCX library missing | `pip install python-docx` |\n| Can't OCR an image | OCR not installed | `pip install pytesseract Pillow` + install Tesseract |\n| Can't convert to PDF | Pandoc / LaTeX missing | install Pandoc + a LaTeX engine |\n| Skill activates for unrelated requests | Trigger patterns too broad | edit `manifest.json → trigger.patterns` |\n| Output truncated mid-section | Context pressure | switch to multi-file output (`--depth comprehensive`) |\n\n---\n\n## License\n\nMIT — see [LICENSE](LICENSE).\n\n## Contributing\n\nIssues and PRs welcome at\n[github.com/aniketkrs/research-paper](https://github.com/aniketkrs/research-paper).\n\nWhen filing an issue, please include:\n- Your runtime (Claude Code, OpenCode, etc.)\n- The slash command or prompt that triggered the issue\n- The contents of `Known-gaps.md` if any\n- The relevant `validation-report.md` if any\n\n---\n\n## Versioning\n\n| Version | Highlights |\n|---|---|\n| **2.3.0** | Multi-format file I/O: read PDF/DOCX/PPTX/XLSX/EPUB/images; write PDF/DOCX/HTML/LaTeX/EPUB/RTF/ODT/PPTX |\n| 2.2.0 | Added `read-research-paper` skill (visual paper reading) |\n| 2.1.0 | Added `get-research-paper` skill (paper discovery) |\n| 2.0.x | Initial release of `research-paper` skill (paper writing) |\n\nSee [CHANGELOG.md](CHANGELOG.md) for the full history.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Faniketkrs%2Fresearch-paper","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Faniketkrs%2Fresearch-paper","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Faniketkrs%2Fresearch-paper/lists"}