What this resource does
Core uses
This toolkit combines a research runtime with many separate Skills
The official README currently describes 79 bundled Skills in the ARIS toolkit. This page highlights representative Skills for literature work, experiment review, and paper preparation.
The component cards below cover representative academic workflows, not the toolkit's full bundled inventory.
Literature, ideas, and research planning
Provides Skills for literature surveys, novelty checks, idea generation, research refinement, proposals, and experiment plans.
- Typical inputs
- Research topic, existing literature, and constraints; Desired workflow and effort level
- Output
- Research plans, literature findings, idea evaluations, or proposal drafts from the selected Skill.
Best for: Moving from a topic to an evidence-aware research plan.
Components for this task
research-lit
Searches and analyses research papers, finds related work, and summarises key ideas across configured sources.
Suggested requests (4)
Find papers on [research topic] and summarise the most relevant related work.
Search my Zotero library for papers on [research topic].
Search the web and Semantic Scholar for papers on [research topic].
Find papers on [research topic] and download up to 10 relevant arXiv PDFs.
Adapted into complete requests from official trigger wording in skills/research-lit/SKILL.md, Source Selection examples.
idea-creator
Generates and ranks research ideas from a broad direction.
Suggested requests (3)
Brainstorm research ideas for [research direction].
Generate and rank research ideas for [research direction].
Suggest research ideas we could work on within [research area].
Adapted into complete requests from official trigger wording in skills/idea-creator/SKILL.md frontmatter.
experiment-plan
Turns a research proposal or method idea into a claim-driven experiment roadmap and evaluation plan.
Suggested requests (5)
Create a detailed experiment plan for [research proposal or method].
Design an ablation matrix for [method or model].
Define an evaluation protocol for [research claim or method].
Set the run order for experiments testing [research claim].
Estimate the compute budget for this experiment plan: [paste plan].
Adapted into complete requests from official trigger wording in skills/experiment-plan/SKILL.md frontmatter.
grant-proposal
Drafts a structured grant proposal from research ideas and literature for documented funding formats.
Suggested requests (4)
Draft a grant proposal for [research direction] using [grant format].
Turn these research ideas and literature notes into a structured grant proposal.
Draft a KAKENHI proposal for [research project].
Draft an NSF proposal for [research project].
Adapted into complete requests from official trigger wording in skills/grant-proposal/SKILL.md frontmatter.
Experiments and evidence checks
Includes workflows for experiment execution, result-to-claim mapping, integrity auditing, statistical review, and reproducibility checks.
- Typical inputs
- Code, experiment configuration, results, and logs; Selected audit or execution mode
- Output
- Experiment artifacts, audit findings, claim mappings, or corrective actions.
Best for: Checking whether computational results support the intended research claims.
Components for this task
experiment-audit
Audits experiment integrity before results are used to support research claims.
Suggested requests (2)
Check the integrity of the experiments in [experiment directory or results path].
Audit these experiment results for integrity problems before I use them to support a claim: [results].
Adapted into complete requests from official trigger wording in skills/experiment-audit/SKILL.md frontmatter.
result-to-claim
Compares completed experiment results with intended claims and identifies missing evidence or next actions.
Paper preparation and review
Contains Skills for paper planning, writing, figures, slides, citations, reviewer simulation, rebuttals, and improvement loops.
- Typical inputs
- Research evidence, manuscript files, citations, or reviewer comments; Target venue and selected paper workflow
- Output
- Draft paper sections, figures, presentations, audit reports, or revision responses.
Best for: Preparing and checking a computational research paper before or after review.
Components for this task
paper-write
Drafts LaTeX paper sections from an existing outline or paper plan.
Suggested requests (2)
Draft a LaTeX paper section from this outline: [paste outline].
Write the [section name] section for [target venue] using these research materials: [materials].
Adapted into complete requests from official trigger wording in skills/paper-write/SKILL.md frontmatter.
paper-figure
Generates data-driven plots, comparison tables, and multi-panel figures from experiment results.
Suggested requests (2)
Generate paper figures and tables from these experiment results: [data or path].
Create the paper figures described in this figure plan: [plan].
Adapted into complete requests from official trigger wording in skills/paper-figure/SKILL.md frontmatter.
citation-audit
Checks whether bibliography entries are real, correctly attributed, and used in a supporting context.
Suggested requests (3)
Check every citation in [paper directory or bibliography file].
Audit this bibliography for fabricated or incorrect references: [file or entries].
Verify whether these references support the claims that cite them: [paper or citations].
Adapted into complete requests from official trigger wording in skills/citation-audit/SKILL.md frontmatter.
rebuttal
Parses external reviews and drafts a grounded, text-only response under documented venue limits.
Suggested requests (4)
Draft a rebuttal to these reviewer comments: [paste comments].
Prepare a point-by-point reply to reviewers using [paper and review bundle].
Draft an ICML rebuttal from these reviews: [reviews].
Draft separate OpenReview responses for these reviewer threads: [reviews].
Adapted into complete requests from official trigger wording in skills/rebuttal/SKILL.md frontmatter.
Use boundaries
Limits and checks
Automated local changes
A workflow can change code, manuscripts, or generated artifacts.
Use version control, review the installer, and inspect diffs before accepting changes.
Model and service costs
Subscriptions, API fees, or compute charges may apply.
Choose routes explicitly and check provider terms and budgets.
Research validity
A polished artifact may still be scientifically invalid.
Verify experiments, statistics, sources, and claims independently.
More boundaries
- ARIS is not designed to replace domain methods, human supervision, or every non-computational stage of a research project.
- Automated workflows cannot validate unavailable datasets, private services, or experiments that were not actually run.
- Novelty, statistical validity, citation support, and publication readiness require independent expert judgement.
Technical details
- Resource type
- Toolkit
- Author or maintainer
- wanshuiyin
- Version
- v0.4.22
- Latest release
- v0.4.22
- Source last updated
- 13 Jul 2026
- Last verified
- 13 Jul 2026
- Documented applications
- Claude Code, Codex, Cursor, OpenClaw, Trae, Google Antigravity, GitHub Copilot CLI, ARIS-Code
- Documented AI models
- Claude Opus 4.8, GPT-5.6-Sol, DeepSeek
- Licence
- MIT
- Access
- Publicly available
- Additional costs
- Platform terms or usage limits may apply. API usage fees may apply for selected components. External services, software, compute, or data access may have separate costs.
- Skill instruction language
- English
- Documentation language
- English
- Repository languages
- Python, Markdown
- Dependencies
- Git; Claude Code or Codex for the documented Skill routes; Optional model CLIs, APIs, Python packages, and compute services
- Review status
- Not tested
Continue exploring