AGENT SKILL

reproducibility-statement

Prepare the reproducibility, transparency and open-science parts of a paper, including data and code availability statements, reproducibility checklists (NeurIPS, ICML, ICLR, ACL Responsible NLP, Nature reporting summaries), research artifact packaging (Zenodo DOI, CITATION.cff, environment lockfiles, seeds), ethics/broader-impact statements, CRediT author contributions, competing interests and AI-use disclosures. Use when a venue requires any of these statements or checklists, or when preparing code/data for release alongside a paper.

License 许可
MIT
最近更新
2026-11-08
Stars
★ 0
Forks
0

01 前言 Front Matter

name
reproducibility-statement
license
MIT
metadata
{ "tags": "reproducibility, data availability, code availability, open science, FAIR, checklist, CRediT, Zenodo, ethics statement", "author": "Kalaris Labs (Sayan Chowdhury)", "version": "1.0", "category": "research-writing", "maintainer": "Kalaris Labs" }
description
Prepare the reproducibility, transparency and open-science parts of a paper, including data and code availability statements, reproducibility checklists (NeurIPS, ICML, ICLR, ACL Responsible NLP, Nature reporting summaries), research artifact packaging (Zenodo DOI, CITATION.cff, environment lockfiles, seeds), ethics/broader-impact statements, CRediT author contributions, competing interests and AI-use disclosures. Use when a venue requires any of these statements or checklists, or when preparing code/data for release alongside a paper.
compatibility
No runtime dependencies.

03 SKILL.md 正文

Reproducibility and Transparency Statements

Workflow

  1. List every statement and checklist the venue requires (section 1).
  2. Deposit data and code, and obtain persistent identifiers before writing the statements.
  3. Draft each statement with concrete locations, licenses and access conditions.
  4. Verify every link and accession while logged out. Never claim availability that does not exist yet.
  5. Answer checklists honestly, with a justification for each "No".

Editors increasingly desk-reject for missing or vague statements, and reviewers check them. Treat each as a precise, verifiable claim, never boilerplate.

1. Find out what the venue requires

Check the current author guidelines and checklist for the target venue (the venue-templates and journal-format skills point to them). Common requirements:

Venue type Typical requirement
ML conferences (NeurIPS, ICML, ICLR) Paper checklist in the PDF (reproducibility, compute, licenses, broader impact, LLM usage), code encouraged
ACL / ARR venues Responsible NLP checklist, limitations section (mandatory), ethics statement
Nature Portfolio Data and code availability statements, reporting summary, source data for figures
PLOS Mandatory data availability statement; data must be available without restriction unless legally exempt
Clinical journals Trial registration, data sharing statement (ICMJE), reporting guideline checklist (CONSORT, STROBE, ...)
Cell Press STAR Methods with Key Resources Table and resource availability section

2. Data availability statement

State where, how and under what conditions, with persistent identifiers:

The processed data supporting the findings are available at Zenodo (https://doi.org/10.5281/zenodo.XXXXXXX) under CC BY 4.0. Raw sequencing reads are deposited in the SRA under BioProject PRJNAXXXXXX. Patient-level clinical data cannot be shared publicly due to [regulation]. Requests can be made to [data access committee contact], and access requires a data use agreement.

  • Use domain repositories where they exist (GEO/SRA/ENA, PDB, PRIDE, OpenNeuro, Dryad, ICPSR, PANGAEA). Otherwise use general ones (Zenodo, Figshare, OSF, Dataverse).
  • "Available upon reasonable request" alone is increasingly rejected. Explain the restriction and the access route.
  • Reserve DOIs before submission (Zenodo supports reserved DOIs). Use private reviewer links if the venue allows.

3. Code availability and artifact packaging

Checklist for the released repository:

  • [ ] Tagged release matching the paper version, archived to Zenodo (GitHub integration mints a DOI)
  • [ ] README with installation, a minimal example, and the exact commands that reproduce each table/figure
  • [ ] Pinned environment: uv.lock/requirements.txt with versions, environment.yml, or a container image digest
  • [ ] Random seeds, number of runs, and hardware noted, with variance reported (mean ± std over N seeds)
  • [ ] Data download/preprocessing scripts (never ship data you're not licensed to redistribute)
  • [ ] Pretrained weights or checkpoints with checksums, if central to the claims
  • [ ] LICENSE (OSI-approved for code; CC BY for text/data), and third-party licenses respected
  • [ ] CITATION.cff so GitHub shows "Cite this repository"
  • [ ] Anonymized for double-blind review (anonymous GitHub mirrors), de-anonymized for camera-ready

4. Compute and environmental reporting (ML and simulation)

Report hardware (GPU/CPU type and count), total compute (GPU-hours) for main experiments and the full project, including hyperparameter search. Report software versions of key frameworks.

5. Other statements

  • Author contributions (CRediT): map each author to the 14 roles (Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, Funding acquisition).
  • Competing interests: declare financial and non-financial interests, or state "none".
  • Funding: funder names and grant numbers exactly as the funder requires.
  • Ethics: IRB/ethics committee name and approval number, consent procedures, animal welfare approvals, dual-use considerations.
  • Use of AI tools: follow the venue's policy. Most require disclosure of generative AI use in writing or analysis and forbid listing AI as an author. Describe what was used and how outputs were verified.
  • Limitations / broader impact: concrete failure modes, populations or settings where results may not hold, potential misuse, and mitigations.

Rules

  • Every statement must be true at submission time. Verify each link and accession works (logged out, private window).
  • Don't claim "fully reproducible" unless someone other than the author has rerun it from the README.
  • Answer checklists honestly. "No" with a justification beats an unjustified "Yes".

Related skills

ml-paper-writing, scientific-writing, venue-templates, datalad, lamindb, research-grants (data management plans).

04 同仓库更多 Skills

Tashima-Tarsh
Tashima-Tarsh/Disha6.6

youtube-full

Use when the user needs YouTube transcripts, video search, channel browsing, playlist extraction, or content monitoring. Trigger phrases: 'get the transcript for', 'search YouTube for', 'what are the latest videos on', 'list this playlist', 'monitor this channel', or any request involving a YouTube URL, video ID, or @handle. Do NOT use for downloading video or audio files, YouTube engagement data (likes, comments), or private/age-restricted videos.

图像与多媒体
★ 0
Tashima-Tarsh
Tashima-Tarsh/Disha6.6

x-twitter-growth

X/Twitter growth engine for building audience, crafting viral content, and analyzing engagement. Use when the user wants to grow on X/Twitter, write tweets or threads, analyze their X profile, research competitors on X, plan a posting strategy, or optimize engagement. Complements social-content (generic multi-platform) with X-specific depth: algorithm mechanics, thread engineering, reply strategy, profile optimization, and competitive intelligence via web search.

★ 0
Tashima-Tarsh
Tashima-Tarsh/Disha6.6

write-a-skill

Create new agent skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, build, or author a new skill.

Skill与智能体开发
★ 0
Tashima-Tarsh
Tashima-Tarsh/Disha6.6

wrap-up

Close out a launched Claude Managed Agent — recap every primitive the founder now owns, regenerate the single-file overview page, and suggest the next 1-2 upgrades. Use when the user says "wrap up", "close this out", "what do I own now", "give me the summary", "recap the agent", or when the orchestrator routes phase=wrap-up. primitives_inventory.py tables everything owned (agent, environment, session, memory, outcome, deployment); overview_page.py regenerates a self-contained ./my-agent/agent-overview.html; upgrade_suggester.py ranks the next moves from recorded deferrals plus standing hardening steps. Companion to run-without-you; the last stop before phase=done.

★ 0
Tashima-Tarsh
Tashima-Tarsh/Disha6.6

workflow-builder

Design and write deterministic multi-agent workflow scripts (.js files in .claude/workflows/) for Claude Code's Workflow tool. Use when a user wants to build, create, author, scaffold, or run a custom Claude Code workflow, orchestrate sub-agents (fan-out, pipeline, loop, judge-panel), or automate a repeatable multi-step task across fresh-context agents.

Skill与智能体开发
★ 0
Tashima-Tarsh
Tashima-Tarsh/Disha6.6

weekly-review

Use when someone wants to run a weekly review, close open loops, audit stalled projects and commitments, get their system back to trusted, restart a lapsed review habit, or says "/cs:weekly-review". Walks David Allen's three-phase loop — GET CLEAR, GET CURRENT, GET CREATIVE — with deterministic scripts that inventory open loops, gate the checklist with named gaps, and score commitment health 0-100.

★ 0