meta-research

Deepdive skill for Claude Code — 12-phase research pipeline: plan-review gate, parallel sub-agent search, claims-ledger triangulation with dissent protection, relevance × authority evidence filter, multi-angle red team, four-layer citation verification. 105 blocks, 29 channels, 460+ stat sources, 47 APIs, 1072 verified endpoints.

Automatically modelling and distilling knowledge within AI. In other words, summarising the AI research firehose.

Automated assessment of transparency across the biomedical literature.

This is a repository that contains all code and data needed to reproduce the results of a meta-research study looking at code (and data) sharing in ecology journals with and without code-sharing policies

This research project is part of a PhD by Lee Jones and will examine statistical quality in health and biomedical research by focusing on linear regression models. The goal of this study is to explore current statistical practice and identify common statistical misconceptions and errors made by researchers. We expect our results will highlight t…

Simulating the impact of random audits on the research world

从20个项目中系统提取可复用方法论模式的实验记录。包含10轮正式审查实证（4后端）、58项发现、G5可追溯审计。不成熟框架，诚实的实验记录。

AWARE is a decision support tool for Avoiding WAste in health REsearch. It allows users to upload studies and assess their value for decision-making. In development.

Open-source bibliometric analysis platform for medical education research. Search PubMed & OpenAlex, analyze trends, visualize networks, export PRISMA-compliant methodology logs.

paper: citation sentiment classification

Governance framework defining institutional authority boundaries and policy contracts for AI-assisted workflows.

Automated GRIM-style consistency checker on RCT baseline tables. Flags integer means that aren’t possible to mathematically add up, screened at scale across PMC trials.

like homebrew but with less fizz. install binaries as fast and as easy as possible. no package manager, no recipes or metadata to update, just install and go bin, binaries, fast, installer

An R package for automated identification and extraction of transparency indicators.

Text mining and analysis of “data available upon request” statements in scientific articles.

Every clinical trial declares what it will measure before it enrols anyone. ClinicalTrials.gov keeps every edit to that promise, dated. This checks the registry for edits that landed after the data was already in.

USENIX 2025 paper artifacts

Data and analysis code for a cross-sectional study of generative-AI transparency statements in 97,916 medical journal articles, 2015-2026 (Research Integrity and Peer Review)

AI-assisted meta-research repository by David Osipov. Synthesizing scientific literature on environmental health, AI ethics, cybersecurity, and more, with human validation. Licensed under CC-BY-4.0. Uses SPDX and automated SBOM

Could these celiac trials have detected their drugs? TAK-101 prevented 71% of gluten injury and was written up as a failure. At 13 patients per arm it needed 95%. Endpoint noise is not constant: SD = 0.40 + 0.30 × injury.