Observed arrival · 2026-10-11
Reproducible Science ties research claims to their evidence
A set of command-line tools records and checks links between research plans, experiment runs, paper claims, and cited passages.
- For
- Researchers tracking claims and experiment outputs
- Worth noticing
- It hashes inputs before a run and lets a claim point to a run ID and manuscript location, such as “Table 2.”
Field notes
The workflow starts by freezing a preregistration, then hashes inputs before a run and records outputs against a run ID. Researchers can attach manuscript claims to those runs and locations, while citation checks test quoted passages against cited sources; `repro verify` checks the chain and `repro reproduce` reruns declared commands. The project explicitly limits its claim: it checks links between claims and artifacts, rather than deciding whether a paper is reproducible.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue