⚡AgentSkills
🔬 Scientific Research · Reproducibility

Make computational results reproducible

Environment pinning, deterministic pipelines, and one-command reruns from raw data.

intermediate~40 minResearchersPhD StudentsPostdocs

Steps

  1. 1Pin environments: lockfiles plus container image recorded per result
  2. 2Version control everything including parameter files and seeds
  3. 3Structure pipeline raw → processed → figures with explicit stages
  4. 4Never edit intermediate files manually; regenerate instead
  5. 5Test one-command rerun on clean machine before submission
  6. 6Publish code/data under DOIs via archival repositories

Common Pitfalls

  • ▲'It works on my laptop' dependency webs
  • ▲Figures hand-tweaked after generation

Commands

Install with skills CLI
$ npx skills add aniruddhaadak80/skills --skill reproducibility-computational-reproducibility
Install globally
$ npx skills add aniruddhaadak80/skills --skill reproducibility-computational-reproducibility -g

Tags

#reproducibility#pipelines#research-code#scientific-research

Related skills

Daily entries linking intent, method versions, observations, and next actions.

🔬 Scientific Research·~25m