| name | research-data-availability |
| description | Use when auditing or drafting thesis data availability, dataset provenance, result-data traceability, data hashes, access restrictions, data dictionaries, or final submission/defense data checks. |
Research Data Availability
Use this skill when thesis claims, figures, tables, or experiments need traceable data sources. It adapts useful nature-data ideas to a graduation-thesis workflow.
Core Rules
- Keep
docs/thesis/data-availability.md as the local data availability source of truth.
- Do not store private datasets, credentials, or restricted data in git.
- Every result claim should trace to source data, processed data, script/notebook, and output artifact.
- Use
DATA-* for datasets/snapshots and connect them to CLM-*, EXP-*, and FIG-* records according to evidence-promotion-policy.md.
- Use
docs/thesis/material-passport.md for evidence material identity cards when datasets, experiment outputs, figures, readings, or document artifacts become important thesis evidence.
- Record why data cannot be shared when privacy, license, advisor, or project constraints apply.
- Data availability is an audit gate, not a substitute for experiment reproducibility.
Workflow
Read references/workflow.md for the checklist and statement pattern. Read references/source-map.md for provenance.
- Identify claims, figures, tables, and experiments that depend on data.
- Register each dataset as
DATA-* with version, path, access level, license/permission, and hash/manifest.
- Map each
CLM-* claim to source data, processed data, script/notebook, and artifact.
- Register evidence-critical datasets, outputs, figures, readings, and final document artifacts in
material-passport.md.
- Mark sharing status as public, private, restricted, or TBD.
- Run or recommend
scripts/audit_data_availability.py.
- Route missing experiment details to
$research-experiment-engineering and missing claim mapping to $research-results-analysis.
- Hand unresolved final issues to
$research-final-audit.
Output Contract
Always include:
- dataset records to add or update
- claim-to-data traceability gaps
- access and license/permission status
- hash or manifest status
- data dictionary status
- final availability statement draft or required fixes
- suggested updates to
docs/thesis/data-availability.md
- material passport updates when data or artifacts become final evidence