cumcm-data-loop
Skill4 starsChanged 3 months agoMITView on GitHub
Build and audit a complete data-code-result loop for CUMCM projects: raw data provenance, cleaning logs, feature definitions, model outputs, figure/table traceability, hashes, reproducible commands, and appendix support-material lists. Use when the user asks for data closure, reproducibility, data cleaning, result verification, code-output consistency, or support-material preparation.
In this repository8 other files it gives its agents.
- cumcm-figureSkill
- cumcm-format-packagerSkill
- cumcm-model-plannerSkill
- cumcm-paper-writingSkill
- cumcm-problem-readerSkill
- cumcm-research-intelligenceSkill
- cumcm-reviewerSkill
- cumcm-sharedSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 18
Create a project inventory: raw data, processed data, scripts, notebooks, outputs, figures, paper drafts, support archive.
- Very specific: line 19
Run scripts/audit_data_lineage.py --project <folder> when files exist.
- Distinctive: line 21
Check for leakage, duplicate rows, unit drift, train/test contamination, manual overwrites, and orphaned outputs.
- Distinctive: line 46
Do not overwrite raw data.
- Distinctive: line 47
Do not present manually edited result tables as reproducible outputs without disclosure.
- Distinctive: line 48
Do not leave figures without source data and generation script.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CUMCM Data Loop
This skill ensures that every number in the paper can be traced and reproduced.
Workflow
- Load
../cumcm-shared/references/data-provenance.md. - Create a project inventory: raw data, processed data, scripts, notebooks, outputs, figures, paper drafts, support archive.
- Run
scripts/audit_data_lineage.py --project <folder>when files exist. - Require a lineage ledger linking raw source -> cleaning script -> processed data -> model script -> output table -> figure/paper claim.
- Check for leakage, duplicate rows, unit drift, train/test contamination, manual overwrites, and orphaned outputs.
- Produce an appendix-ready support-material file list.
Output Contract
Data loop status
- Raw data locked:
- Processed data reproducible:
- Outputs reproducible:
- Figures linked:
- Paper claims linked:
Lineage ledger
| Artifact | Source | Script/command | Hash | Used in paper | Notes |
Blocking issues
- [P0] ...
Appendix support-material list
- ...Red Lines
- Do not overwrite raw data.
- Do not present manually edited result tables as reproducible outputs without disclosure.
- Do not leave figures without source data and generation script.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

