cumcm-research-intelligence
Skill4 starsChanged 3 months agoMITView on GitHub
Search and organize recent papers, public datasets, formula sources, benchmark metrics, and source provenance for CUMCM or mathematical-modeling tasks. Use when the user asks for latest literature, data lookup, 8+ references, dataset citations, formula sources, research background, benchmark baselines, or a pre-contest dossier for a modeling topic.
In this repository8 other files it gives its agents.
- cumcm-data-loopSkill
- cumcm-figureSkill
- cumcm-format-packagerSkill
- cumcm-model-plannerSkill
- cumcm-paper-writingSkill
- cumcm-problem-readerSkill
- cumcm-reviewerSkill
- cumcm-sharedSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Search papers with scripts/search_papers.py unless the user names a different source.
- Very specific: line 15
Search datasets with scripts/search_data.py when external data may be useful or when the problem asks for real-world supplements.
- Distinctive: line 47
Do not cite a paper only because the title sounds relevant.
- Distinctive: line 48
Do not invent DOI, author names, dataset coverage, licenses, or formulas.
- Distinctive: line 49
Do not browse or summarize live problem-discussion pages during contest mode.
- Distinctive: line 50
Do not let the final paper include external data without source URL, access date, and processing trace.
≈ “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 Research Intelligence
This skill builds a source-grounded evidence base for mathematical modeling. It is for literature, data, formulas, and provenance. It does not solve contest problems from online discussions.
Workflow
- Load
../cumcm-shared/references/competition-ethics.mdand decide whether the task is pre-contest, post-contest, or in-contest. - Convert the topic into bilingual search terms: Chinese domain terms, English academic terms, method terms, metric terms, and data-source terms.
- Search papers with
scripts/search_papers.pyunless the user names a different source. Require at least 8 credible references when external literature supports the paper. - Search datasets with
scripts/search_data.pywhen external data may be useful or when the problem asks for real-world supplements. - For every formula, metric, or benchmark, record the source, exact role, and whether it is copied, adapted, or only conceptually similar.
- Return a literature/data matrix, not a loose bibliography.
- Hand off to
cumcm-model-plannerwith only source-supported modeling implications.
Output Contract
Return these sections by default:
Scope and rule mode
- Contest mode:
- Search terms:
- Source boundary:
Literature matrix
| ID | Claim/model use | Title | Year | Source | DOI/URL | Evidence strength | Notes |
Dataset matrix
| ID | Use | Publisher | Coverage | URL | License/terms | Access date | Caveats |
Formula and metric ledger
| Formula/metric | Source | How it will be used | Citation needed in paper |
Modeling implications
- Candidate algorithms:
- Validation metrics:
- Risks:Red Lines
- Do not cite a paper only because the title sounds relevant. Check abstract or metadata relevance.
- Do not invent DOI, author names, dataset coverage, licenses, or formulas.
- Do not browse or summarize live problem-discussion pages during contest mode.
- Do not let the final paper include external data without source URL, access date, and processing trace.
Resources
| Resource | Use |
scripts/search_papers.py | Query OpenAlex, Semantic Scholar, Crossref, and arXiv, merge candidates, and emit Markdown/JSON. |
scripts/search_data.py | Query public dataset catalogs and emit a provenance table. |
references/source-evaluation.md | Source-quality grading and citation/data-use rules. |
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.

