Agent Skills by echo-ice
Showing 28 vetted skills indexed across 1 GitHub repositories.
matlab-model-code-generator
Generate MATLAB or Beita Tianyuan code from validated method plans and cleaned data.
robustness-checker
Automate robustness, sensitivity, error, and baseline comparison checks for modeling contest solutions.
model-code-analyzer
Translate mathematical modeling plans into Python or MATLAB code logic and layout.
consistency-auditor
Audit numbers, file names, symbols, and parameters across documents, code, and results.
model-assumptions-builder
Extract and organize model assumptions from problem parses and method pools.
data-auditor-cleaner
Audit and clean raw contest data for mathematical modeling readiness.
workflow-orchestrator
Manage mathematical modeling contest workflows from problem parsing to paper writing.
python-model-code-generator
Generate executable Python modeling code from validated method plans and cleaned data.
figure-table-planner
Plan figures and tables for mathematical modeling papers by audience and purpose.
paper-polisher
Refine mathematical modeling contest paper sections in Markdown or LaTeX drafts.
python-code-reviewer
Review Python code for method plan compliance and traceable outputs.
matlab-code-reviewer
Review MATLAB and Beita Tianyuan modeling code for runnability and compatibility.
symbol-table-builder
Generate a conflict-free global symbol table from problem parses and candidate method pools.
decision-prompt-builder
Generate tailored decision prompts framed as trade-offs at modeling gates.
problem-parser
Parse mathematical modeling problems into structured specifications with goals, constraints, and outputs.
problem-classifier
Classify mathematical modeling subproblems into standard types and output a structured report.
math-figure-generator
Generate publication-quality mathematical figures with matplotlib and predefined layouts.
method-selector
Compare and recommend candidate modeling schemes for mathematical modeling subproblems.
modeler-decision-logger
Log human modeling decisions into an append-only log with validation.
result-report-generator
Generate structured experiment reports comparing candidate methods from model outputs.
related-paper-analyzer
Analyze user-provided papers for method selection cues without fabricating references.
completeness-auditor
Verify completed skills have audit artifacts with at least five pass items.
code-reviewer
Route Python and MATLAB code to language-specific reviewers.
solution-package-builder
Integrate modeler's method explanations and programmer's result analyses into one document.