Skip to content

Lucasuiii/modeling-workbench / .agents/skills/cumcm-workflow

cumcm-workflow

Verified

Build or resume mathematical-modeling competition work, including CUMCM, MCM/ICM, graduate and regional contests, data challenges, and open-topic statistical modeling. Guide problem framing, modeling, computation, validation, Chinese or English LaTeX writing, and reviewed PDF/source delivery using current official requirements. Not for ordinary paper polishing.

FileKindSize
SKILL.mdskill md9.0 KB
agents/openai.yamlconfig427 B
assets/latex-template/generic-ctex/macros.texother859 B
assets/latex-template/generic-ctex/main.tex.tmplother2.5 KB
assets/latex-template/generic-ctex/metadata.tex.tmplother272 B
assets/latex-template/generic-ctex/planned-section.tex.tmplother1.0 KB
assets/latex-template/generic-ctex/references.bibother155 B
assets/latex-template/generic-ctex/sections/00_abstract.texother935 B
assets/latex-template/generic-ctex/sections/98_references.texother221 B
assets/latex-template/generic-ctex/sections/99_appendix.texother264 B
assets/latex-template/generic-ctex/template.jsonconfig304 B
assets/latex-template/generic-en/macros.texother868 B
assets/latex-template/generic-en/main.tex.tmplother2.2 KB
assets/latex-template/generic-en/metadata.tex.tmplother278 B
assets/latex-template/generic-en/planned-section.tex.tmplother1013 B
assets/latex-template/generic-en/references.bibother155 B
assets/latex-template/generic-en/sections/00_abstract.texother454 B
assets/latex-template/generic-en/sections/98_references.texother201 B
assets/latex-template/generic-en/sections/99_appendix.texother240 B
assets/latex-template/generic-en/template.jsonconfig249 B
assets/latex-template/huawei-ctex/main.tex.tmplother3.3 KB
assets/latex-template/huawei-ctex/template.jsonconfig276 B
assets/plot-style/cumcm.mplstyleother953 B
references/01-intake.mdmarkdown3.9 KB
references/02-problem-analysis.mdmarkdown2.8 KB
references/03-model-design.mdmarkdown7.6 KB
references/04-computation.mdmarkdown13.4 KB
references/05-validation.mdmarkdown5.9 KB
references/06-paper-writing.mdmarkdown14.3 KB
references/07-compile-delivery.mdmarkdown6.1 KB
references/artifact-contracts.mdmarkdown3.1 KB
references/competition-adaptation.mdmarkdown6.8 KB
references/competition-writing.mdmarkdown5.6 KB
references/diagnostics.mdmarkdown4.5 KB
references/evidence-rules.mdmarkdown1.6 KB
references/handoffs.mdmarkdown3.5 KB
references/huawei-delivery.mdmarkdown10.1 KB
references/latex-template.mdmarkdown6.2 KB
references/mechanism-validation.mdmarkdown11.6 KB
references/open-topic-research.mdmarkdown4.9 KB
references/task-driven-modeling.mdmarkdown7.2 KB
schemas/claim-ledger.schema.jsonconfig5.5 KB
schemas/common.schema.jsonconfig2.5 KB
schemas/compile-receipt.schema.jsonconfig3.7 KB
schemas/cross-question-ledger.schema.jsonconfig1.3 KB
schemas/delivery-manifest.schema.jsonconfig7.9 KB
schemas/figure-manifest.schema.jsonconfig1.9 KB
schemas/handoff.schema.jsonconfig2.1 KB
schemas/independent-review-package.schema.jsonconfig3.4 KB
schemas/independent-review-result.schema.jsonconfig4.0 KB
schemas/latex-template-manifest.schema.jsonconfig2.4 KB
schemas/model-contract.schema.jsonconfig5.0 KB
schemas/paper-plan.schema.jsonconfig3.3 KB
schemas/paper-quality-report.schema.jsonconfig5.7 KB
schemas/paper-revision-log.schema.jsonconfig2.2 KB
schemas/paper-visible-text-report.schema.jsonconfig1.6 KB
schemas/problem-facts.schema.jsonconfig1.7 KB
schemas/results-index.schema.jsonconfig1.4 KB
schemas/run-manifest.schema.jsonconfig7.1 KB
schemas/source-manifest.schema.jsonconfig1.9 KB
schemas/task-capabilities.schema.jsonconfig3.9 KB
schemas/workflow-state.schema.jsonconfig2.4 KB
scripts/backend_selection.pyscript6.3 KB
scripts/build_handoff.pyscript20.4 KB
scripts/build_independent_review_package.pyscript14.9 KB
scripts/canonical_evidence.pyscript8.4 KB
scripts/compile_sources.pyscript2.4 KB
scripts/cumcm_check.pyscript2.9 KB
scripts/delivery_archives.pyscript7.0 KB
scripts/doctor.pyscript7.5 KB
scripts/index_result.pyscript10.9 KB
scripts/init_latex_paper.pyscript17.2 KB
scripts/init_project.pyscript11.1 KB
scripts/inventory_artifacts.pyscript4.1 KB
scripts/json_pointer.pyscript1.0 KB
scripts/official_materials.pyscript3.0 KB
scripts/paper_visible_text_check.pyscript5.4 KB
scripts/plan_redo.pyscript18.8 KB
scripts/project_status.pyscript5.8 KB
scripts/provenance.pyscript2.2 KB
scripts/record_compile.pyscript13.2 KB
scripts/record_decision.pyscript12.0 KB
scripts/record_run.pyscript28.1 KB
scripts/refresh_evidence.pyscript4.7 KB
scripts/set_mode.pyscript2.0 KB
scripts/submission_check.pyscript10.6 KB
scripts/workflow_checks.pyscript146.0 KB

references/mechanism-validation.md

View on GitHub
# Mechanism-triggered validation

Use only sections whose mechanism can change the current answer. Start from the official subquestion, its explicit simplifications and derived necessary conditions; a contest label or method keyword is not a trigger. These are risk prompts, not mandatory experiments or automatic model upgrades.

Put the selected falsifiable condition in the existing `verification_plan` and, when useful, `model/VALIDATION_PLAN.md`: identify the quantity, scope, units, failure condition and tolerance basis. Tolerances follow measurement resolution, numerical error and the decision margin; there is no universal cutoff. Reuse evidence only when its inputs, implementation, check definition and scope still match. Reading this guide or naming an assertion is not verification. Computable checks must inspect actual results and be written by the executed program through `--assert-file`; mathematical applicability and proof strength remain reviewer judgments. Writing and compilation consume the reviewed evidence rather than starting these experiments again.

## Continuous events and first arrival

**When:** the answer depends on a continuous trajectory crossing, touching or staying outside a boundary.

Endpoint safety does not establish interval safety. Sign-change searches can miss tangent roots or multiple crossings within one step. Check event direction, an initially satisfied condition and ordering of competing events. Refine critical intervals or use derivative/interval bounds; denser printed output alone does not improve the integrator. A finite scan without a root is not a proof of no event.

**Discriminating check:** for contact defined by `g(t) <= 0`, `g(t) = (t - 0.5)^2` on `[0, 1]` touches at `0.5` although both endpoints are positive. Check the stationary point as well as endpoints. Use problem-specific error bounds for a numerical trajectory.

**Boundary:** a task defined only at discrete observation times needs no continuous interpolation guarantee. If contact requires strict penetration, this tangent alone is not penetration.

## Inverse problems and calibration

**When:** observations determine unknown parameters or a simplified forward model is inverted.

Reconstruct the corresponding observations with the fitted parameters and original units/domain. Small residuals do not imply unique parameters: inspect scaled sensitivities/Jacobian rank, profiles or alternative parameter combinations where uniqueness matters. Shared-object measurements must use consistent shared parameters unless the task supplies state differences. Separate noise, discretization and approximation error; fitting data are not independent validation.

**Boundary:** non-identifiable parameters do not automatically invalidate an identifiable prediction. Preserve the requested approximation and report its domain; do not force uniqueness with invented bounds or replace the requested model merely because a richer fit is possible.

## Sequential sampling and reliability boundaries

**When:** simulation or sampling decides reliability, minimum capacity or a stopping threshold.

Define the probability event and sampling unit. An interval crossing the target leaves the decision undecided. Repeatedly inspecting fixed-sample intervals and stopping on success does not preserve nominal coverage. Use a prespecified validation design or a sequential guarantee whose assumptions match; validate an optimized candidate independently or account for selection. Check monotonicity before binary search and distinguish feasible, infeasible and unresolved boundary candidates when the budget ends.

**Boundary:** a truly fixed validation sample does not require sequential machinery. Shared random numbers can help paired comparisons, but identical repeats are not independent sample size. Neither a universal sample count nor a forced pass/fail verdict is justified.

## Discrete optimization and exported plans

**When:** the answer is a schedule, allocation, route, layout or rounded/postprocessed optimization result.

Read the final exported plan and independently recompute its objective and applicable resource, timing, integrality and terminal constraints from the original definition. Solver status describes the internal solve: distinguish incumbent feasibility, valid bound, stopping reason and claimed optimality. A timeout is not infeasibility, and successful exit is not global optimality. Report violations with the affected object, time/resource, observed value and requirement.

**Discriminating check:** the relaxation `x = y = 0.6` satisfies `x + y <= 1.2`; rounding each variable to one violates it. Replay the exported integers, not the retained relaxation. A repaired feasible plan still needs its own objective and optimality scope.

**Boundary:** do not impose integrality on genuinely divisible decisions, full service where omissions carry a permitted penalty, or rest/setup rules absent from the task. Back-to-back half-open intervals are compatible when no additional gap is required.

## Objective priority and available information

**When:** objectives have an explicit priority or questions/policies have different information sets.

Map the current question to variables, objective order and observations available at each decision. Lexicographic optimization is not an arbitrary weighted sum: establish bounds and weight dominance for an equivalent encoding, or preserve preceding objectives with justified tolerances in sequential solves. An unproved first-stage optimum cannot support a claim of overall lexicographic optimality. Simulated future weather/demand/failures may be known to the world generator but unavailable to the policy.

**Boundary:** a mathematically equivalent formulation is valid despite a different method name. An explicitly labeled omniscient benchmark is permissible; it is not an executable policy. Do not import later-question information or old-contest constants into this question.

## Dynamics and physical quantities

**When:** constitutive laws, evolving stocks, accumulated output or time averages drive the answer.

Distinguish coefficient, force and power: for a passive damper with relative velocity `u` and nonnegative coefficient `c(u)`, a consistent convention gives `F = -c(u)u` and dissipated power `-Fu = c(u)u^2`. Check sign, units and relevant zero/limit cases. Recompute state balances and the objective from the trajectory: stock differs from flow, an integral differs from an average, and unequal time spacing generally requires time weights. Apply the specified transient window and terminal/periodic conditions. For discretized fields, solver residual, grid/time/domain error and model error are distinct; test the answer quantity where those errors could change it.

**Boundary:** active energy input is not passive dissipation; an open system needs the correct flux balance, not forced constant energy. A finite-horizon task does not automatically require a steady-state simulation or invented terminal penalty.

## Grouped data and surrogate optimization

**When:** repeated records share an object/batch/trajectory, or a fitted predictor selects a decision.

Choose the experimental unit and split to match deployment. Fit imputation, scaling, feature selection and tuning inside the appropriate training folds. Check feature availability at prediction time. Single-factor data do not automatically identify interactions; inspect estimable effects when interpreting parameters. For surrogate optima, check support domain, feasible controls and the final candidate against the original physical definition or independent observations; high training fit alone cannot validate an extremum or intervention benefit.

**Discriminating check:** for prediction on unseen objects, intersect the object IDs in train and test; row-disjoint sets may still share every object. Preserve the offending IDs as diagnostic evidence.

**Boundary:** predicting future records for known objects can permit shared IDs if temporal and feature availability constraints are respected. Low rank limits particular parameter interpretations, not every prediction. Without candidate validation, report an unverified prediction rather than fabricating experimental confirmation.

## Data transformations and evaluation targets

**When:** cleaning, joining, aggregating or filling data changes the observations supporting an answer, or a reported metric chooses a model.

Reconcile keys, multiplicities, units, denominators and time coverage across the raw-to-result path. Check whether an operation changes the intended population or weights, and whether preprocessing uses information unavailable at deployment. Compare models on the same target, population and information set; choose errors or costs that match the requested use. A global average can hide failure on the group or horizon that matters.

**Discriminating check:** a supposedly one-to-one join duplicates an entity because the lookup table contains two rows for its key; a sum or mean can then change even though every original key remains present. Inspect expected key cardinality and the affected records, not only nonempty output.

**Boundary:** one-to-many joins and population-changing filters can be intended. Verify their mathematical meaning and resulting weights; do not force row-count invariance on every transformation. A zero denominator or unavailable outcome may make a metric undefined, not zero error.

## Statistical interpretation and causal claims

**When:** the answer interprets an interval, test, regression coefficient or policy effect as a statement about a population or intervention.

Identify the estimand, sampling/assignment unit, dependence and assumptions needed for that interpretation. Uncertainty must match the design; treating correlated records as independent can manufacture precision. Account for outcome-driven selection or multiple comparisons where they support the claim. Distinguish association, prediction and intervention effects. A design name such as DID or IV is not evidence that its identification assumptions hold.

**Discriminating check:** duplicating every observation should not create independent information; a narrower interval from treating copies as new independent samples exposes a unit-of-analysis error. Inspect the data-generating dependence and variance calculation rather than deciding from the p-value alone.

**Boundary:** a nonsignificant difference is not equivalence; a meaningful equivalence claim needs a justified margin and suitable evidence. Descriptive or predictive tasks need no invented causal experiment. Missing identification can limit the claim, but dropping a required causal question does not answer it.

## Ranking and constructed indicators

**When:** normalized indicators and chosen weights produce a ranking or recommendation.

Trace indicator direction, units, missing-data handling, correlated or duplicate constructs, weights and aggregation. Explain whether the weights express stakeholder preference, data variation or an estimated relation. Test plausible choices near a consequential rank reversal; report stable conclusions separately from preference-dependent ones. Recompute the published ranking from its actual published inputs and transformations.

**Discriminating check:** duplicating an indicator and splitting or retaining weights can change its effective influence; check whether the resulting rank change reflects a justified preference or accidental double counting. Calling a method “objective weighting” does not settle that question.

**Boundary:** there need not be a universal correct ranking. Sensitivity to legitimate preferences can be the answer rather than a bug; do not manufacture a single winner if the task permits conditional recommendations.
canvas-designanthropics/skillsCreate beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
docxanthropics/skillsUse this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files) or Word templates (.dotx files). Triggers include: any mention of 'Word doc', 'word document', '.docx', '.dotx', or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx or .dotx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'template', or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
pdfanthropics/skillsUse this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
pptxanthropics/skillsUse this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx or .potx file (even if the extracted content will be used elsewhere, like in an email or summary); editing, modifying, or updating existing presentations; combining or splitting slide files; working with templates (.potx), layouts, speaker notes, or comments. Trigger whenever the user mentions "deck," "slides," "presentation," or references a .pptx or .potx filename, regardless of what they plan to do with the content afterward. If a .pptx or .potx file needs to be opened, created, or touched, use this skill.Stars: 198kUpdated Claude CodeCodexVerifiedInstall