vibeflow-report

name

vibeflow-report

description

Use when planning, drafting, revising, validating, or completing a VibeFlow/VibeCAE report from 0-caseDict/caseDict and artifacts in A-geo, B-mesh, C-run, and D-post, including problem statement, assumptions, geometry, mesh, physical models, materials, boundaries/loads/initial conditions, numerical setup, results, limitations, conclusions, evidence traceability, figure/table selection, and report readiness checks.

VibeFlow Report

This skill is responsible for organizing the case process into a deliverable, traceable, reviewable report. A report is not a simple file summary; it organizes the unified physical/numerical facts from caseDict, upstream artifacts, and post-processing evidence into a logical CAE argument.

When to Use

• Results are already sufficient to support conclusions.
• Evidence from A-geo, B-mesh, C-run, and D-post needs to be summarized into a document.
• Limitations and applicability scope need to be made explicit.
• It is necessary to judge whether the report lacks geometry, mesh, boundary, numerical-setting, result, or error evidence.
• A paper-like Chinese or English narrative needs to be formed from charts, logs, and caseDict.

What to Read First

• caseMeta, intent, physicalProblem, geometry, materials, regions, interfaces, boundariesLoadsInitials, mesh, numericalMethod, execution, postProcessing, report, provenance, and errors in 0-caseDict/caseDict
• Key artifacts already generated under A-geo/, B-mesh/, C-run/, and D-post/
• Report structure and evidence checklist: references/report-checklist.md

Inputs

• 0-caseDict/caseDict
• Geometry artifacts under A-geo/
• Mesh artifacts under B-mesh/
• Dictionaries/configuration and logs under C-run/
• Charts and metrics under D-post/

Outputs

• E-report/report.md
• When necessary, E-report/figures/ copied indexes, E-report/evidence_index.md, or data tables for the report
• When necessary, updates to caseDict.report, postProcessing.artifacts, provenance.artifactTrace, and errors

Core Workflow

1. Extract problem definition, parameters, assumptions, and configuration facts from caseDict.
2. Extract evidence from the four stage directories.
3. Distinguish:
   • verified facts
   • engineering assumptions
   • unresolved limitations
4. Check gaps: when core evidence is missing, return upstream instead of forcing conclusions in the report.
5. Organize a readable report rather than merely pasting figures.
6. Establish traceability among key charts, tables, logs, and conclusions.

Rules

• Do not package missing results as conclusions.
• Do not write guesses as facts.
• Every key numerical value should be traceable as much as possible to artifact files, caseDict fields, or logs.
• If core evidence is still missing, complete upstream stages first.
• Writing language follows caseDict.report.language; if unspecified, use the user's language.
• Mathematical formulas follow project conventions: display equations use $$...$$ on independent lines.
• Figures, tables, and text support each other. Each key figure explains field, time, slice/region, and physical meaning.
• Limitations must be stated; do not report only favorable results.
• Unresolved blocking errors must not be ignored; they can only be described as unfinished status or limitations.

Good Report Criteria

• The full workflow can be understood without browsing directories.
• Conclusions and limitations appear together.
• Figures, tables, and text support each other.
• Configuration, paths, units, directions, and caseDict are consistent.
• It independently explains why the case was modeled, meshed, and solved this way, and what the results mean.

Fallback Conditions

• Key parameters or assumptions are missing: return to vibeflow-case-dict.
• Geometry/mesh evidence is missing: return to vibeflow-geometry or vibeflow-mesh.
• Solve logs or results are insufficient: return to vibeflow-openfoam.
• Charts, metrics, or scenes are insufficient: return to vibeflow-postprocess.
• Toolchain issues prevent chart or document generation: return to vibeflow-toolchain.