Skip to content

Model Cards

Overview

The FCC framework auto-generates 173 model cards following the Mitchell et al. (2019) and Gebru et al. (2021) standards. Each card documents a persona's intended use, limitations, ethical considerations, and R.I.S.C.E.A.R. specification.

Model Card Categories

Persona Model Cards (147)

Individual model cards for the 102 core personas plus 45 vertical personas, organized by category:

Category Count Description
Core 5 Foundation personas (Research Crafter, Blueprint Crafter, etc.)
Integration 8 System integration specialists
Governance 4 Compliance and policy personas
Stakeholder 5 Communication and coordination personas
Champion 4 Elevated orchestration personas
Data Engineering 6 Data pipeline and quality personas
ML Lifecycle 9 End-to-end ML workflow personas
ML Models 11 Model-specific specialist personas
DevOps 3 Infrastructure and deployment personas
App Development 2 Application building personas
Open Science 4 FAIR data and reproducibility personas
Responsible AI 5 Ethics, explainability, and compliance personas
JV Collaboration 4 Partnership and IP management personas
Docs-as-Code 4 Documentation engineering personas
Privacy 3 Data protection and consent personas
Knowledge Graph 3 Ontology and semantic personas
Local-First AI 4 Edge computing and federation personas
UX & Visualization 6 Interface and visualization personas
Protocol Engineering 6 A2A/MCP protocol personas
JV Governance 6 Joint venture governance personas
Healthcare (vertical) 11 Healthcare and life sciences personas
Finance (vertical) 11 Financial services personas
Legal (vertical) 5 Legal and regulatory affairs personas
Insurance (vertical) 6 Insurance personas
Energy (vertical) 6 Energy personas
Government (vertical) 6 Government personas

Workflow Model Cards (6)

Cards documenting each workflow action type:

  • Scaffold — Project initialization and structure generation
  • Refactor — Code and architecture improvement
  • Debug — Issue identification and resolution
  • Test — Verification and validation
  • Compare — Multi-option analysis and recommendation
  • Document — Documentation generation and maintenance

Category Model Cards (20)

Aggregate cards for each persona category, summarizing collective capabilities, risk profiles, and interaction patterns.

Generating Model Cards

# Generate all 173 model cards
fcc model-card --personas all --output docs/model-cards/

# Generate for a specific persona
fcc model-card --persona RC --output docs/model-cards/

# Generate for a category
fcc model-card --category core --output docs/model-cards/

Card Structure

Each model card includes:

  1. Model Details — Name, version, type, owner
  2. Intended Use — Primary and out-of-scope use cases
  3. R.I.S.C.E.A.R. Summary — Role, inputs, style, constraints, outputs, archetype
  4. Factors — Relevant demographic, environmental, and domain factors
  5. Metrics — Evaluation criteria and performance measures
  6. Training Data — Data sources and preprocessing (for AI-powered personas)
  7. Ethical Considerations — Bias risks, fairness, and mitigation strategies
  8. Caveats & Recommendations — Known limitations and usage guidelines
  9. Quantitative Analyses — Discernment matrix and design target scores