Runtime — SOFIA's first implementation uses Claude Code and CLAUDE.md files. The concepts described here (persona, isolation, artifacts) are provider-agnostic — only the runtime layer is specific to a provider. Some sections describe Claude Code specifics. Latter versions implement other providers.

Persona Archetypes

Starting points for persona design — not prescriptions. Adapt to your project.

Each archetype is a template with the 7 persona dimensions filled for a common role. Use them as inspiration when bootstrapping personas, not as rigid definitions.

Archetype Role Key prohibition File
Architect System architecture, specs, ADRs Doesn't code persona-architect.md
Designer UI design, visual craft, design system Doesn't code persona-designer.md
Developer Code, tests, implementation Doesn't decide architecture persona-dev.md
Meta-challenger Anti-echo-chamber, challenges premises No operational artifact, no memory persona-meta-challenger.md
Method inspector Method conformity, structural audit Doesn't make product decisions persona-method-inspector.md
Researcher Scientific validation, sourcing Doesn't code persona-researcher.md
Strategist Positioning, risk, messaging Doesn't code persona-strategist.md
UX User flows, behaviors, acceptance criteria Doesn't code persona-ux.md
Writer Editorial distillation, long-form content Doesn't decide architecture persona-writer.md

How to use

  1. Pick the archetype closest to your need
  2. Copy the file into your instance (shared/orga/personas/persona-{name}.md)
  3. Replace {Name} with your persona's name
  4. Adjust the prohibitions — they must create tension with your other personas
  5. Create the matching context file (shared/orga/contextes/contexte-{name}-{product}.md)

The prohibition column is the most important. If two personas have no prohibitions in tension, they won't produce useful friction.