persona

Consensus Persona Respawn

Replaces weak personas with successors learned from historical ledger mistakes to keep evaluator quality high.

Why persona lifecycle matters

A static evaluator panel degrades. consensus-persona-respawn keeps decision quality from plateauing by replacing low-performing personas with evidence-informed successors.

Respawn process

  1. detect dead/weak persona (threshold or explicit trigger)
  2. mine recent decision artifacts
  3. extract recurring mistake patterns
  4. generate successor persona profile
  5. write persona_respawn + updated persona_set

Reliability benefits

  • anti-stagnation for long-running agent systems
  • transparent mutation history
  • explicit lineage between old and new evaluators
  • better long-term arbitration quality

SKILL.md perspective

Simple trigger-based API with strict JSON and board-native writes.

AI-SELF-IMPROVEMENT perspective

This is evaluator maintenance: adaptation driven by observed failure, not random persona churn.