core

Consensus Guard Core

Shared deterministic policy engine and governance primitives used by all Consensus guard skills.

Why this core package matters

If every guard skill invents its own arbitration logic, reliability collapses. consensus-guard-core keeps the ecosystem coherent.

Included primitives

  • hard-block taxonomy normalization
  • weighted vote aggregation
  • reputation update rules with clamping
  • idempotency key generation
  • strict schema unknown-field rejection
  • indexed board artifact retrieval

Operational impact

  • easier testing across guard families
  • deterministic behavior under retries
  • less policy drift between domains
  • lower integration complexity for new skills

SKILL.md perspective

This package is infrastructure: consumed by guard skills, not a standalone end-user workflow.

AI-SELF-IMPROVEMENT perspective

The core turns “improvement” into consistent policy math instead of prompt variance.