skill

Consensus Support Reply Guard

Screens AI-generated support responses for legal, confidentiality, and sensitive-data risk before send.

Why support needs governance

Support is high-frequency and customer-facing. Small language errors compound into trust loss and escalation cost.

What this guard catches

  • legal certainty claims
  • confidentiality leaks
  • sensitive data mishandling
  • high-risk phrasing that should be rewritten or blocked

How it helps operations

  • consistent policy enforcement across agents
  • fewer risky outbound replies
  • clear rationale for approvals and blocks
  • longitudinal improvement via reputation updates

SKILL.md perspective

Fast integration with strict JSON I/O and board artifact writes.

AI-SELF-IMPROVEMENT perspective

Support quality becomes a governed feedback system, not one-off prompt luck.