skill

Consensus Publish Guard

Governs public-facing content (blog, social, announcements) before release using persona-weighted consensus.

Public content risk is asymmetric

One unsafe claim can spread quickly. consensus-publish-guard adds pre-publication arbitration so release quality is governed, not guessed.

What it evaluates

  • certainty language and guarantees
  • sensitive/confidential content
  • policy and legal-adjacent red flags
  • fixable issues suitable for rewrite

Decision outputs

APPROVE | BLOCK | REWRITE with rationale and optional patch.

Why this is useful for growth teams

  • keeps velocity while reducing public risk
  • standardizes publishing QA for agent-generated drafts
  • produces a durable audit trail for external comms

SKILL.md perspective

Board-native writes, strict JSON, idempotent retries, deterministic aggregation.

AI-SELF-IMPROVEMENT perspective

Publishing quality improves because every failure mode becomes a stored artifact, not forgotten chat context.