agent-guardrailslisted

# Agent Guardrails ## Overview An unconstrained GTM agent is a liability — it can send bad emails, exceed budgets, violate compliance, and damage your brand. Guardrails are the safety mechanisms that keep agents effective without letting them run wild. This skill covers the full guardrail stack. ## Frameworks Referenced This skill is grounded in public frameworks and source material relevant to the task: - **AI Safety Framework.** Use the relevant method or published guidance where it improves the requested deliverable; do not cite it as decoration. - **Anthropic Constitutional AI Principles.** Use the relevant method or published guidance where it improves the requested deliverable; do not cite it as decoration. - **Anthropic — Agent Skills Progressive Disclosure.** Use the relevant method or published guidance where it improves the requested deliverable; do not cite it as decoration. ## When to Use - "Add guardrails to my agents" - "Agent safety controls" - "Prevent agent from [bad behavior]" - "Agent compliance rules" - "Agent kill switch" ## Step-by-Step Process ### Phase 1: Content Guardrails Before any agent-generated content reaches a human or external system: **Tone & Appropriateness:** - Must pass professionalism check: no profanity, no aggression, no manipulation - Must pass brand voice check: consistent with brand guidelines - Must pass factual accuracy check: no fabricated statistics, no made-up customer names **Compliance:** - Must include unsubscribe li