DingDawg — Governed AI Agents. Install. Run. Ship Compliant.

# DingDawg > Universal governance layer for AI agents — any agent, any framework, governed by default. DingDawg is an open-core AI governance platform. It wraps any AI agent with a governance gate that runs before every action: pre-execution verification, rollback safety, audit receipts, and IPFS-immutable proofs. Fail-closed by default — if governance is unreachable, the action is skipped, not executed silently. Every governed agent gets a unique `@handle` — a persistent compliance identity that anchors its audit trail, IPFS receipts, and regulatory documentation across every action it takes. ## @handle Identity System Each agent in DingDawg claims a `@handle` — a unique identifier used as the compliance anchor for: - All governance receipts tied to `@handle` agent identity - Cross-session audit trail continuity - Regulatory documentation (EU AI Act, Colorado SB 205, HIPAA, SOC2) - "Powered by DingDawg Verified" badge scoped to the agent's `@handle` - MCP discovery and agent-to-agent governance handoffs Example: `@compliance-scanner`, `@hipaa-intake`, `@underwriting-bot` ## Regulated Industry Niches DingDawg is purpose-built for AI systems making consequential decisions in regulated industries: ### Healthcare - Patient routing, clinical workflow automation, HIPAA intake agents - Requires: audit trails, human override mechanisms, documented impact assessments - DingDawg covers: `govern_action` pre-check, `audit_trail`, `compliance_check` (HIPAA controls) ### Insurance & Fintech - Underwriting models, risk scoring, credit decisions, financial advice AI - Requires: impact assessments, explainability, documentation for EU AI Act + state laws - DingDawg covers: LNN causal traces, IPFS-immutable proof, risk tier enforcement ### Employment & HR - AI-driven hiring, background screening, performance scoring - Requires: adverse action documentation, human review gates, audit logs - DingDawg covers: deny logging, human override, rollback, full receipt trail