io.github.dingdawg/dingdawg-governance

Official

by dingdawg · JavaScript

Universal governance layer for AI agents. MCP-native, fail-closed, audit proofs and rollback.

Universal governance layer for AI agents with fail-closed controls and audit trails.

Topics

ai-agentsai-governanceaudit-trailcolorado-ai-actcomplianceeu-ai-actllm-governancemcpmodel-context-protocolnist-ai-rmfreceiptsresponsible-aitamper-proofgovernance-receipts

README.md4 KBView on GitHub

DingDawg Governance

Every AI action receipted. Capability-gated. Rollback-ready.

AI Governance-as-a-Service via MCP. Every write, shell command, and state-changing operation your agent makes is validated before it runs and cryptographically receipted after it completes — giving you a tamper-evident audit trail you can query, export, or roll back at any time.

Works with Claude Code, Codex, Cursor, Windsurf, and any MCP-compatible agent.

Install

bash

npm install dingdawg-governance

Quick Start

bash

# Authenticate (free — no credit card)
npx dingdawg-governance auth login

# Start the governance MCP server
DINGDAWG_API_KEY=your_key npx dingdawg-governance

Add to your MCP client config:

json

{
  "mcpServers": {
    "dingdawg-governance": {
      "command": "npx",
      "args": ["dingdawg-governance"],
      "env": {
        "DINGDAWG_API_KEY": "your_key_here"
      }
    }
  }
}

Get your free API key at app.dingdawg.com/settings/api.

What You Get

Pre-execution validation — Every action checked against your policy before it runs. Blocked actions never touch your filesystem.
Cryptographic receipts — Tamper-evident record for every action with a unique ID, output hash, and public verification URL.
One-command rollback — Undo any file write by referencing its receipt ID. Prior state is captured automatically before every write.
Real-time trust scores — Behavioral pattern tracking across sessions surfaces anomalies before they become incidents.
Compliance reports in seconds — SOC 2 and ISO 27001-aligned audit reports as signed PDF or structured JSON. One API call.

MCP Tools

Tool	What it does
`validate_action`	Pre-execution check. Returns `APPROVED`, `FLAGGED`, or `BLOCKED` with reason code and risk score.
`generate_receipt`	Post-execution cryptographic receipt with tamper-evident output hash.
`capture_rollback_state`	Snapshot current state before destructive or high-risk operations.
`rollback_action`	Restore prior state from a receipt ID.
`query_receipts`	Search receipts by date, agent, action type, or status. Export-ready.
`check_status`	Current tier, daily usage, quota, and active governance alerts.
`generate_audit_report`	On-demand compliance report — SOC 2, ISO 27001, or custom policy framework.

Plans

Plan	Governed actions/day	Rollback window	Audit reports
Free	200	—	—
Pro — $29/mo	10,000	30 days	Unlimited
Business — $149/mo	Unlimited	90 days	Compliance PDF

No credit card to start. Full pricing: dingdawg.com/governance#pricing

Documentation

Full docs at dingdawg.com/governance/docs

Contributing

Issues and pull requests welcome. See CONTRIBUTING.md.

github.com/dingdawg/governance-mcp

Built by DingDawg — Trust layer for the agentic internet.

Install

Configuration

Environment variables

DINGDAWG_API_KEYsecret

API key for paid tier access — get free at dingdawg.com

claude_desktop_config.json

{
  "mcpServers": {
    "dingdawg-governance": {
      "command": "npx",
      "args": [
        "-y",
        "dingdawg-governance@2.1.6"
      ],
      "env": {
        "DINGDAWG_API_KEY": "<YOUR_DINGDAWG_API_KEY>"
      }
    }
  }
}