What does fail-closed mean in the context of this governance layer?

Fail-closed means that if the governance system cannot verify a policy or encounters an error, it denies the action by default rather than allowing it. This ensures safety and compliance even when systems are degraded.

How do I authenticate and get started with dingdawg-governance?

Install via `npx dingdawg-governance@2.0.6`, then set the `DINGDAWG_API_KEY` environment variable. Free tier keys are available at dingdawg.com; paid tiers unlock additional features.

Can I audit and prove what actions my agents took?

Yes. The server provides immutable audit proofs that record every agent action, when it occurred, and the governance decision applied. These proofs are cryptographically verifiable for compliance reporting.

Does this work with any AI agent framework or only specific ones?

It is MCP-native, meaning it integrates with any system that supports the Model Context Protocol. Check the documentation for framework-specific setup guides.

What happens if I need to undo an agent action that was already executed?

The rollback feature allows you to revert executed operations to a previous state. Rollback rules can be configured as part of your governance policy to automate recovery from failed or unwanted actions.

MCP ServerSTDIOOfficialv2.0.6

Dingdawg Governance MCP Server

Enforce governance policies and audit trails for AI agent actions with fail-closed safety guarantees. Built for developers who need compliance, rollback capabilities, and immutable proof of agent decisions.

io.github.dingdawg/dingdawg-governance

Try MCP with AI models View source

Hosted URL

Local install

Transport

STDIO

Auth

No auth required

What the Dingdawg Governance MCP server does

How models use it and what it is built for.

Connect to Dingdawg Governance

Local install — runs as a subprocess.

npx dingdawg-governance@2.0.6

Environment variables

Configuration this server reads at startup.

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

Resources

Where to find authoritative docs and source for Dingdawg Governance.

Source repository

Example prompts for Dingdawg Governance

Paste any of these into Agent Studio after connecting Dingdawg Governance.

Set up a governance policy that requires approval before agents execute high-risk operations
Query the audit log to see which agent actions were executed and when they occurred
Configure rollback rules so failed agent operations automatically revert to the previous state
Explain how to integrate DINGDAWG_API_KEY for production governance enforcement

Dingdawg Governance MCP server — FAQ

Common questions about connecting and running Dingdawg Governance.

What does fail-closed mean in the context of this governance layer?
Fail-closed means that if the governance system cannot verify a policy or encounters an error, it denies the action by default rather than allowing it. This ensures safety and compliance even when systems are degraded.
How do I authenticate and get started with dingdawg-governance?
Install via `npx dingdawg-governance@2.0.6`, then set the `DINGDAWG_API_KEY` environment variable. Free tier keys are available at dingdawg.com; paid tiers unlock additional features.
Can I audit and prove what actions my agents took?
Yes. The server provides immutable audit proofs that record every agent action, when it occurred, and the governance decision applied. These proofs are cryptographically verifiable for compliance reporting.
Does this work with any AI agent framework or only specific ones?
It is MCP-native, meaning it integrates with any system that supports the Model Context Protocol. Check the documentation for framework-specific setup guides.
What happens if I need to undo an agent action that was already executed?
The rollback feature allows you to revert executed operations to a previous state. Rollback rules can be configured as part of your governance policy to automate recovery from failed or unwanted actions.

Skip the local setup — run MCP in your browser

MCP Playground runs 10,000+ hosted MCP servers — GitHub, Linear, Notion, Stripe, Sentry and more — across Claude, GPT, Gemini, DeepSeek and 60+ AI models. Compare model answers side-by-side, save agent presets, share runs. Zero install.

Open Agent Studio