Who it is forDeveloper building an agent that calls MCP tools such as database, release, or package publishing tools.
The problemAn agent that can call a tool is not the same as being authorized to run that action in the current context.
The outcomeEach privileged tool call returns an explicit decision before it runs, and the agent can fetch the matching Decision Dossier.
What you need first
- A Docker environment, such as Docker Desktop, where you can run the Decionis components
- An AI agent or MCP client that can send proposed actions to Decionis before running a tool
- A written policy for which actions need PROCEED, HOLD, or BLOCK, since Decionis evaluates against versioned policy
- Awareness that the README says components are released incrementally, so some listed features may not be available yet
Steps
- Run Decionis as a containerized MCP service in your Docker environment.
- Register the Decionis MCP server with your MCP client so the agent can call it.
- Instruct the agent to request authorization before any database mutation, package publish, or release creation.
- Stop the tool call if the response is HOLD or BLOCK, and wait for PROCEED before continuing.
- Retrieve the Decision Dossier for each governed call and store it with your run logs.
Prompt for your agent
Using the decionis/docker repository, request an authorization decision from the Decionis MCP server for a database.drop action on <target database>. Expected output: a PROCEED, HOLD, or BLOCK verdict plus an evaluation ID. Verify that the agent does not call the drop tool unless the verdict is PROCEED.
Caveat. The README describes the MCP server as a planned component released incrementally, so confirm it is available in your version before you build on it.
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Disclaimer
Deduped lists open-source repositories that belong to other people. We do not own, maintain, audit, or vouch for any of them, and a repo appearing here is not an endorsement or a safety check. The use cases, steps, and prompts are written by AI and may be incomplete, outdated, or just wrong. Where a use case is marked "Seen in the wild", we link to public posts where someone described something similar, but we have not verified those posts and the use case itself is still our AI's own write-up. Use all of this at your own risk. Only run code, commands, or agents on systems and accounts you are authorized to use, and read the code and every command before you run it, especially anything that touches credentials, files, or the network. Nothing here is professional, security, legal, or financial advice. All trademarks and project names belong to their respective owners, and their use here does not imply affiliation. Content is provided as is, without warranties of any kind. Found a mistake, a safety problem, or something you want removed? Email [email protected] and we will look at it promptly. Full disclaimer page