Draugr — Claude Code MCP server: what it does, how to install
Draugr coordinates security tools from a single configuration file and ranks findings by actual exposure. It helps teams evaluate release risks and gates pull requests on newly introduced flaws.
Open on GitHubWhat Draugr does
Draugr runs established engines such as Trivy, Semgrep, and Gitleaks through a unified draugr.saga.yaml file. Instead of configuring separate scanners, you describe your services, images, and infrastructure once. The tool selects applicable checks for dependencies, exposed secrets, container images, and infrastructure code.
Traditional tools flag severity in isolation. Draugr calculates priority from P1 to P4 by checking whether an affected component faces the public internet. A high-severity flaw in a private internal helper will not derail your release.
In pull requests, draugr diff inspects only new issues. That keeps teams from getting stalled by inherited technical debt. Scanners must be fetched beforehand with draugr tools install.
What it helps with
- scan application dependencies for vulnerabilities with Trivy
- detect leaked secrets and committed credentials with Gitleaks
- check Terraform and Kubernetes manifests for misconfigurations
- block pull requests only when new security issues appear
Example prompts for the agent
- Run a Draugr scan on this repo and report P1 priority issues
- Execute
draugr diffto evaluate security findings introduced in this branch - Create a
draugr.saga.yamldescriptor for the project usingdraugr survey
How to install Draugr
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Draugr
How to install Draugr in Claude Code?
Run curl -fsSL https://draugr.dev/install.sh | sh to place the binary into ~/.local/bin. Then run draugr init to generate draugr.saga.yaml and execute draugr tools install to download required scanners.
What does Draugr do?
It orchestrates SAST, SCA, and secret scanning through a single descriptor. All detected vulnerabilities are weighted and prioritized based on exposure and component reachability.
How does P1–P4 priority differ from severity?
Severity measures a bug's theoretical worst-case impact. Draugr's priority considers whether the flawed component is reachable from the network and how critical it is to operations.