What it does
The server lets a coding agent lint one Dockerfile or Kubernetes manifest. Its checks cover root in the final stage, unpinned base images, curl piped to sh, secrets in ENV or ARG, Pod Security Standards, resources and probes. Each lint returns an A–F grade, a score, a passed flag set by a configurable failure threshold, and every finding with its line, a fix and a reference.
It suits a point where an agent is about to hand over a container image definition or a Kubernetes manifest and you want those findings reviewed first. Files named like Dockerfile or Containerfile are treated as Dockerfiles when given by path.
The server runs locally over stdio and is started with uv. It needs an API key from RapidAPI or api.market, set in the environment; if both keys are set, a variable selects the marketplace. Without a key it still starts, and the first tool call explains how to get one. Each tool call counts as one request on your plan, and the free Basic plan allows 15 requests a month. Local files are read only when a tool is given their path, and they are sent to the API for that single request, which processes them in memory without storing them. Your key is sent only to the marketplace you chose. The server is MIT-licensed open source.
Limitations and requirements
Requires your own RapidAPI or api.market key for the paid Shipshape API; the free Basic plan allows 15 requests a month with a hard limit. Files must be up to 512 KiB. Helm templates must be rendered first (helm template). NetworkPolicies, PodDisruptionBudgets and RBAC are not checked. Requires Python 3.10 or later.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.Bytestay/shipshape-mcp
Dockerfile and Kubernetes lint for AI agents: Pod Security Standards, A–F grade, a fix per finding.
v0.2.9As shown on registry.modelcontextprotocol.io when we last read it