MCP Server with Docker tools and resources. This server provides tools for interacting with Docker containers and images. Uses the docker-py library for Docker operations.
Docker MCP server has basic tool definitions with schemas and descriptions present, but significant gaps limit production readiness. All 5 tools have descriptions (10-50 chars, below ideal 50-200 baseline) and input schemas with types. However, descriptions are generic and lack actionable context for LLM selection. No output schemas documented. Error handling returns raw strings without recovery guidance. No tool annotations (readOnlyHint/destructiveHint). Parameter descriptions exist but lack format constraints, ranges, or validation rules. docker_run has complex parameters (ports, volumes) with format examples in descriptions but no formal validation or enum constraints. No pagination support despite tools returning potentially large lists (docker_ps, docker_images). Composition is reasonable (5 focused tools), but missing idempotency guarantees and confirmation patterns for destructive operations (docker_run, docker_stop).
List available Docker images.
Return detailed information about a container.
List running Docker containers.
Run a Docker container.
Stop a running container.
No output schemas documented. Tools return unstructured strings (e.g., docker_ps returns tab-delimited text, docker_inspect returns raw JSON). LLMs cannot parse or chain results reliably.
Descriptions are too short (10-50 chars vs. baseline 50-200) and lack context. 'List running Docker containers' does not explain WHEN to use docker_ps vs docker_inspect, or what the output structure is. LLMs cannot reliably select the right tool.
docker_run and docker_stop lack destructive operation warnings and confirmation patterns. docker_run can create containers with side effects; docker_stop terminates running services. No dry-run or confirmation step to prevent accidental execution.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | 2026-07-28+ | v2 |
Complex parameters (ports, volumes in docker_run) have format examples in descriptions but no formal enum, pattern, or validation constraints. LLMs may pass invalid formats like 'port:8080' instead of '8080:80'.
Error handling returns raw strings ('Error: {str(e)}') without recovery guidance. LLMs cannot determine if errors are retryable, user-fixable, or fatal. No actionable next steps provided.