Harbor Registry Mcp
by mshegolevcommunityv0.1.0
MCP server for Harbor Registry that lets an LLM list projects, repositories, and artifacts, generate storage reports, identify cleanup candidates, and delete untagged or old artifacts. It supports Harbor 2.x instances and uses dry-run by default for bulk deletion.
From the registry: Harbor Registry MCP — projects, repos, artifacts, storage reports, cleanup (with dry-run).
Install & configure
Prerequisites
- requiredInstall Python 3.10+Python 3.10 or newer is required to run the server.
- requiredInstall uvxRecommended launcher for running the MCP server package without manual installation.
- requiredCreate Harbor credentialsYou need a Harbor instance URL plus a Harbor username and password or robot token. A robot account is recommended.https://goharbor.io/
Install in your client
Paste into the AI you're chatting with — it will figure out the rest
Please install the `harbor-registry-mcp` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install Python 3.10+** — Python 3.10 or newer is required to run the server.
- **Install uvx** — Recommended launcher for running the MCP server package without manual installation.
- **Create Harbor credentials** — You need a Harbor instance URL plus a Harbor username and password or robot token. A robot account is recommended. (https://goharbor.io/)
Canonical MCP server config (stdio transport):
- command: `uvx`
- args: ["--from","harbor-registry-mcp","harbor-registry-mcp"]
- required environment variables:
- `HARBOR_URL`: Harbor URL without a trailing slash. (example: `<your-harbor-url>`)
- `HARBOR_USERNAME`: Harbor username; robot account recommended. (example: `<your-harbor-username>`)
- `HARBOR_PASSWORD`: Harbor password or robot token. (example: `<your-harbor-password-or-token>`)
- optional environment variables:
- `HARBOR_SSL_VERIFY`: Whether to verify SSL certificates. Use true or false. Defaults to true. (example: `<true-or-false>`)
Note: README also mentions pipx installation, but the recommended and most portable config uses uvx with the PyPI package.
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.Note: README also mentions pipx installation, but the recommended and most portable config uses uvx with the PyPI package.
Tools (8)
harbor_list_projects
List Harbor projects with repository counts and visibility.
harbor_list_repos
List repositories within a project.
harbor_list_artifacts
List artifacts in a repository with tags, size, and scan status.
harbor_storage_report
Generate a full project storage breakdown across repositories and artifacts.
harbor_cleanup_candidates
Suggest artifacts to delete, such as untagged, never-pulled, or old versions.
harbor_delete_artifact
Delete a single artifact by tag or digest.
harbor_delete_untagged
Delete all untagged artifacts in a project or repository.
harbor_delete_old_artifacts
Keep the latest N artifacts per repository and delete the rest, with dry-run default.
Environment variables
HARBOR_URLrequiredHarbor URL, without a trailing slash.HARBOR_USERNAMErequiredHarbor username, typically a robot account.HARBOR_PASSWORDrequiredPassword or robot token for Harbor authentication.HARBOR_SSL_VERIFYWhether to verify SSL certificates; true/false, default true.Prerequisites
- • Python 3.10+
Packages
harbor-registry-mcp
pypiv0.1.0transport: stdio
Details
- Language
- Python
- License
- MIT
- Released
- Apr 18, 2026
- Last commit
- May 2, 2026
- Registry name
- io.github.mshegolev/harbor-registry-mcp