• SECURE
  • REMOTE
  • LOCAL
  • HUMAN

How to manage pipeline secrets and files with your AI

Connect Dagger to your AI client to let it handle sensitive credentials and remote assets. Once connected, your agent can prep your environment for any task.

Ask AI about this page

Short answer

How can my AI access secrets and files for my pipeline?

Your AI uses Dagger to pull secrets from your environment and fetch files from the web or your local host. This allows your agent to build and run tasks with the exact context it needs. You'll see how the agent handles these inputs in the breakdown below.

Where every outcome lands.

The results of your AI's actions.

Depending on what you ask, your agent will move data between your environment, the web, and your local machine.

SECURE

Ready secrets

The AI pulls specific credentials needed for a run. You get a prepared environment without ever seeing the raw secret values.

REMOTE

Fetched files

Your agent grabs files from a URL and brings them into the execution context. This sets up your dependencies or config files automatically.

LOCAL

Host access

The AI reads files directly from your local machine to use as input. This bridges the gap between your files and the Dagger engine.

HUMAN

Manual review

If a secret is missing or a URL is unreachable, the AI stops. You'll need to step in to fix the path or provide the right permission.

The workflow

What your AI does when the request arrives.

The agent follows a specific path to gather the data required for your programmable CI tasks.

  1. Fetch remote assets

    The agent reaches out to a specific URL to download necessary files or configurations.

    query_http

  2. Retrieve credentials

    The agent pulls sensitive environment variables or tokens needed to authenticate a process.

    query_secret

  3. Read local data

    The agent looks at your current machine to find files that need to be part of the pipeline.

    query_host

  4. Verify environment

    The agent checks that all required files and secrets are present before starting the execution.

    query_hostquery_secret

Try it

Copy these to start.

Use these prompts to kick off your first Dagger task.

Starting points

Swap out the repo names, URLs, and secret keys with your own specific details.

Dagger Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_jU749naeDKubVrRR38BTiJPB8XzxtT6YXppJtFAW/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Dagger capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "dagger-programmable-ci-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_jU749naeDKubVrRR38BTiJPB8XzxtT6YXppJtFAW/mcp"
    }
  }
}
Copy into chat04

  • Use Dagger to fetch the config file from https://example.com/config.yaml and prepare my environment.

  • Pull the AWS_ACCESS_KEY_ID secret from my environment and use it to run a Dagger build.

  • Read the contents of ./src/main.py from my host and use it in a Dagger container.

  • Download the latest build artifact from the remote URL and check if it matches my local file.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Dagger once, then ask.

You only need to link the MCP once. Your credentials stay encrypted on the Vinkius side, and you can find the full list of capabilities on the Dagger page.

Connect Dagger to your AI

FAQ

How the Dagger MCP behaves.

  • 01

    Can the AI change my environment variables?

    No. The AI can only query existing secrets and files to use them in a task. It cannot modify your system's environment variables.

  • 02

    Will my secrets be exposed in the chat history?

    No. The AI uses the tools to pull the data into the execution context, but it does not print the raw secret values back to you in the chat.

  • 03

    Can the AI delete files on my computer?

    No. The tools only allow the AI to read files from your host or fetch them from the web. It has no capability to delete or overwrite local files.

  • 04

    How does it handle a 404 error when fetching a file?

    The tool will fail to find the resource, and the AI will report that the file could not be retrieved from the provided URL.

  • 05

    Can I use this to upload files to a server?

    No. The current tools are designed for querying and fetching data, not for pushing or uploading files to remote destinations.

  • More questions about Dagger? The Connector page answers them. See everything the Dagger Connector can do

Connect Dagger to Claude, Cursor, ChatGPT & more