More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    llm-env-vault: keep .env secrets out of your AI agent

    1

    Your agent sees DATABASE_URL="value 3"; your app gets the real value. Secrets live in an encrypted local vault, released only through a password dialog you answer.

    $7.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    llm-env-vault: keep .env secrets out of your AI agent

    llm-env-vault: keep .env secrets out of your AI agent

    Example session with this skill installed

    Protect the .env files in this project so you never see the real values, then start my app with the real database URL and Stripe key.

    • Read your context and instructions
    • Compiled the llm-env-vault

    I moved the real values out of your project's .env files and into the encrypted vault. You confirmed each change in the vault's own dialog; I never saw the values.

    FileVariableWhat I can see now
    .envDATABASE_URLpostgres://placeholder-1.invalid
    .envSTRIPE_SECRET_KEY"value 2"
    .envSMTP_PORT3
    api/.envOPENAI_API_KEY"value 4"

    vault_status() now lists 4 variables by name only. The real values live in vault.enc (AES-256-GCM), which I cannot read.

    Your app still gets the real values when you run it through the vault:
    run_with_env(["npm", "start"], only_vars=["DATABASE_URL", "STRIPE_SECRET_KEY"])
    You type the master password into a native dialog, and the command's output is redacted before it reaches me.

    Next steps

    1. Start your app through run_with_env with only the variables it needs.
    2. Back up vault.enc and keep the optional paper recovery key somewhere safe.
    3. Commit the placeholder .env if you like: it documents which variables the project needs and holds no secrets.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Telling your AI agent "don't read .env" is a suggestion, not a control.

    llm-env-vault is an MCP server that replaces the secret values in your .env files with harmless numbered placeholders, stores the real values in a password-encrypted local vault, and lets you run real commands with real secrets injected. Every time, a native password dialog that only you can answer is the gate.

    What it does

    • Placeholders, not secrets. /llm-env-vault:protect finds every .env in your project and moves the real values into the vault. The agent keeps working with the variable names.
    • Run with real secrets. run_with_env injects the real values into one command you approve, scoped to only the variables it needs. Output is redacted before it reaches the agent.
    • Typed placeholders (new in 2.0). SMTP_PORT=17, SMTP_USE_SSL=false, DATABASE_URL=postgres://placeholder-3.invalid: config loaders like pydantic-settings, django-environ and Spring still parse the file, and it still holds no real value.
    • Whole-file encryption. Encrypt a secrets file into a .levault that is safe to commit.
    • Nothing happens silently. Every operation that touches a real value shows a dialog with exactly what will change, and nothing is written until you click Allow.

    Security

    scrypt key derivation, AES-256-GCM with the header authenticated, envelope encryption with an optional paper recovery key, and atomic writes. The master password never touches disk and is never a CLI argument or environment variable. Adversarial review passes closed a leak into the agent's context, a git-config code-execution vector and TOCTOU windows.

    Known limitations

    Windows is the tested and supported platform; macOS and Linux need python on the PATH and are untested. The first run installs the Python dependencies with pip. Restart Claude Code after installing.

    Backed by 640+ tests.

    Windows is the only fully tested platform. macOS/Linux require 'python' on the PATH. Requires a manual slash command installation and restart to activate the MCP server.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Requires

    macOS
    Python
    Aware MCP
    Node.js

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Environment Variables
    Listed1 month ago

    What's inside

    Frequently Asked Questions

    It replaces sensitive values in your .env files with harmless numbered placeholders, storing the actual secrets in an encrypted local vault. This ensures that even if an AI agent scans your workspace, it only sees variable names and dummy data, not your production credentials.

    When you need to execute a command, the 'run_with_env' tool injects the real secrets into that specific process after you approve a native password dialog. The output is then redacted before being sent back to the agent to prevent accidental leaks.

    The skill uses industry-standard AES-256-GCM encryption and scrypt key derivation, ensuring your master password never touches the disk or CLI history. It has also passed adversarial reviews to mitigate risks like context leaks and git-config execution vectors.

    While it is an MCP server designed to work with tools like Claude Code, it currently officially supports and is tested on Windows. macOS and Linux users can run it if Python is in their PATH, but these platforms are currently listed as untested.

    You receive the full MCP server package, which includes tools for protecting files, running commands with injected secrets, and whole-file encryption. The purchase also includes access to the latest v2.0 features like typed placeholders for improved compatibility with config loaders.

    New in LLM Ops