More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    SecretSafe – use API keys without seeing them

    2

    Your agent calls OpenAI, GitHub, Stripe & any API with your keys, without ever seeing them. Keys stay in a Terminal 3 TEE vault; only allow-listed hosts get them, and every echo is redacted.

    Free

    Security scanned
    SecretSafe – use API keys without seeing them

    SecretSafe – use API keys without seeing them

    Example session with this skill installed

    Use my github secret to list my 3 most recently updated repositories.

    • Read your context and instructions
    • Compiled the secretsafe

    Your 3 most recently updated repositories

    I called the GitHub API through SecretSafe. Your token was injected by SecretSafe at request time; I only saw it as [REDACTED:github].

    Request: secure_request → GET https://api.github.com/user/repos?sort=updated&per_page=3
    Secret: github (allowed host: api.github.com, injected into Authorization header)
    Status: 200 OK

    #RepositoryVisibilityLast updated
    1ElTanque81/secretsafepublictoday
    2ElTanque81/hodlguardpublic6 days ago
    3ElTanque81/prejudgepublic4 days ago

    What happened behind the scenes

    • list_secrets showed a secret named github, allowed only for api.github.com. No value was returned.
    • I wrote the request without the key; SecretSafe read it from your Terminal 3 vault (bound to your did:t3n) and added the Authorization header.
    • Any echo of the token in the response would have been replaced with [REDACTED:github].

    Blocked attempt (for comparison)

    A request with the same secret to https://webhook.site/... was refused before sending:
    Host "webhook.site" is not on this secret's allow-list (api.github.com).

    Next steps

    • Want me to open an issue or check CI status on one of these repos? I can use the same secret.
    • To use another API, add its key yourself with secretsafe add <name> --host <api.host.com>; never paste it into the chat.

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

    About this skill

    Stop pasting API keys into your agent.

    SecretSafe is an MCP server + skill that keeps your API keys in the Terminal 3 T3 Network TEE vault, bound to your own did:t3n identity. Your agent only sees secret names and writes {{secret}}. SecretSafe injects the real key at call time, only for hosts you allowed, and replaces any echo with [REDACTED].

    This skill teaches your agent to:

    • never ask for keys in chat, read .env files or hard-code secrets
    • use list_secrets / secure_request for any authenticated API call
    • tell you which secretsafe add command to run when a key is missing
    • respect the host allow-list (no workarounds, prompt-injection resistant)
    • walk you through setup in Claude Desktop, Claude Code or Cursor

    Includes ready made request templates for OpenAI, Anthropic, GitHub, Gemini, Stripe (test mode), Hugging Face, Notion and Slack.

    Requires the free, open-source SecretSafe MCP server and a free Terminal 3 account (test credits included). Setup takes about 3 minutes.

    Author: ElTanque81, creator of SecretSafe. Source: https://github.com/ElTanque81/secretsafe (MIT) Demo: https://youtu.be/dcWbVviVK7E

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.

    Supports only HTTPS on port 443. Secrets are briefly held in local memory during execution. Requires a Terminal 3 account and credits for vault reads.

    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 buying a skill.

    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

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 6 days ago

    • Free with an Agensi account

    Accounts
    Terminal 3 account (free)
    Software
    GitNode.js 20+
    Optional
    macOS Keychain (for storing the T3 key)

    Install the SecretSafe MCP server first (3 minutes). Your keys are stored by you in the terminal and never entered in chat.

    Permissions

    Terminal / Shell
    Read Files
    Network Access
    Listed6 days ago

    What's inside

    Frequently Asked Questions

    No. It only sees secret names and writes {{secret}}; SecretSafe injects the key at request time and redacts any echo.

    In your own vault on the Terminal 3 T3 Network (TEE), readable only by your did:t3n.

    Each key has an exact host allow list. Requests to any other host are refused before the key is used.

    New in Security

    Free