janee

What This Skill Does

Janee serves as a secure proxy and secrets manager specifically designed for AI agents. Rather than storing sensitive API keys in plaintext configuration files that are vulnerable to prompt injection or accidental exfiltration, Janee keeps your credentials encrypted at rest. When your agent needs to interact with an external API, it delegates the request to the Janee service. Janee then handles the authentication internally, injecting the necessary keys into the HTTP request headers before forwarding it to the target service. This 'zero-exposure' architecture ensures the AI agent itself never actually 'sees' or has access to your raw credentials. Additionally, Janee provides audit logging, path-based access control, and a centralized kill switch to revoke access immediately if a compromise is suspected.

Installation

To begin, ensure you have Node.js installed on your system. First, install the global CLI tool: npm install -g @true-and-useful/janee Initialize the configuration store with janee init. You can add new service credentials by running janee add. Finally, to integrate this capability into your OpenClaw environment, install the dedicated plugin using the OpenClaw CLI: openclaw plugins install @true-and-useful/janee-openclaw This provides the necessary MCP tools for your agent to communicate with the Janee daemon.

Use Cases

Janee is ideal for developers and users building autonomous agents that need to perform actions across multiple third-party services like Stripe, GitHub, or specialized web platforms. By using Janee, you can safely give an agent permission to read your Stripe balance without providing the agent with your full 'write' or 'delete' credentials. It is also perfect for team environments where you want to allow an agent to use shared organizational keys without exposing the secret strings to every developer or agent instance running on local machines.

Example Prompts

1. "Janee, please list all the services I currently have configured and check their status."
2. "Use the moltbook service to post an update to the /posts endpoint with the following body: 'Hello world from my automated agent!'"
3. "Janee, verify if the stripe service is restricted to read-only access and confirm the current authentication rules."

Tips & Limitations

To maximize security, always define strict path-based policies for your services. For example, if an agent only needs to fetch data, explicitly deny 'POST', 'PUT', and 'DELETE' methods in your configuration file. Remember that Janee acts as a middleman; while this provides immense security benefits, ensure the machine hosting the Janee daemon is secure, as it holds the encrypted key store. Always keep your local Janee CLI updated to the latest version to benefit from security patches.