OneCLI – Vault for AI Agents in Rust
Agents never see real keys, but Vault already does secret injection.
A lightweight HTTP proxy that injects authentication credentials into outbound API requests. Designed to run alongside OpenClaw to provide AI agents in sandboxed environments (like Docker containers) with authenticated API access without exposing secrets directly.
It replaces real API keys with short proxy tokens (e.g. PROXY:openai) and uses an X-Upstream-Host allowlist to ensure containers can only hit approved endpoints — neat and low-friction. Comes as a tiny Rust CLI with init/secret set/start commands and clear SDK examples, so you can bolt it onto OpenClaw or Docker agents without adopting a full secrets vault.
Backend developers and infrastructure/security engineers running sandboxed AI agents who need to keep API keys off the agent
* Put all auth tokens into a secrets directory
* Run OpenClaw in sandbox-exec mode using a shell wrapper. OpenClaw process is blocked by the OS from accessing secrets.
* OpenClaw routes API requests to HTTP proxy that injects auth tokens.
Agents never see real keys, but Vault already does secret injection.
Delegation chains with full visibility beat JWT, but switching costs are massive.
Agents never see credentials — brokered access beats retrieval for prompt injection safety.
Formal verification + 20 threat layers for AI agent security at sub-5ms latency.
Three UIs (TUI, terminal, web) simplify MitM debugging, but mitmproxy and Burp already own this.
Macaroon-based budget enforcement for AI agents—fills a real economic governance gap.