# OneCLI: Secure Credential Gateway for AI Agents Open source solution to keep sensitive data safe from AI agents. - **Source:** [Hacker News: Show HN](https://news.ycombinator.com) - **Published:** 2026-07-23 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/715f93d4-6eb4-43b4-aa41-3da0b5b7be8b ## Introduction OneCLI is an open source network gateway that sits between AI agents and the services they use, keeping credentials secure. Unlike traditional vaults, it maintains control over how secrets are used after delivery to agents. ## Problem AI agents need sensitive information like API keys to perform tasks. However, storing these in agent memory or local files creates vulnerabilities. Malicious actors can manipulate agents to disclose these credentials. ## Solution OneCLI prevents direct credential access by agents. It matches requests by host/path, verifies access permissions, replaces placeholders with actual credentials, and forwards requests. All secrets remain encrypted at rest. ## Technology The Rust-written proxy features a Next.js dashboard. It uses AES-256-GCM encryption and runs in Docker containers. Integrates with popular password managers like Bitwarden or 1Password. Compatible with any AI agent framework. ## Use Cases Users can require human approval for sensitive operations. For example, an agent can review PRs but can't merge without approval. Companies currently use it mostly for coding agents, preventing permission escalation attempts. ## Significance AI model behavior can't be fully trusted. OneCLI maintains control through deterministic rules. While agents may find creative solutions, tight security policies minimize risks by balancing their capabilities with necessary restrictions. ## Sources covering this story - https://onecli.sh — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._