Skill Detail

Broker API credentials to AI agents without exposing raw keys using OneCLI

Store credentials once, then inject them into outbound agent requests at runtime so agents can call services without receiving raw secrets.

Security & VerificationMulti-Framework
Security & Verification Multi-Framework Published
⭐ 1.9k GitHub stars
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill broker-api-credentials-to-ai-agents-without-exposing-raw-keys-using-onecli Copy
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Tools required
OneCLI gateway and dashboard, stored service credentials, AI agents making HTTP calls through the gateway
Install & setup
curl -fsSL https://onecli.sh/install | sh
Author
onecli
Publisher
Open Source Project
Last updated
Apr 23, 2026
Quick brief

Use OneCLI when an agent must call external APIs but should not receive raw credentials directly. The upstream project is specific about the workflow: store credentials once, issue scoped agent access, match outbound requests by host or path, and inject the real secrets at request time through the gateway.

How it works

What this skill actually does

Invoke this instead of wiring secrets into agents, prompts, or app config normally when the job is controlled credential brokering for agent HTTP calls. The scope boundary is narrow enough to stay skill-shaped: OneCLI is being used here for per-agent secret injection and access control on outbound requests, not as a generic vault, dashboard, or platform listing.