RBAC Access Control for AI Agents with Airtable & Telegram

This workflow provides granular access control for AI agents and multi-agent setups by enforcing Role-Based Access Control (RBAC). User permissions for specific tools are managed in an Airtable base, where tools are linked to roles and users. Incoming requests via Telegram (or adaptable to WhatsApp/IMAP) trigger queries to Airtable using the user's Telegram username, retrieving allowed tools.

A LangChain Code node intelligently compares requested tools against permitted ones, replacing unauthorized tools with dummy nodes that respond with authorization denial messages. Authorized tools pass through unchanged to the main agent or sub-workflows. Responses return seamlessly to the Telegram conversation, maintaining a natural user experience.

Benefits include enhanced security for AI deployments, preventing unauthorized tool access without recoding agents. It's ideal for teams managing shared AI agents, dev tools, or enterprise automation, saving time on custom permission logic. Use cases: Secure multi-tenant AI services, internal tool gating, or compliant agent orchestration.

Setup is straightforward: Clone the workflow, configure OpenAI/Airtable/Telegram credentials, import the provided Airtable template, follow in-workflow instructions, and activate.