Skip to main content

AI Agent Instructions

Instructions for AI coding agents (Claude Code, Cursor, etc.) to use the airs CLI programmatically.

The canonical file lives at AGENTS.md in the repository root. It is designed to be loaded into an AI agent's context window so the agent can operate the CLI autonomously.

How to Use​

Claude Code​

Add to your project's CLAUDE.md:

See AGENTS.md for instructions on using the `airs` CLI.

Or reference it directly — Claude Code automatically reads CLAUDE.md and can be pointed to AGENTS.md.

Other Agents​

Copy the contents of AGENTS.md into your agent's system prompt or context. The file is self-contained and requires no external references.

What's Covered​

The agent instructions document covers:

  • Authentication — which credentials are needed for which commands, how to verify them
  • Output formats — how to use --output json for machine-parseable results
  • Complete command reference — every command, flag, default, and required parameter
  • Common workflows — step-by-step recipes for scanning, CRUD, red teaming, auditing, guardrail optimization
  • Agent loop protocol — AGENTS.md documents the autonomous guardrail optimization loop
  • Error handling — common errors and fixes
  • Platform rules — critical AIRS behaviors agents must know (profile naming, propagation delays, etc.)
  • Config file — location, format, and field reference

Design Principles​

The instructions are written for agents, not humans:

  1. Structured for parsing — tables, code blocks, consistent formatting
  2. --output json emphasized — agents should always request JSON for programmatic parsing
  3. Non-interactive patterns — all commands documented with flags that bypass interactive prompts
  4. Error recovery — common failure modes and how to handle them
  5. Credential scoping — clear mapping of which credentials each command needs