# AI Agent Awareness (AX)

> Integrate coding agents and LLM scrapers safely with Kubeletto.

- **Category**: Platform & Reference
- **Last Verified**: 2026-07-14

Kubeletto is designed to be highly agent-friendly. We support machine-readable formats and agentic guidelines to assist AI coding partners.

## Agentic Capabilities

- **/llms.txt** - Clean overview of common operations.
- **/llms-full.txt** - Comprehensive reference of commands and schemas.
- **JSON Output** - Append `--output json` to parse responses programmatically.
- **Accept Header** - Send header `Accept: text/markdown` or fetch `/docs/[slug].md` to download raw Markdown pages.

## Agent Safety Rules

- Always pass `--yes` or `--non-interactive` to prevent blocking the terminal in automated agents.
- Never trigger destructive actions (`env delete`, `keys revoke`) without explicit confirmation.
- Run `kubeletto doctor [app] --output json` first when diagnosing failures.