---
title: The agent-native standard
description: 'One bar for agent-readiness, measured on two surfaces: your CLI and your website. Scored MUST / SHOULD / MAY, not asserted.'
url: https://anc.dev/
dateModified: '2026-11-15'
---

# The agent-native standard

One bar for agent-readiness, measured on two surfaces: your CLI and your website. Scored MUST / SHOULD / MAY, not asserted.

To audit a CLI, install the linter (`anc`) locally: see [/install](https://anc.dev/install). To audit a website and its MCP server, use
[/audit](https://anc.dev/audit) or the `audit_website` MCP tool. To install the agent-native-cli **skill bundle** (the Claude Code /
Codex / Cursor / OpenCode skill), see [/skill](https://anc.dev/skill). Agents that speak MCP can connect directly to the
streamable-HTTP server at [https://anc.dev/mcp](https://anc.dev/mcp) (`GET` returns 405; clients `POST`). See
[/mcp-docs](https://anc.dev/mcp-docs) for the endpoint description and [/mcp-skill](https://anc.dev/mcp-skill) for the wire contract and tool
catalog.

## CLI leaderboard

| # | Tool | Score |
|---|------|-------|
| 1 | [anc](https://anc.dev/score/anc) | 100% |
| 2 | [bird](https://anc.dev/score/bird) | 99% |
| 3 | [xr](https://anc.dev/score/xr) | 96% |
| 4 | [watchexec](https://anc.dev/score/watchexec) | 86% |
| 5 | [fd](https://anc.dev/score/fd) | 85% |

## Web leaderboard

| # | Site | Score |
|---|------|-------|
| 1 | [anc.dev (anc)](/score/anc.dev) | 96% |
| 2 | [modelcontextprotocol.io (Model Context Protocol)](/score/modelcontextprotocol.io) | 94% |
| 3 | [docs.box.com (Box Docs)](/score/docs.box.com) | 94% |
| 4 | [supabase.com (Supabase)](/score/supabase.com) | 90% |
| 5 | [stripe.dev (Stripe Developers)](/score/stripe.dev) | 90% |


## Audit a CLI tool or a website, live.

Pick CLI or Website on the homepage form or the [audit page](https://anc.dev/audit) and enter a target. A CLI target is a tool name, an install command, or a GitHub URL: `ripgrep`, `cargo binstall ouch`, `npm install -g cowsay`, `pip install black`, `uv tool install rclone`, `github.com/cli/cli`. A website target is a domain or a URL: `anc.dev`, `modelcontextprotocol.io`. [Install `anc` locally](https://anc.dev/install) for source and project depth.

## Principles

- [P1: Non-Interactive by Default](https://anc.dev/p1) — Every automation path MUST run without human input. A CLI tool that blocks on an interactive prompt is invisible to an agent: the agent hangs, the user sees nothing, and the operation times out silently.
- [P2: Structured, Parseable Output](https://anc.dev/p2) — CLI tools MUST separate data from diagnostics and offer machine-readable output formats. Mixing status messages with data forces agents into fragile regex extraction that breaks on any format change.
- [P3: Progressive Help Discovery](https://anc.dev/p3) — Help text MUST be layered so agents (and humans) can drill from a short summary to concrete usage examples without reading the entire manual. The critical layer is the one that appears after the flags list, because that is where readers look for invocation patterns.
- [P4: Fail Fast with Actionable Errors](https://anc.dev/p4) — CLI tools MUST detect invalid state early, exit with a structured error, and tell the caller three things: what failed, why, and what to do next. An error that says "operation failed" gives an agent nothing to act on.
- [P5: Safe Retries and Explicit Mutation Boundaries](https://anc.dev/p5) — Every CLI with write operations MUST support --dry-run so agents can preview a mutation before committing it. Commands MUST make the read-vs-write distinction visible from name and --help alone, and destructive writes MUST require explicit confirmation. An agent that cannot distinguish a safe read from a dangerous write will either avoid the tool or execute mutations blindly. Both are failure modes.
- [P6: Composable and Predictable Command Structure](https://anc.dev/p6) — CLI tools MUST integrate cleanly with pipes, scripts, and other tools. That means handling SIGPIPE, detecting TTY for color and formatting decisions, supporting stdin for piped input, and maintaining a consistent, predictable subcommand structure.
- [P7: Bounded, High-Signal Responses](https://anc.dev/p7) — CLI tools MUST provide mechanisms to control output volume. Agent context windows are finite and expensive: a tool that dumps tens of thousands of lines of unfiltered output wastes tokens on every request and can exceed smaller context windows entirely, breaking the conversation that invoked it. "High-signal" here means the bytes that survive --quiet are the ones the caller asked for (data and errors), not progress, decoration, or chatter.
- [P8: Discoverable Through Agent Skill Bundles](https://anc.dev/p8) — Without a skill bundle, every fresh agent invocation begins the same way: pull --help, infer the idioms, try a command, parse the error, try again. A skill bundle (canonical names AGENTS.md or SKILL.md) collapses that loop: agent-discoverable through filesystem convention rather than through --help, loaded once, recognized thereafter.
