# CLIARE

CLIARE means CLI Agent Readiness Evaluation.

CLIARE audits command-line interfaces for agent readiness. It measures a released CLI binary as a black box, records runtime evidence, infers command shape, detects side effects, scores agent-readiness signals, and emits artifacts that maintainers and agent harnesses can use.

Canonical pages:

- Home: https://cliare.sh/
- Why CLIARE: https://cliare.sh/why-cliare.html
- GitHub: https://github.com/modiqo/cliare
- Crates.io: https://crates.io/crates/cliare

Core concepts:

- Agents can explore terminals, but exploration is not a design contract.
- Maintainers use CLIARE to improve help coverage, argument grammar, diagnostics, parseable output, precondition clarity, and safe discovery behavior.
- Harnesses use CLIARE artifacts such as command-index.json and shape.json as evidence-backed CLI surface inputs before routing commands.
- CLIARE treats help text as evidence, not truth. Runtime confirmation and bounded evidence carry stronger weight.
- Scores are deterministic for a model version and evidence set. The current model is cliare-score-v0 with experimental_partial status.

Recommended install:

```sh
curl -fsSL https://github.com/modiqo/cliare/releases/latest/download/install.sh | sh
cliare metadata --format text
```

Quick start:

```sh
cliare measure mycli --out .cliare/mycli --profile standard --refresh
cliare summary --out .cliare/mycli
cliare report maintainer --out .cliare/mycli --format markdown
cliare report harness --out .cliare/mycli --write
```