@prave/cli

v1.2.1

Published

2 days ago

Prave CLI — discover, install, version, test, and ship Claude Skills. The developer platform for the complete Skill lifecycle.

`@prave/cli`

The Prave CLI — discover, install, author, version, test, and ship Skills to any AI agent from your terminal.

Website · Documentation · CLI Cheat Sheet · Status

What is Prave?

Prave is the developer platform for the complete lifecycle of AI agent Skills — a registry to discover them, an editor to write them, an intelligence layer to keep your library lean, a tester to verify them, and this CLI to wire it all into your agent's skills directory on every machine you work from.

The Skill format originated with Claude Code (a folder with a SKILL.md and frontmatter). The same format works out of the box with any agent that auto-loads instruction files from a folder — Prave is agent-aware: each Skill is tagged with which agents understand it, and prave deploy mirrors a Skill across Claude Code · OpenAI Codex · Cursor · Gemini CLI · Cline · Amp, with more added as the ecosystem grows.

This package is the command-line surface. The web app, the public registry, billing, and the API live at prave.app.

Install

npm install -g @prave/cli
# or
pnpm add -g @prave/cli
# or
yarn global add @prave/cli

Requires Node.js 18+.

Quick start

prave login              # device-code auth in your browser
prave search "review pull requests"
prave install pr-reviewer    # writes into your agent's skills folder
prave deploy pr-reviewer     # mirror it to every agent you use
prave list                   # what's now installed locally

That's enough to be useful. The full vocabulary is below.

Multi-agent by default. prave install writes to your primary agent's skills folder (Claude Code by default; configurable via prave settings). prave deploy then mirrors the Skill to every other agent you've enabled — Cursor, Codex, Gemini, Cline, Amp — converting between their respective rule-file formats automatically.

`prave sync` in practice

prave sync shows you when you last ran it and confirms before doing any work — under 30 minutes ago and the default flips to n. Skills then install in parallel (concurrency 4) with a per-Skill progress counter:

$ prave sync
Last sync: 2 hours ago — sync again? [Y/n] y
Syncing 12 Skills — this takes about 15 seconds.
✓ 12 / 12 — pr-reviewer
Synced 12 · failed 0

The last_sync_at watermark persists at ~/.prave/state.json. Pass --yes to skip the prompt in CI scripts.

Commands

Auth & account

| Command | What it does | | -------------- | --------------------------------------------------------------------------------- | | prave login | Browser device-code auth. Tokens stored chmod-600 at ~/.prave/credentials.json. | | prave logout | Forget local credentials. | | prave whoami | Show signed-in user, plan, and credentials path. |

Discover & install

| Command | What it does | | -------------------------------------------------------- | --------------------------------------------- | | prave search <query> | Search the public Skill registry. | | prave install <slug> [--no-deps] | Install into your agent's skills folder. | | prave uninstall <slug> | Remove a locally installed Skill. | | prave list [--remote] [--verbose] | What's installed locally (or remote). | | prave export <slug> -o file.md | Dump a Skill's SKILL.md without installing. |

Authoring & sync

| Command | What it does | | ----------------------------------------------- | -------------------------------------------------------------------------------------- | | prave import [--upload --public\|--private] | Scan your agent's skills folder, optionally publish to Prave. | | prave deploy <slug> [--agent <name>] | Mirror a Skill across enabled agents (Claude / Cursor / Codex / Gemini / Cline / Amp). | | prave sync | Pull updates for every locally installed Skill. | | prave update [<slug>] [--dry-run] | Diff installed Skills against the registry, pull what's outdated. | | prave diff <slug> | Local vs registry side-by-side diff. |

Intelligence

| Command | What it does | | ------------------------------------ | ----------------------------------------------------- | | prave overview [--json] | Summary of token cost, conflicts, and library health. | | prave whatdoes <slug> | Triggers, tokens, conflicts for a Skill. | | prave conflicts | Cross-Skill collision check. | | prave optimize [--remove-unused] | Recommendations: heavy / underused / mergeable. |

Usage tracking (Pro+)

| Command | What it does | | --------------------------------- | --------------------------------------------------------- | | prave usage hook install | Real-time invocation tracking for agents that support it. | | prave usage hook uninstall | Remove the real-time hook. | | prave usage scan [--since 7d] | Transcript scanner — backfill recent invocations. | | prave usage status | Hook health, recent counts, and top Skills. |

Coverage today: Real-time tracking lives natively on Claude Code via its plugin contract. For every other agent, the transcript scanner auto-runs at the tail of prave sync and reads the agent's local conversation history to extract Skill invocations — so Intelligence and the optimiser stay accurate everywhere. Real-time hooks for Cursor / Codex / Gemini / Cline / Amp ship as soon as those agents expose an equivalent contract.

Settings

| Command | What it does | | ---------------- | ----------------------------------------------------------- | | prave settings | Configure agents, providers, and credits from the terminal. | | prave cheat | Print the full one-page command reference. |

Auto-update

prave login retries on 401 once with a fresh access token swapped via your stored refresh_token. You won't see "expired token" mid-install — re-authenticating is only required when the refresh token itself is rejected (you signed out everywhere or were idle for 30+ days).

Privacy & telemetry

The CLI sends a fire-and-forget event per command (e.g. cli_install, cli_sync, cli_login_started) to PostHog EU Cloud so we can see which commands matter to users. Events carry your Supabase user-id once you've logged in (or a stable per-machine UUID at ~/.prave/anon-id if you haven't), plus OS family, OS release, Node.js version, and the CLI version — no command arguments, no Skill content, no file paths.

Opt out by setting:

PRAVE_TELEMETRY=0

(also accepts false, off, no). When set, no event ever leaves your machine.

Full data-protection details: Privacy Policy.

Where things live

The exact Skill location depends on the agent — prave settings lets you reconfigure per-agent paths, and prave list always reflects what is actually on disk. The defaults:

Claude Code   ~/.claude/skills/<slug>/SKILL.md
Cursor        ~/.cursor/rules/<slug>.mdc
Codex         ~/.codex/skills/<slug>/SKILL.md
Gemini CLI    ~/.gemini/skills/<slug>/SKILL.md
Cline         ~/.cline/rules/<slug>.md
Amp           ~/.amp/skills/<slug>/SKILL.md

Local Prave state:

~/.prave/credentials.json          # access + refresh tokens (chmod 600)
~/.prave/anon-id                   # anonymous telemetry id (pre-login)
~/.prave/usage-cursor.json         # transcript scanner watermark

Status

API health and uptime: status.prave.app — auto-refreshes every 30 seconds, shows the live build version + commit SHA, and surfaces incidents from the last 30 days.