Skill

Fail

Audit score 45

belt

belt-sh/cli

Cloud platform CLI for AI agents: run 250+ apps, manage knowledge, search skills, and connect MCP servers.

What is belt?

belt is a lightweight (~4MB) CLI that gives AI agents access to a cloud platform with 250+ AI models and applications, persistent knowledge management, reusable skill workflows, and MCP server connectors. Use it to run image generation, video, search, audio, and other AI tasks; store and retrieve agent memory; discover and install skills; and integrate external tools.

Run 250+ AI apps (image generation, upscaling, video, search, audio, and more) with schema discovery and sample input generation
Search and manage persistent agent knowledge with semantic search across concepts, observations, references, preferences, and skills
Discover, install, and use reusable skill workflows from a registry or GitHub repositories
Connect and run MCP server tools (Slack, etc.) with unified command interface
Search across apps, skills, and knowledge simultaneously with `belt suggest`
Export structured output and save media files directly with `--json` and `--save` flags

How to install belt

npx skills add null --skill belt

Prerequisites

curl (for installation)
Valid belt account (requires `belt login` after install)

Claude Code

Cursor

Windsurf

Cline

How to use belt

1.Install belt: `curl -fsSL cli.inference.sh | sh`
2.Authenticate: `belt login` and verify with `belt me`
3.Search for an app: `belt app search "image"` or `belt suggest "your task"`
4.View app schema: `belt app get namespace/app-name`
5.Generate sample input: `belt app sample namespace/app-name --save in.json`
6.Run the app: `belt app run namespace/app-name --input in.json --save output.ext`
7.Optionally install skills: `belt skill add namespace/skill-name`
8.Optionally connect MCP servers: `belt mcp connect service-name`

Use cases

Good for

Generate images or upscale/edit them using OpenAI, Reve, or Pruna models
Create videos with Google Veo-2 or Seedance models
Perform web searches with Tavily or Exa integrations
Synthesize speech with ElevenLabs TTS
Store and retrieve learned concepts, observations, and preferences for agent memory

Who it's for

AI agent developers (Claude Code, Cursor users)
Automation engineers building multi-step AI workflows
Teams needing persistent agent memory and knowledge management
Developers integrating multiple AI models and external services

belt FAQ

How do I find which app to use for my task?

Use `belt app search "keyword"` to find apps by capability, or `belt suggest "what you want to do"` to search across apps, skills, and knowledge simultaneously.

How do I save output from an app?

Use the `--save filename` flag when running an app, e.g., `belt app run openai/gpt-image-2 --input '{...}' --save output.png`.

Can I use belt skills without installing them?

Yes, use `belt skill use namespace/skill-name` to run a skill on-demand without installation. Use `belt skill add` to install persistently.

What types of knowledge can I store?

belt supports five knowledge types: skill, concept, observation, reference, and preference. Create them from files or stdin with `belt know create`.

How do I connect external services like Slack?

Use `belt mcp connect service-name` to connect an MCP server, then list available tools with `belt mcp tools service-name` and run them with `belt mcp run`.

Full instructions (SKILL.md)

Source of truth, from belt-sh/cli.

name: belt description: "Use the belt CLI — run 250+ AI apps, manage knowledge, search skills, connect MCP servers" allowed-tools: Bash(belt *)

belt cli

belt is the cloud platform cli for ai agents. single ~4mb binary, no runtime dependencies.

install

curl -fsSL cli.inference.sh | sh
belt login
belt me

apps — run 250+ ai models

belt app search "image"                      # find apps
belt app get openai/gpt-image-2              # view schema
belt app sample openai/gpt-image-2 --save in.json  # generate sample input
belt app run openai/gpt-image-2 --input in.json    # run it
belt app run openai/gpt-image-2 --input '{"prompt": "..."}' --save output.png

common apps:

image: openai/gpt-image-2, reve/create, pruna/p-image
upscale/edit: pruna/p-image-upscale, pruna/p-image-edit
video: google/veo-2, seedance/seedance-2-i2v
search: tavily/search, exa/search
audio: elevenlabs/tts

knowledge — persistent agent memory

belt know search "query"                  # semantic search
belt know list --type observation         # list by type
belt know get namespace/name              # get details
belt know create ./file.md --type concept # create from file
echo "learned X" | belt know create - --name x --type observation  # from stdin
belt know delete <id>

types: skill, concept, observation, reference, preference

skills — reusable workflows

belt skill search "deployment"            # search registry
belt skill store --featured               # browse featured

# use on-demand (no install, stdout)
belt skill use namespace/skill-name       # from store
belt skill use github.com/user/repo       # from github
belt skill use user/repo --skill name     # pick from multi-skill repo

# install persistently
belt skill add namespace/skill-name       # auto-detects agents
belt skill add ns/name --agent claude-code
belt skill list                           # list installed

belt skill upload ./my-skill              # publish

connectors — mcp servers

belt mcp list                             # available connectors
belt mcp search "slack"                   # search
belt mcp connect slack                    # connect
belt mcp tools slack                      # list tools
belt mcp run slack send_message --input '{"channel": "#general", "text": "hello"}'

suggest — unified search

belt suggest "how to generate images"     # searches apps + skills + knowledge

tips

use --json for structured output when piping
use --save filename to save media outputs directly
belt app sample generates valid input — start there for unfamiliar apps
belt update to get latest version

Related skills

More from belt-sh/cli and the wider catalog.

adaptive-communication

bencium/bencium-marketplace

Detect communication style and adapt responses to match user intent—relational or transactional.

1.4k installsAudited

Agentic UX Design - Relationship-Centric Interfaces

bencium/bencium-marketplace

Design AI-first interfaces that build lasting relationships through memory, trust evolution, and collaborative planning.

1.4k installsAudited

bencium-aeo

bencium/bencium-marketplace

Optimize content for AI citations and answer engine visibility across ChatGPT, Claude, Gemini, and AI Overviews.

1.4k installs

bencium-code-conventions

bencium/bencium-marketplace

Bence's opinionated code style, tech stack, and workflow conventions for modern web development.

1.4k installsAudited

bencium-controlled-ux-designer

bencium/bencium-marketplace

Expert UI/UX design guidance that prioritizes user collaboration and accessible, distinctive interfaces.

1.9k installsAudited

bencium-impact-designer

bencium/bencium-marketplace

Create distinctive, production-grade frontend interfaces with bold design choices that avoid generic AI aesthetics.

1.8k installsAudited