QVeris CLI Skill

Why CLI Skill When MCP Skill Already Exists?#

The qveris-cli Skill complements the existing qveris MCP Skill, offering distinct advantages for certain scenarios:

1. Process Isolation & Stability#

CLI runs as a separate process. If a tool call fails or encounters corrupt data, the fault is confined to the CLI process without affecting the host agent's main session.

2. Low Prompt Tokens#

Unlike MCP which injects tool schemas into every LLM prompt (consuming 200-500 tokens per tool), CLI runs as a subprocess — no schema definitions are added to the prompt, resulting in significantly lower prompt token usage.

3. No MCP Dependency#

• Works in environments without MCP server setup
• Useful for quick prototyping, debugging, or environments where MCP isn't available
• Simplifies CI/CD pipelines — just install the CLI binary

4. Scriptable & Automatable#

CLI commands can be:

• Embedded in shell scripts
• Chained with other Unix tools (|, grep, jq)
• Scheduled via cron or integrated into build systems

5. Transparent Debugging#

• See exact command being executed
• Easily replay commands for troubleshooting
• --dry-run validates params without consuming credits

Installation#

One-liner (recommended)#

curl -fsSL https://qveris.ai/cli/install | bash

npm#

npm install -g @qverisai/cli

npx (no install)#

npx @qverisai/cli discover "weather API"

Requirements: Node.js 18+

Quick Start#

The examples below omit --json for readability. Add --json when scripting or using with agents.

# 1. Authenticate (saves key to ~/.config/qveris/config.json)
qveris login
 
# 2. Discover tools
qveris discover "weather forecast API"
 
# 3. Inspect a tool (use index from discover results)
qveris inspect 1
 
# 4. Call it
qveris call 1 --params '{"wfo": "LWX", "x": 90, "y": 90}'

API Endpoint Override#

The CLI uses its built-in API endpoint unless you explicitly pass --base-url or set QVERIS_BASE_URL. API keys do not select or replace the endpoint.

qveris discover "weather forecast API" --base-url "$QVERIS_BASE_URL"

Use the complete HTTP(S) API root supplied by the active deployment. Resolution order is --base-url > QVERIS_BASE_URL > built-in default.

Complete Workflow Example#

# Discover weather tools
qveris discover "weather forecast API" --json --limit 5
 
# Inspect first result
qveris inspect 1 --json
 
# Validate params without consuming credits
qveris call 1 --params '{"wfo": "BOU", "x": 50, "y": 30}' --dry-run --json
 
# Execute the call
qveris call 1 --params '{"wfo": "BOU", "x": 50, "y": 30}' --json
 
# Generate production code snippet (only produced on successful calls)
qveris call 1 --params '{"wfo": "BOU", "x": 50, "y": 30}' --codegen curl
qveris call 1 --params '{"wfo": "BOU", "x": 50, "y": 30}' --codegen python

Always use --json for structured output in scripts.

Interactive Mode#

For hands-on exploration, launch a REPL session that keeps discovery state in memory:

qveris interactive

qveris> discover "crypto price API"
qveris> inspect 1
qveris> call 1 {"symbol": "BTC"}
qveris> codegen python
qveris> exit

Available REPL commands: discover, inspect, call, codegen <curl|js|python>, history, help, exit.

When to Use Which#

Troubleshooting#

Authentication Error (Exit Code 77)#

# Check auth status
qveris whoami
 
# Re-login
qveris login

Missing API Key (Exit Code 78)#

# Check config location
qveris config path
 
# Set key directly
qveris login --token "sk-xxx"

Service Unavailable (Exit Code 69)#

# Run diagnostics
qveris doctor
 
# Check the effective API endpoint
qveris whoami

Session Expired#

Sessions expire after 30 minutes. Re-run discover to refresh:

qveris history          # View session state
qveris history --clear  # Reset session

Exit Codes#

Sources#

• The Advantages of Command Line Interfaces Over SDKs and DLLs
• Why Developers Love API and CLI Tools
• The Return of the CLI
• CLI vs API: What's the Difference?

Official Documentation#

• CLI Reference
• REST API
• MCP Server
• Get API Key