Agent Toolkit

Agent Toolkit

Machine-readable contracts for building AI agents on top of kraken-cli.

Source of Truth

tool-catalog.json is the canonical command contract.

• Coverage: 151 commands with parameters, types, defaults, and examples
• Every command includes: group, description, auth requirement, dangerous flag, and parameter schemas
• 34 commands are marked dangerous: true (orders, withdrawals, transfers, cancel-all, staking)

error-catalog.json is the error routing contract.

• 9 error categories with retry/backoff guidance

Contents

File	Purpose
tool-catalog.json	Canonical command catalog (151 commands)
error-catalog.json	Error categories with retry guidance
examples/	Runnable shell examples

Using the Catalog

Load commands in Python

import json

with open("agents/tool-catalog.json") as f:
    catalog = json.load(f)

for cmd in catalog["commands"]:
    print(f"{cmd['group']}/{cmd['name']}: {cmd['description']}")
    if cmd.get("dangerous"):
        print("  ⚠ requires human confirmation")

Build tool definitions for any platform

The catalog contains everything needed to generate tool schemas for OpenAI, Anthropic, MCP, LangChain, or any other framework. Each command entry includes:

{
  "name": "order-buy",
  "group": "trade",
  "command": "kraken order buy <PAIR> <VOLUME>",
  "description": "Place a buy order.",
  "auth_required": true,
  "dangerous": true,
  "parameters": [
    { "name": "PAIR", "type": "string", "required": true, "positional": true },
    { "name": "--type", "type": "string", "required": false }
  ],
  "example": "kraken order buy BTCUSD 0.001 --type limit --price 50000 --validate"
}

Filter by safety

safe_commands = [c for c in catalog["commands"] if not c.get("dangerous")]
dangerous_commands = [c for c in catalog["commands"] if c.get("dangerous")]

Related Files

• ../CONTEXT.md: Runtime context for tool-using agents
• ../AGENTS.md: Full integration guide
• ../skills/: Goal-oriented workflow skills
• ../gemini-extension.json: Gemini CLI extension manifest