Launching today
BlvkWare Agent Kits
Every file your AI agent needs, designed from your answers
1 follower
Every file your AI agent needs, designed from your answers
1 follower
Pick the job an agent should own. Answer a few questions. Get its instructions, tools (OpenAI, Anthropic, MCP), workflows, records, tests and a Python runtime that enforces its rules in code. Every file shown before you pay. $79 or $199, one-off.
Hi PH. I build AI agents for small businesses, and the part nobody wants to pay an agency for is the specification: what the agent is allowed to do, in what order, what it must never do, and how you know it works.
So that is what a kit is. You pick the job (chasing quotes, answering the inbox, booking, collections: 23 of them), tell it about your business and the rules it must never break, choose which of your systems it touches, and set how much it may do alone. It starts at Draft: anything a customer would see waits for a person.
The kit page lists every file your kit will contain, with the price, before you pay. If you want to read one first, there is a complete sample kit here, every file, free: https://blvkware.dev/sample-kit/
The part I care most about: the rules are not just asked for in a prompt. Every kit ships a small Python runtime (nothing to install) that checks each action before it runs, holds anything customer-facing in an approval queue with its own approvals page, keeps a tamper-evident log, and honours opt-outs and quiet hours. It also runs as an MCP server, so you can try the agent in Claude Desktop.
What it is not: hosted software. You keep your own model account and data, and wire it into OpenAI or Anthropic tool calling, any MCP client, or n8n, Make and Zapier.
One person built this. I would genuinely like to know which job you would give an agent first.