Security
Agent Action Pentest
We attack the real actions your AI agent can take, and deliver a written report within 72 hours of access.
- Written report, every finding with a severity
- Steps to reproduce it
$2,500
Buy, $2,500 DetailsThe P(rogrammable)law Company
We make AI that can be trusted with real work.
Making AI smarter doesn’t make it accountable. We trust people with work because there are rules, records, and someone to answer to. AI needs the same.
Rules are specific to the job. So we work one industry at a time, starting with the trades.
We have spent more than a year on this problem: how to trust autonomous software. Previously, VulnZap checked the security of AI-written code for hundreds of engineers. Veto, now open source, checks each action an AI agent takes against a written policy and holds the consequential ones until a person approves.
Most companies today build user-facing agents first and add controls later. We built the controls first. Guddun is built on them. We also signed the EU AI Pact before AI Act enforcement began.
Not sure which? Book 15 min with Yaz. Invoice same day. 50% deposit available.
Security
We attack the real actions your AI agent can take, and deliver a written report within 72 hours of access.
$2,500
Buy, $2,500 DetailsSecurity
We attack a Lovable, Bolt, or Cursor app, fix the Critical and High issues, and re-test.
$1,500
Buy, $1,500 DetailsSecurity
Risky MCP calls wait for approval or stop. We review your tools and leave the guard rules.
$1,500
Buy, $1,500 DetailsAgent guardrails
Refunds over your cap wait for a human. Smaller ones run.
$2,000
Buy, $2,000 DetailsAgent guardrails
Disclosure rules that fire when a person must be told.
$750
Buy, $750 DetailsCommerce
Your checkout breaks on the one weekend that matters. We check scripts, apps, speed and payment errors, then ship the fixes.
Book by Oct 31 to be live before Black Friday.
$1,500
Buy, $1,500 DetailsBook 20 minutes with a founder [email protected] +1 (628) 296-7377