TrustYourAgent
Test Your AI System
Resources

TrustYourAgent Trust Intelligence

Guides, research and product updates on certifying and monitoring AI agents.

Upcoming events →

Latest Intelligence

Bastyn Scan: open-source command-line scanner for AI-agent code, available on GitHub at github.com/bastyn-labs/bastyn-scan
Engineering

93% precision and the fewest false alarms: Bastyn Scan against seven other scanners on AI-agent code

Bastyn Scan had the highest precision (93%) on five deliberately vulnerable AI-agent projects and raised the fewest false alarms on five real ones. We compared it with seven other scanners on 181 known vulnerabilities.

BASTYN team · 28 Sept 2026 · 21 min read

Blogs

team reviewing AI agent governance alongside human workforce policy
Blog

Agentic vs human processes

This piece works through the similarities and differences between managing digital and human workforces, what transfers to an agentic AI workforce, what needs strengthening, and what has no human precedent at all, drawing on how BNY Mellon and Microsoft are governing agents in production today.

11 Aug 2026 · 3 min read

Illustration showing provider and deployer responsibilities under the EU AI Act and the vendor accountability gap for AI systems.
Blog

The Vendor Accountability Gap explained | EU AI Act obligations

The EU AI Act and new AI regulations makes AI deployers legally responsible for monitoring, oversight and incident reporting, even when the AI model comes from a third-party vendor. Accountability rules and fines have been updated

6 Aug 2026 · 4 min read

Illustration of AI credit decisioning agents analysing applicant data, affordability, fraud, identity, credit history and risk signals to produce automated loan approvals or declines with continuous agent trust monitoring by BASTYN.
Blog

No rules broken but loan book reshaped.

AI agents are beginning to transform loan and credit decisioning. Discover the hidden agentic risks across data, policy, permissions, timing and multi-agent workflows, and why continuous assurance is becoming essential.

20 Jul 2026 · 5 min read

Guides

Global AI Regulations and Standards Registry
Guide

Introducing BASTYN Global AI Regulations and Standards Registry

The first public release of the BASTYN Global AI Regulations and Standards Registry brings together AI obligations, standards and assurance frameworks that are in operation across major markets. Standards are mapped to BASTYN coverage.

24 Sept 2026 · 2 min read

EU AI Act compliance checklist for AI agent builders - visual guide covering classification, obligations, and deployer requirements
Guide

The EU AI Act: a practical checklist for agent builders

Cut through the legalese. A plain-English checklist of what the EU AI Act actually asks of teams shipping AI agents.

5 Jun 2026 · 9 min read

Research

Abstract diagram showing an AI agent gradually drifting away from its approved behavioural path as its data, memory, tools, permissions, goals and actions change.
Research

Behavioural drift: the silent way agents fail in production

An agent that passed every test on launch day can quietly degrade over time. We dig into what drift is, why drift happens and how to catch it early.

28 May 2026 · 7 min read

More resources

Bastyn Scan: open-source command-line scanner for AI-agent code, available on GitHub at github.com/bastyn-labs/bastyn-scan
Engineering

93% precision and the fewest false alarms: Bastyn Scan against seven other scanners on AI-agent code

Bastyn Scan had the highest precision (93%) on five deliberately vulnerable AI-agent projects and raised the fewest false alarms on five real ones. We compared it with seven other scanners on 181 known vulnerabilities.

28 Sept 2026 · 21 min read

Subscribe for latest BASTYN Trust Intelligence

For risk, compliance, product and security experts