// ai security · monthly
AI red teaming and AI penetration testing that runs alongside your releases: one AI application, scheduled test rounds each month, and the changes you flag in time tested within the agreed scope before they reach your users.
A change to your AI application that you flag by email or in a shared channel before the next test round, for example:
Changes that are not flagged are not tested automatically, so flagging is what keeps the testing aligned with your releases.
Written to be handed straight to whoever asked for it: security questionnaires, SOC 2 or ISO 27001 evidence, ISO/IEC 42001 and NIST AI RMF work, procurement or investor review.
AI SaaS teams, agencies that build chatbots for their clients, and firms in finance, insurance and law: anyone whose AI talks to customers or touches their data, with developers but no one whose job is to attack it.
The application, the test-round cadence, the monthly testing capacity and the turnaround are fixed in a written quote before any work starts. The plan is invoiced monthly in advance and can be cancelled on one month’s notice.
Public, credited research in the same area as this service. All AI Infrastructure disclosures →
The Qwen2-VL and Qwen3-VL video samplers bound on the request-controlled max_frames value, which the server-side frame ceiling does not reach. One unauthenticated request to /tokenize exhausts API-server memory before admission control.
The embed handler passed a visitor-supplied username where a user object was expected, so the memory lookup collapsed to the single-user scope and the instance owner's private Memories were appended to the system prompt on an unauthenticated route.
// scope & quote
Describe your chatbot, agent or the component you care about most. A written plan with a fixed price, one-off or monthly, comes back before any work begins.
