AckuityRequest assessmentRESOURCES
Research and practical perspectives for teams building, deploying and securing AI agents.
AI AGENT SECURITY GUIDES
Agent security spans permissions, tool calls, data access and the sequence of actions an agent takes. Start with the question your team needs to answer. New to the category? Compare AI security and AI agent security.
PRACTICAL EVALUATION
Twelve checks for access, approvals, enforcement and downstream evidence. Includes a downloadable worksheet.
INSIDE THE PRODUCT
Eight screens for discovery, investigation, policies and audit context. See what each view helps your team evaluate.
MONITOR ACTIVITY
Read tool-call traces, connect multi-step attacks and distinguish an alert from an enforced block.
CONNECT CONTEXT
Inspect identity, intent, targets, tools and history with the Agent Security Context Graph.
VERIFY CONTROL
Evaluate authorized work, blocked actions, review and recovery against target-system evidence.
Rogue Pixel · Latest insight
Microsoft researchers edited seven neurons in a tested model to produce a planted command on a trigger. The model still passed its evaluations. The technique requires access to change model weights.
Source: Microsoft Security
Explore Rogue Pixel“Every test” refers to the evaluations used in the research, not all possible tests. Payload-output rates are not end-to-end attack success rates. The closing check is proposed; Ackuity was not involved.
FEATURED ARTICLE · INCIDENT ANALYSIS
A sourced account of the July 2026 incident, with confirmed impact, 23 techniques, a video timeline and lessons for AI, security and platform teams.
By Vinod Vasudevan · September 28, 2026
Read the featured articleHAVE YOUR SAY
Working with AI agents? Tell us what’s working, what worries you, and where you want more control.
Take the 4-minute surveyTRY THE CONTEXT CHALLENGE
Three decisions. One missing piece of context in each. Discover what changes when you see the whole picture.
ARTICLES
ARTICLE · 10 MIN READ
An AI agent incident response runbook for queued jobs, delegated agents, target-system evidence and safe restart. Includes a blank response worksheet.
ARTICLE · 10 MIN READ
Measure false positives in AI agent security with paired tests, clear denominators, task completion and review burden. Includes a blank evaluation worksheet.
ARTICLE · 9 MIN READ
Can an AI agent keep running after credential revocation? Check tokens, cached authorization, accepted jobs and delegated access with six proposed tests.
WHITEPAPERS
ACKUITY RESEARCH
RESEARCH PAPER · 8 PAGES
SECURITY RESEARCH
Explore four classes of threats and 21 threat vectors across data interaction, privilege misuse, agent manipulation and overreach.
A practical starting point for examining the risks that emerge when AI systems can act.
Explore the whitepaperVIDEOS
EXPLAINER
See why AI agent security needs action-level context, and how Ackuity’s Agent Execution Control Switch approaches it.
Watch video
EXPLAINER
When should you block one action, and when should you contain the agent?
Watch video
EXPLAINER
What changes when AI moves from answering questions to taking action?
Watch videoNEWS
COMPANY NEWS · OCTOBER 2, 2026
An integration with the OpenShell supervisor will let teams check every agent action against who asked, what they asked for and what the agent has done before, before the action leaves the sandbox
Read announcementCOMPANY NEWS · SEPTEMBER 24, 2026
Former S&P Global head of cybersecurity and technology risk will help Ackuity build for the security leaders who approve AI agents for production
Read announcementNEXT STEP
Start with visibility. Discuss the execution controls you need.
