Security · CISO
Govern every agent's behavior, posture, and credentials — and prove it.
Your seat answers one question: across every AI agent running in the org, what is unsafe right now and what needs a human? Start here, then drill where the numbers point.
Bottom line
Control coverage 63% (5/8 agents governed & monitored). 2 agents are compound risk — Critical & ungoverned. 2 high findings and 1 shadow agent are waiting on you. Posture 76, +2 this week → 80 in ~2 weeks at this slope.
63%
control coverage
5/8 governed & monitored
1
critical agents
posture < 50
2
open high findings
need a human now
2
compound risk
Critical & ungoverned
Fleet trajectory
6-week posture index · target ≥ 80Posture 76, +2 this week → 80 in ~2 weeks at this slope.
Your surfaces
Every agent as one row — what it is, does, costs, how safe.
The worklist of agent behavior that needs a human.
How safe the fleet is, what lowers the score, where it heads.
Each agent's machine credential and how trustworthy it is.
Which sensitive files agents reached into, and how much moved.
What an enforcing policy would have blocked — a dry run.
Derived: controlCoverage · compoundRiskAgents · postureIndexTrend · weeksToTarget — from app/data/c16-fleet.ts.