Overview — Posture Overview
The Posture Overview is the first screen users see after logging in to Theseus Guard.
It provides a real-time summary of the platform's governance health across all active guards, agents, and policy lanes.
What is the Posture Overview?
The Posture Overview is a live command dashboard that aggregates key performance indicators (KPIs) from every active governance lane. It is designed to give operators, engineers, and compliance managers an instant read on the current state of AI safety across the tenant.
The dashboard refreshes continuously and surfaces critical metrics without requiring navigation into individual lanes.
Key Metrics & KPI Cards
The top row of the dashboard displays a set of summary cards, each representing a critical governance signal:
Guard Coverage
Shows the total number of active Guards currently deployed and protecting AI agent interactions. A lower-than-expected count may indicate agents running without governance coverage.
Drift Events (Last 24h)
Counts the number of drift events detected by the L1/L2/L3 engine stack within the past 24 hours. A spike here warrants investigation in the Detect → Live Event Stream or Detect → Anomaly Radar pages.
L3 Escalations
Displays the number of interactions that triggered the L3 (TG-Sentinel:Fast or Max Model) evaluation layer. L3 escalations indicate edge cases that exceeded L1 and L2 confidence thresholds.
Open Runtime Reviews
Shows pending decisions in the Route → Runtime Reviews queue that require human operator approval. This metric has direct SLA implications — operators should prioritize resolving open reviews.
Open Alerts
The count of unresolved alerts across all lanes. Alerts are generated by system-level anomalies such as probe degradation, L2/L3 divergence, or department format mismatches. Navigate to Prove → Alert Center to manage them.
Governance Health Index (GHI)
A single executive metric (0–100) combining guard coverage, alert health, review status, and drift trends. The Governance Health Index represents the overall AI safety and compliance status of the tenant in real time.
Dashboard Sections
Drift Trend Chart
A time-series chart showing drift event volume over the past 7 or 30 days. Useful for identifying patterns, spikes after agent deployments, or gradual degradation of model behavior.
Top Violating Guards
A ranked list of Guards that have generated the most drift events or escalations. This helps prioritize which Guards require tuning or policy updates.
Recent Alerts
A summary of the most recently created alerts, with severity and status. Clicking an alert navigates directly to the Alert Center detail view.
Runtime Review Queue Preview
A snapshot of the top items in the Runtime Reviews queue, including SLA time remaining. This ensures that critical decisions are not missed even when the operator hasn't navigated to the Route lane.
Who This Is For
| Role | Use |
|---|---|
| Admin | Overall platform health at a glance |
| Approver | Monitor open reviews and alert count |
| Compliance Manager | Track Governance Health Index (GHI) and escalation rates |
| Engineer | Observe drift trends after deploying or updating agents |
| Auditor | Read-only view of platform posture |
Related Pages
- Detect → Live Event Stream — Drill into individual events driving drift metrics
- Detect → Anomaly Radar — Visual scatter plot of L3-triggered anomalies
- Route → Runtime Reviews — Resolve open review queue items
- Prove → Alert Center — Manage and resolve open alerts
- Admin → Policy Compliance — Per-agent compliance status and grace period tracking