Dashboard
LIVEReal-time overview of Acme Corp's AI agent fleet.
⚠
3 agents missing heartbeat — no check-in in the last 24 hours
All three were discovered as shadow AI and remain unassessed. Silence isn't nothing — silence is a signal.
Email Classifier · 26h
Document Summarizer · 31h
Sales Assistant · 2d
0
Total agents
5 DISCOVERED AS SHADOW AI
0
High risk
2 ANNEX III DETERMINATIONS
$0
Monthly cost
LAST 30 DAYS · ALL PROVIDERS
0
Open incidents
2 × P1 · ART. 73 CLOCK RUNNING
Risk distribution 12 AGENTS
12AGENTS
High risk3
Limited / at risk4
Minimal / compliant5
Cost — last 30 days $18,427 TOTAL
JUN 8JUN 18JUN 28JUL 7
Top agents by cost 30 DAYS
Recent incidents
Agents assessed
58%
7 of 12 agents under at least one framework
Remediation
21/32
66% of tasks completed — every one owned and dated
Unassessed production agents
⚠ Email Classifier
⚠ Document Summarizer
⚠ Sales Assistant
Framework coverage AGENTS ASSESSED PER FRAMEWORK
This is real data from a real product.
Sign up and see your own AI agent fleet in minutes — the scanner does the rest.
Agents
REGISTRYOne system of record for every AI agent at Acme Corp — including the five discovered as shadow AI.
| AGENT | ENV | OWNER | RISK | EU AI ACT | HEARTBEAT | COST/30D |
|---|
Compliance
Every agent scored against six frameworks. Click any agent row in the matrix to open its profile.
Compliance matrix SCORE % · — = NOT ASSESSED
Remediation plan 21 OF 32 COMPLETE · CLICK TO TOGGLE
EU AI Act
Regulation (EU) 2024/1689 — Acme Corp's posture against the framework that carries the fine.
FULL ENFORCEMENT · AUGUST 2, 2026
--D --H --M --S
Maximum penalty: €35M or 7% of global revenue — whichever is higher. High-risk systems must satisfy Articles 9–15 before the deadline. Two Acme agents carry Annex III determinations.
Annex III determinations
HR Screening AgentHIGH RISK — ANNEX III §4(a)
Fraud Detection BotHIGH RISK — ANNEX III §5(b)
Data Pipeline AgentHIGH RISK — INTERNAL POLICY
→ Annex III §4(a): AI used in recruitment or selection of natural persons is high-risk by definition. §5(b): AI evaluating creditworthiness / essential services. Articles 9–15 obligations apply to both.
Article obligations — high-risk agents
Discovery
The CLI scanner sweeps apps, SDK packages, API keys, VS Code extensions, network egress, and local inference. Run it — this is the moment shadow AI stops being invisible.
acme-corp — agentpmo scan · bash
Press ▶ Run scan to sweep the estate…
12
agents discovered
5
shadow — no owner, no record
Detection surfaces
SDK packages · 7 repos
API keys · 4 active
VS Code ext. · 41 machines
Network egress · 2 endpoints
Ollama · dev-ml-02
LM Studio · —
llama.cpp · —
Incidents
2 × P1 OPENP1–P4 incident management with automatic routing and Article 73 evidence capture. Click an incident for its timeline.
| SEV | INCIDENT | AGENT | DETECTED | STATUS |
|---|
Token Charter
No agent deploys without a signed charter: an expected monthly spend, a stretch threshold that alerts, and a runaway threshold that trips the circuit breaker. The cap is set at registration — not discovered in a post-mortem.
Charters — current cycle ALERT AT STRETCH · CIRCUIT BREAKER AT RUNAWAY · HUMAN AUTH TO RESUME
Dependencies
Which agents break when a provider, model, or key changes. Single points of failure are flagged.
Provider & model dependency map
Reports
CISO-ready PDFs and regulator evidence packages, assembled from live data. What used to take a quarter now takes one click — try it.
EU AI Act readiness package
PDF
AgentPMO_EU_AI_Act_Readiness_AcmeCorp.pdf
47 pages · 12 agents · 6 frameworks · SHA-256 sealed
Audit Log
Every governance action, hash-chained and immutable. This is what "evidence, not assurances" means.
Recent entries CHAIN VERIFIED ✓
✓ CHAIN INTEGRITY VERIFIED — 1,847 ENTRIES · GENESIS 2026-05-11 · NO GAPS, NO REWRITES