QP
Fleet Overviewagent-as-actor · grounded in qcontrol telemetry
Assurance · Fleet Overview
Every AI agent observed across the fleet — what they are, do, cost, and how safe they are.
8
agents
7 instrumented
4
active runs
29.1M
tokens
3.9M in 24h
$684.50
spend
$142.60 in 24h
61
sessions
6
MCP servers
8
agent types
52.4k
events captured
Fleet posture
score 76 / 100 · target ≥ 80Hardened
4
At risk
3
Critical
1
Spend by agent
$684.50 totalClaude Code
$214.50
support-copilot
$168.40
Cursor
$96.20
Codex CLI
$78.00
data-pipeline-agent
$52.30
marketing-gpt
$44.10
What agents do
event familiestool
11.2k
llm
8.9k
mcp
4.2k
file
3.8k
process
2.1k
run
61
Top risks
lowest posture firstmarketing-gptApprovals & sandbox bypassed (yolo)25
Zed AgentNo tool-approval policy60
support-copilotRunning as root65
Codex CLISandbox disabled70
CursorNo tool-approval policy85
Agents in use — by tool & owner
8 agents · 61 sessionsAgentOwnerModelTokensSpendPosture
Claude CodeAnthropic · climark@qpoint.ioclaude-opus-4-86.9M$214.50StandardCursorAnysphere · idejane@qpoint.ioclaude-sonnet-4-63.1M$96.20StandardCodex CLIOpenAI · clidev-7f3@duck.comgpt-5-codex2.4M$78.00Elevatedsupport-copilotInternal · servicesvc-support@qpoint.iogpt-4o5.6M$168.40Elevatedmarketing-gptInternal · servicerootgpt-4o1.1M$44.10Criticaldata-pipeline-agentInternal · servicesvc-data@qpoint.ioclaude-haiku-4-59.2M$52.30StandardClaude DesktopAnthropic · desktopmark@qpoint.ioclaude-opus-4-8800.0k$31.00StandardZed AgentZed Industries · idejane—0$0.00ElevatedFields: security_score · danger_level · tokens · cost · sessions · activity.families — from app/data/c16-fleet.ts (qcontrol /api shape).