REPAIR CONTROL

Failures explained. Repairs controlled.

Loading repair cases…
SECURITY POSTURE

What needs fixing first.

One place for dependency advisories, local scan findings and the exact next safe action.

Evidence-led · never auto-fix
POSTURE SCORE—/100
Finding trendlatest checks
CRITICAL—
HIGH—
MEDIUM—
Fix nextEvery action stays reviewable
Loading posture…
Scan an affected code, config or prompt excerpt
Scan resultLocal only
Run a focused check when you need more evidence.
LIVE MONITORING

Watch the project and its servers.

Project health covers dependency risk. Server health covers everything detected after connection — including containers and orchestration — without adding extra screens.

Monitoring status loading…
PROJECT HEALTH

Watch dependency risk.

Import the package manifest from the project you care about. Sentinel checks known advisories on a schedule.

No monitored projects yet.
SERVERS—
ONLINE—
NEEDS ATTENTION—
STALE—
Your serversSigned heartbeat every 60 sec
No connected server yet.
CONNECT A SERVER

Add your servers.

Connect as many servers as you run. Sentinel inspects each one, configures monitoring for the detected stack automatically, and lists every connection above.

EVIDENCE-LED IDEAS

Useful ideas, never blind automation.

Sentinel proposes small repairs and larger product improvements only when it can name the observed evidence behind them.

AI receives only an operational summary after you explicitly request it — never source code, secrets or customer data.

Loading recommendations…
SKILL STUDIO

Turn incident-response rules into agent instructions.

Generate a versioned Markdown skill for Codex, Claude Code or any coding agent, so a team responds to recurring failures consistently.

safe-crm-repair.md
Generate a skill to see the reviewed draft here.
Suggested for your projects
Loading suggestions…
SAFETY RULES

Separate rules for code and infrastructure.

Project rules decide how Sentinel handles vulnerabilities and repair plans. Infrastructure rules decide what a connected server, Docker host or Kubernetes cluster may do.

PROJECT GUARDRAILS
VULNERABILITY RESPONSE
INFRASTRUCTURE MODE
ALERT THRESHOLDS
ACTIONS SENT TO THE NODE AGENT
PLAN & BILLING

Understand incidents before they cost more.

Start with one FlowWatch workspace, then add more incident volume and shared repair decisions.

Checkout is wired as a provider handoff. Connect Stripe in deployment before collecting payment details.