Cloak HarborExposure Defense
Production pilot · authorized data only

Personal Exposure Defense, operationalized.

Cloak Harbor turns privacy exposure into an intelligence-grade workflow: discover exposure, verify evidence, score risk, create work, schedule remediation, route human approvals, and prove progress.

Open Live Cockpit Start Client OnboardingOpen Client ReportOpen Management PortalReview Human QueueView Executive Report
Backend
Postgres
persistent case state
Storage
S3 / Blob
long-term evidence vault
Calendar
Cron
scheduled OSINT checks
Kanban
Auto
due tasks advance
Workflow
Vercel
durable protected runs
Evidence
Linked
clickable source records
Auth
Server
PLATPHORM_API_KEY only
Approval
Human
before external action
Defense loop ledger

One inspectable line from principal to report.

This ledger is the plain-English audit trail for the demo. Every stage names the real fixture-backed proof, the scaffolded boundary, and whether a human approval gate is required before the outside world is touched.

Principal

Jordan Hale

synthetic-principal-001

Loop stages

10

same principal

Evidence

12

redacted records

Score

54 -> 78

demo calibration

Work

12

case tasks

Approvals

8

human gates

Evidence

Fixture-backed and labeled.

Risk

Fixture-backed and labeled.

Approval

Fixture-backed and labeled.

Report

Fixture-backed and labeled.

Checkpoint persistence

Protected operators can POST /api/v1/loop with PLATPHORM_API_KEY to record a Postgres-backed snapshot. Vercel production is durable when DATABASE_URL is configured and degraded when it is not.

Configured
Postgres: vercel_postgresVercel: persistent

externalActionTaken:false

01

Principal

Synthetic

The whole demo is anchored to one synthetic executive principal, Jordan Hale, with consent and scope visible before any work begins.

0

Findings

0

Evidence

0

Packets

externalActionTaken: false

Open stage

02

Discovery

UI scaffold

The discovery stage shows what would be searched and why, while refusing to run a live people-search crawl.

12

Findings

0

Evidence

0

Packets

externalActionTaken: false

Open stage

03

Evidence

Connected

Each exposure has evidence with attribution, confidence, redaction state, and trace linkage before it can drive action.

12

Findings

12

Evidence

0

Packets

externalActionTaken: false

Open stage

04

Risk

Connected

StealthScore turns the evidence set into a plain before/after risk posture with transparent drivers.

12

Findings

0

Evidence

0

Packets

externalActionTaken: false

Open stage

05

Work

Connected

The risk posture becomes owned work, with every task linked back to an exposure and lane.

12

Findings

0

Evidence

0

Packets

externalActionTaken: false

Open stage

06

Cadence

Connected

The calendar makes the 30-day defense rhythm visible: immediate hygiene, partner review, re-checks, and reporting.

7

Findings

0

Evidence

0

Packets

externalActionTaken: false

Open stage

07

Action

Dry-run

Action packets are drafts only: they help a human decide what to send, but they do not touch outside systems.

2

Findings

0

Evidence

2

Packets

externalActionTaken: false

Open stage

08

Approval

Human review

The review center separates useful AI/operator drafting from the authority to act outside Cloak Harbor.

7

Findings

0

Evidence

2

Packets

externalActionTaken: false

Open stage

09

Report

Ready for review

The report compresses the case into an executive-ready proof package: status, risk, actions, open issues, and limits.

11

Findings

0

Evidence

2

Packets

externalActionTaken: false

Open stage

10

Recurrence

Degraded

The loop does not end at a report; recurring exposures are watched and routed back into evidence, work, and review.

5

Findings

0

Evidence

0

Packets

externalActionTaken: false

Open stage

Synthetic case. Real operating model.

Jordan Hale is fictional, but the workflow mirrors a real principal-defense motion.

Evidence captured, not assumed.

Every major action is tied back to findings, confidence, source class, and review state.

AI drafted. Human approval required.

No message, scan, upload, provider call, or packet submission happens from this UI phase.

Personal Exposure Defense OS

A real operating model for nontechnical people: understand the risk, approve the work, prove the progress.

Cloak Harbor is intentionally not a fake scanner or black-box AI console. The system shows what is known, what is scaffolded, what needs approval, and what persists when Postgres is configured locally or on Vercel.

Capabilities

32

Loop stages

11

Human-gated

18

External actions

0

Exposure-defense loop coverage

Every capability maps back to the same Jordan Hale loop instead of becoming a disconnected tool tour.

Synthetic

Intake

4

Confirm who is protected, what is in scope, and what the system is allowed to touch.

Discover

2

Plan bounded exposure discovery against synthetic or authorized identifiers.

Verify

2

Review evidence before anything becomes a risk or action item.

Score

1

Translate exposure into a clear StealthScore and explain what would improve it.

Queue

1

Turn validated findings into accountable work with owners and due dates.

Operate

10

Run the work through Kanban, service setup, partner handoffs, and communications drafts.

Cadence

1

Keep the 30-day remediation rhythm visible for clients and operators.

Draft

1

Draft opt-out, correction, partner, and client packets without sending them.

Approve

4

Require a person to approve sensitive work before any outside-world touch.

Prove

6

Package the before/after story into reports, screenshots, and audit evidence.

Monitor

0

Watch for re-exposure after cleanup without pretending live monitoring is active.

Prove

Executive landing

Local preview

A boardroom-ready first screen that explains Personal Exposure Defense as one practical operating loop, not a generic dashboard.

Jordan Hale synthetic principal12 deterministic findingsStealthScore 54 to 78

externalActionTaken: false

Open capability

Intake

Client onboarding

Human review

A calm intake wizard that explains what happens next and makes consent boundaries clear before any discovery work.

Responsible-use gateRisk posture stepsSynthetic readiness score

externalActionTaken: false

Open capability

Discover

Discovery planning

UI scaffold

A planning console for bounded discovery that shows what would be searched before any live scan exists.

Source classifierFalse-positive guardRecurrence watch

externalActionTaken: false

Open capability

Verify

Evidence review

Connected

An evidence vault that makes clear what was captured, why it matters, and what still needs human review.

Evidence idsRedaction stateTrace id

externalActionTaken: false

Open capability

Score

StealthScore center

Connected

An explainable risk center that shows why the synthetic principal starts at 54 and can move to 78.

Baseline 54Projected 78Evidence-backed drivers

externalActionTaken: false

Open capability

Approve

Human review center

Human review

The approval desk where sensitive decisions wait for a person before they can leave the system.

Review queuesDecision reasonsRequired approver

externalActionTaken: false

Open capability

Draft

Action packet studio

Dry-run

A drafting studio that turns findings into clear request packets while keeping every send blocked.

AI draft panelCompliance checklistApproval state

externalActionTaken: false

Open capability

Prove

Report studio

Ready for review

A one-page executive report builder that explains what changed, what remains open, and what happens over the next 30 days.

Executive summaryScore movementTop risks

externalActionTaken: false

Open capability

Sensitive action matrix

No external action is taken. Sensitive capabilities are explainable, human-gated, and explicitly marked scaffolded or protected until backend wiring exists.

Human review

Client onboarding

Human review

Synthetic or authorized data only

externalActionTaken: false

Customer onboarding

Waiting client

No document upload success is claimed

externalActionTaken: false

Management onboarding

UI scaffold

No team changes are saved

externalActionTaken: false

Discovery planning

UI scaffold

No live scan has been performed

externalActionTaken: false

Calendar cadence

Protected action

No real calendar event is created

externalActionTaken: false

Human review center

Human review

No auto-submit

externalActionTaken: false

Action packet studio

Dry-run

AI drafted, human approval required

externalActionTaken: false

Communications hub

UI scaffold

No sending without protected server action

externalActionTaken: false

Jitsi consultation room

Protected action

No call history is faked

externalActionTaken: false

Twilio communications

UI scaffold

No fake delivered state

externalActionTaken: false

SendGrid mail provisioning

UI scaffold

No fake delivered state

externalActionTaken: false

Cloudflare posture

Needs env

No Cloudflare API call is made

externalActionTaken: false

Backend support mode

Vercel Postgres/Neon persistence is configured for production state, audit ledgers, and protected operator workflows.

Configuredpostgres

Postgres

vercel_postgres

Vercel mode

persistent

No external action

false

Support flag
Docker
Vercel
State

Deterministic synthetic fixture mode

Never degraded; this is the safe public baseline.

Yes
Yes
Connected

PED OS capability catalog

Catalog remains readable even when persistence is unavailable.

Yes
Yes
Connected

Postgres capability support flags

Vercel reports fixture/degraded mode only until DATABASE_URL is configured.

Yes
Yes
Configured

Local audit event persistence

Deployments without DATABASE_URL show public-safe trace fixtures and do not claim durable audit persistence.

Yes
Yes
Configured

Protected capability bootstrap API

Without Postgres, the route returns postgres_not_configured instead of fake success.

Yes
Yes
Protected action

MCP capability introspection

MCP still reports fixture/degraded runtime when Postgres is absent.

Yes
Yes
Connected

External service actions

Every external action remains scaffolded/protected until intentionally wired.

No
No
UI scaffold

Live web scan execution

No live scan has been performed in demo mode.

No
No
UI scaffold

SMS/email delivery

No message has been sent and no delivered state is shown.

No
No
Dry-run

Provider provisioning

No resource has been provisioned.

No
No
Protected action

Executive report publishing

No report is published to Docs, Sheets, Files, or AWS until those integrations are configured.

Yes
Yes
Ready for review

Responsible-use gates

Safety gates remain active in fixture mode.

Yes
Yes
Connected

Discover exposure

Curated findings across people-search, property, public-record, corporate filings, credentials and identity linkage — never an unbounded crawler.

Verify evidence

Every finding carries source class, capture timestamp, confidence, attribution and a false-positive state before it can drive action.

Score with StealthScore

A transparent, rubric-driven risk model converts raw exposure into a defensible posture score and a projected remediation delta.

Operate & prove

Kanban, calendar cadence, AI-drafted action packets, human approval gates, trace timeline and a one-page executive report.

The operating loop

One repeatable Personal Exposure Defense loop — from intake to recurrence monitoring.

The pilot runs authorized client-defense cases with persisted profiles, evidence, scheduled OSINT checks, human approvals and reversible downstream handoffs.

  1. 01IntakeConsent-scoped principal profile
  2. 02Exposure discoveryCurated findings, no live scrape
  3. 03Evidence validationSource class, confidence, attribution
  4. 04Risk scoringRubric-driven StealthScore delta
  5. 05Case operationsKanban remediation lanes
  6. 06Calendar cadenceOpt-out & partner deadlines
  7. 07Agent-assisted actionAI drafts in dry-run mode
  8. 08Human approvalNo external send without a gate
  9. 09Trace & reportOne-page executive proof
  10. 10Recurrence monitorRe-scan & watch

Proof pattern

The same machinery that runs the platform runs the defense loop.

Discovery, evidence validation, agent-safe operations, case management, auditability and reporting are mapped to real operating capabilities — surfaced behind the scenes, never as a tool tour.

Exposure Discovery

SearchOpsBrowserOpsOpenContentRoot graph

Identity / Evidence Validation

TrustOpsCatalogSpecEvals

Agent-Safe Operations

AgentOpsMCP GatewayClawsSandboxKeysWebhooks

Case Management

KanbanCalendarSheetsDocsClaws

Auditability

TraceMonitorEvalsFingerprintHealthFiles

Reporting / Handoff

DocsSheetsOpenContentWebhooksFiles

Open-core wedge

SigReduce: the open foundation for exposure defense.

Open source

  • Exposure object schema
  • Source evidence model
  • Identity graph structure
  • Risk label taxonomy
  • Synthetic / redacted dataset
  • Scoring rubric
  • API contract
  • Reference CLI
  • Demo fixtures
  • Validation tests

Commercial · Cloak Harbor

  • StealthScore calibration
  • Partner connectors
  • Legal templates
  • Tradecraft
  • Client portal
  • Analyst services
  • Premium reporting
  • Partner workflows
  • Production action routing

We open-source the schema, evidence model and risk taxonomy, never a personal-data scraper. Read the SigReduce overview.

Responsible use is built in

  • Synthetic data only unless client-authorized.
  • Authorized privacy defense only.
  • No stalking, doxxing, harassment, evasion, impersonation, credential theft, or illegal scraping.
  • No automated takedown without human approval.
  • No live scraping of real people in demo mode.
  • No private data leaves the app without explicit operator action.
  • AI drafts are suggestions and require human review.
  • All sensitive workflows are gated.
  • The system is for defensive exposure reduction and audit-ready privacy operations.
Full responsible-use policy

The pilot ask

Let's pick one narrow principal-defense workflow and turn it into a real demo in days, not months.

Open the cockpitHow it works