SCRIMED Cyber Defense Command Center

Turn security into an operating control plane across every SCRIMED route.

SCRIMED Cyber Defense Command Center is a synthetic/no-PHI security-readiness and control-monitoring layer. It does not claim security certification, HIPAA compliance, SOC 2, HITRUST, penetration-test completion, breach immunity, PHI authority, production connector approval, or customer go-live.

Inspect API Evidence Packet API Trust Center Release Continuity Workarounds QA Evidence

Statusscrimed-cyber-defense-active-no-secrets-no-phi

Enforced controls6

Monitored controls2

Critical threats2

High threats3

Diligence cards8

Share-ready cards5

Evidence gaps3

Audit hashscrimed-intel-42903637

Cybersecurity Posture

SCRIMED now treats cybersecurity as an operating control plane: browser hardening, request sanitization, protected-route fail-closed behavior, token redaction, no-PHI safety governance, incident lanes, and buyer diligence cards are visible and contract-tested.

SCRIMED's current cyber layer is designed for no-PHI diligence, synthetic pilots, protected operator evidence, and future regulated deployment preparation.

The current posture preserves fail-closed protected routes, browser hardening, proxy sanitization, token redaction, no-secret tests, and explicit no-authority headers.

External WAF, SIEM, penetration testing, SBOM signing, and customer-specific controls remain required before PHI authority.

Buyer Security Diligence Cards

Procurement-grade answers are packaged as auditable, synthetic-only control cards.

manual_operator_required

AAL2 and role-scoped operator access

How does SCRIMED constrain protected operator and evidence workflows?

Protected AAL2 flows are designed to fail closed without authorized tenant-admin, pilot-lead, or reviewer context, and happy-path writes require a short-lived authorized bearer session.

Category: identity_access
Owner: Security lead + tenant administrator
Review cadence: Before each protected no-PHI pilot evidence run.
Evidence routes: /qa-aal2-run-evidence, /api/scrimed-cyber-defense/evidence-packet
Required next evidence: Run strict AAL2 durable-store smoke with an authorized role and attach only the redacted token fingerprint result.
Human review required: true
Synthetic only: true
Audit hash: scrimed-intel-106ddb8d
Blocked claims: No security certification assertion. No breach immunity assertion. No live PHI authority. No customer go-live authority. No production connector approval. No autonomous clinical authority.
Residual risk: Customer SSO, SCIM, emergency access, and tenant-specific access review evidence remain deployment-specific work.

ready_to_share

Token, secret, and credential exposure controls

How does SCRIMED avoid exposing credentials during demos, smokes, and buyer evidence review?

No-secret tests, AAL2 helper redaction, token fingerprinting, clipboard cleanup support, and forbidden-claim checks keep credentials out of source, logs, docs, and chat output.

Category: token_secret_protection
Owner: Platform security
Review cadence: Every release candidate and before any protected smoke evidence is shared.
Evidence routes: scripts/aal2-bearer-token-helper.mjs, scripts/lib/aal2-token-policy.mjs, npm run test:nonsecret
Required next evidence: Attach managed secret-scan output and rotation runbook metadata before PHI-bearing review.
Human review required: true
Synthetic only: true
Audit hash: scrimed-intel-a660ce49
Blocked claims: No security certification assertion. No breach immunity assertion. No live PHI authority. No customer go-live authority. No production connector approval. No autonomous clinical authority.
Residual risk: Managed secrets scanning and production secret rotation still require external CI/platform configuration.

ready_to_share

Protected route fail-closed posture

Do protected APIs fail closed when authentication, AAL2, tenant context, or role checks are missing?

Protected durable-store and evidence flows return fail-closed responses without valid authorized context, while public buyer surfaces expose metadata-only readiness.

Category: protected_route_boundary
Owner: Platform reliability + security
Review cadence: Every protected-route change.
Evidence routes: /api/workflows/execution-attempts/durable-store, npm run smoke:execution-attempt-durable-store
Required next evidence: Attach customer-specific protected-route access review after SSO/RBAC configuration.
Human review required: true
Synthetic only: true
Audit hash: scrimed-intel-9f8b432f
Blocked claims: No security certification assertion. No breach immunity assertion. No live PHI authority. No customer go-live authority. No production connector approval. No autonomous clinical authority.
Residual risk: Production authorization evidence still needs customer identity-provider integration and deployment logs.

ready_to_share

No-PHI and raw-payload boundary

What prevents synthetic diligence routes from becoming live PHI processing paths?

Safety governance, explicit no-PHI headers, no raw connector payload logging, and retained no-go boundaries keep current routes metadata-only and synthetic-only.

Category: data_boundary
Owner: Privacy lead + clinical governance
Review cadence: Every clinical, connector, or data-ingestion scope expansion.
Evidence routes: /api/scrimed-cyber-defense, app/lib/scrimedSafetyGovernance.ts
Required next evidence: Complete privacy risk assessment, BAA/customer agreement, retention policy, and customer-specific data-flow review.
Human review required: true
Synthetic only: true
Audit hash: scrimed-intel-06ad3044
Blocked claims: No security certification assertion. No breach immunity assertion. No live PHI authority. No customer go-live authority. No production connector approval. No autonomous clinical authority.
Residual risk: Live PHI processing requires legal, privacy, retention, customer, infrastructure, and incident-response approvals outside this code path.

ready_to_share

Proxy and request-header sanitization

How does SCRIMED reduce middleware-bypass and suspicious forwarded-header risk?

Next proxy sanitizes middleware-bypass, debug-token, API-key, and token-like forwarded headers before route handlers receive requests.

Category: request_sanitization
Owner: Platform security + infrastructure
Review cadence: Every proxy, routing, or security-header change.
Evidence routes: proxy.ts, next.config.js, npm run smoke:scrimed-cyber-defense
Required next evidence: Attach deployment WAF/bot-management policy and CSP report workflow before regulated traffic.
Human review required: true
Synthetic only: true
Audit hash: scrimed-intel-734f79e5
Blocked claims: No security certification assertion. No breach immunity assertion. No live PHI authority. No customer go-live authority. No production connector approval. No autonomous clinical authority.
Residual risk: External WAF and bot-management controls remain deployment-specific defense-in-depth requirements.

ready_to_share

API abuse and model-cost guardrails

How does SCRIMED reduce billing-spike and automated abuse risk while provider calls are disabled by default?

Cost guardrail metadata, provider-call kill-switch defaults, request-counting scaffolds, and safe public errors keep current AI/provider usage controlled.

Category: api_abuse_cost_guardrail
Owner: Platform reliability + finance operations
Review cadence: Every model-router, provider, or public API expansion.
Evidence routes: app/lib/costApiGuardrails.ts, app/lib/requestRateLimit.ts, docs/RISK_REGISTER.md
Required next evidence: Attach tenant quota policy, billing alert thresholds, and abuse-response runbook before paid provider exposure.
Human review required: true
Synthetic only: true
Audit hash: scrimed-intel-5034046f
Blocked claims: No security certification assertion. No breach immunity assertion. No live PHI authority. No customer go-live authority. No production connector approval. No autonomous clinical authority.
Residual risk: Production billing anomaly alerting, WAF throttling, and tenant-specific quotas must be configured in hosting/provider systems.

external_evidence_required

Incident response and tabletop readiness

What happens if a token, PHI-risk submission, protected route, or public route is abused?

Incident lanes define triggers, first response, owners, and evidence routes for token leakage, protected-route abuse, PHI exposure risk, and public-route abuse.

Category: incident_response
Owner: Security lead + privacy lead + release steward
Review cadence: Quarterly before regulated pilots and after every security incident.
Evidence routes: /scrimed-cyber-defense, docs/scrimed-cyber-defense.md
Required next evidence: Attach redacted tabletop notes, escalation roster, SIEM/log-drain proof, and incident communications template.
Human review required: true
Synthetic only: true
Audit hash: scrimed-intel-b24fb29a
Blocked claims: No security certification assertion. No breach immunity assertion. No live PHI authority. No customer go-live authority. No production connector approval. No autonomous clinical authority.
Residual risk: Tabletop exercise records, SIEM proof, pager escalation, and customer notification paths are not established by code alone.

blocked_until_approved

Vendor connector and production integration readiness

What remains blocked before EHR, payer, imaging, or device connector activation?

Production connectors remain blocked until customer-specific threat models, BAAs/contracts, data-flow reviews, testing evidence, and formal approvals exist.

Category: vendor_connector_readiness
Owner: Security lead + integration owner + customer security
Review cadence: Per connector and per customer environment.
Evidence routes: /release-continuity, /boundary-release-approvals, /approvals-readiness
Required next evidence: Complete connector threat model, least-privilege scope map, rollback drill, legal approval, and customer security approval.
Human review required: true
Synthetic only: true
Audit hash: scrimed-intel-f2dcc457
Blocked claims: No security certification assertion. No breach immunity assertion. No live PHI authority. No customer go-live authority. No production connector approval. No autonomous clinical authority.
Residual risk: Each connector needs vendor security review, least-privilege scopes, rollback plan, monitoring, and customer signoff.

Enforced Controls

Controls are categorized by surface, status, evidence, protected assets, and retained boundary.

enforced

Global browser security headers

Next.js response headers enforce CSP, frame denial, MIME sniffing prevention, referrer limits, HSTS, CORP, COOP, Origin-Agent-Cluster, and restricted browser permissions.

browser_surface

Evidence: next.config.js, npm run build, npm run smoke:public
Protected assets: public app routes, investor/buyer surfaces, synthetic readiness APIs
Verification: Contract check verifies hardened headers stay present.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

enforced

Proxy request-header sanitizer

Next proxy strips middleware-bypass, debug-token, API-key, and forwarded token headers before App Router handlers receive requests.

request_sanitization

Evidence: proxy.ts, npm run smoke:scrimed-cyber-defense
Protected assets: all non-static routes, protected AAL2 APIs, public diligence routes
Verification: Contract check requires x-middleware-subrequest stripping and proxy guard headers.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

monitored

Protected APIs fail closed

Protected durable-store and buyer evidence paths require authorized AAL2 bearer context and return fail-closed responses without valid sessions.

auth_rbac

Evidence: scripts/execution-attempt-durable-store-authenticated-smoke.mjs, scripts/aal2-token-policy-selftest.mjs
Protected assets: execution attempt evidence, boundary release evidence, tenant-scoped pilot workspaces
Verification: Nonsecret suite runs missing-token and policy checks without logging tokens.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

monitored

Cost and API abuse guardrails

Request counting, cost guardrail metadata, provider-call kill switch defaults, and safe synthetic limits constrain high-volume abuse paths.

api_abuse

Evidence: app/lib/requestRateLimit.ts, app/lib/costApiGuardrails.ts, docs/RISK_REGISTER.md
Protected assets: model-router budget, public APIs, synthetic evaluation endpoints
Verification: Readiness contracts keep provider calls disabled by default unless explicitly configured.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

enforced

Token and secret redaction

Bearer-token helper, smoke paths, and policy tests use fingerprints and redaction rather than writing or logging full credentials.

secret_protection

Evidence: scripts/aal2-bearer-token-helper.mjs, scripts/lib/aal2-token-policy.mjs
Protected assets: AAL2 bearer tokens, Supabase sessions, local .env.local
Verification: Nonsecret suite clears token env vars and validates local helper behavior without exposing secrets.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

enforced

PHI and clinical authority safety gate

Central safety gate blocks live PHI, final clinical decision authority, treatment, prescribing, outreach, payer submission, EHR writeback, connector approval, and certification claims.

phi_safety

Evidence: app/lib/scrimedSafetyGovernance.ts, app/api/scrimed-cyber-defense/route.ts
Protected assets: clinical-facing copy, AI workflow APIs, public diligence materials
Verification: API route evaluates safety governance before returning cyber-defense metadata.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

enforced

No-secret regression suite

Nonsecret test runner clears bearer-token environment variables and validates generated-integrity, safety, AAL2, and SCRIMED control contracts.

supply_chain

Evidence: scripts/scrimed-nonsecret-test-suite.mjs, scripts/check-generated-integrity.mjs
Protected assets: CI quality gates, source tree integrity, nonsecret developer workflows
Verification: npm run test:nonsecret runs the cyber-defense contract with credentials blanked.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

enforced

Security assurance pipeline

A no-secret assurance contract checks hardening headers, proxy sanitization, forbidden claims, token-like leaks, route wiring, and retained authority boundaries.

supply_chain

Evidence: app/lib/scrimedSecurityAssurancePipeline.ts, scripts/scrimed-security-assurance-contract-check.mjs
Protected assets: source code, docs, scripts, public cyber route, nonsecret CI gates
Verification: npm run security:assurance and npm run test:nonsecret keep the assurance gate active.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

planned_external_dependency

Incident response readiness lanes

Incident lanes define first response for token leakage, suspicious agent action, PHI exposure risk, dependency alert, and public-route abuse.

incident_response

Evidence: docs/scrimed-cyber-defense.md, /scrimed-cyber-defense
Protected assets: operators, buyers, evidence routes, future PHI programs
Verification: Requires external owner assignment, SIEM/log drain setup, and tabletop exercise before production PHI authority.
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

Threat Matrix

Every major threat receives a current mitigation, detection signal, escalation path, and residual risk.

high

Prompt injection and tool misuse

Untrusted content, future RAG sources, agent tool calls, MCP connectors.

Safety governance, scoped tool registries, human review, no direct LLM-to-record-system authority.

Detection: Unexpected tool request, missing citation, untrusted source escalation, denied policy decision.
Escalation: Pause automation, route to security lead and clinical governance, convert trace into regression test.
Residual risk: Future live connectors require external penetration testing, red-team evaluation, and customer-specific policy review.

critical

Credential or bearer-token leakage

Local terminals, browser console, protected smoke tests, environment files, logs.

Token helper stores only local env state with restricted mode, redacts fingerprints, proxy strips token-like debug headers.

Detection: Token-like value in logs, invalid/expired token preflight, suspicious forwarded auth header.
Escalation: Clear clipboard, rotate token/session, revoke compromised session, rerun nonsecret suite.
Residual risk: Production secret rotation playbooks and managed secrets scanning must be externally operated.

critical

PHI exposure before approval

Synthetic demos, future import pipelines, user pasted text, connector payloads.

No-PHI boundary, PHI pattern classification, no raw connector payload logging, safety gate blocks live PHI.

Detection: PHI pattern match, live-patient wording, raw payload attempt, connector write request.
Escalation: Block request, isolate artifact, require privacy/security review, document root cause.
Residual risk: Live PHI authority requires BAA, privacy program, logging controls, retention policy, and customer-specific approval.

high

Middleware/proxy bypass header abuse

Inbound HTTP headers and reverse-proxy forwarding behavior.

Proxy strips x-middleware-subrequest and debug-token headers and labels sanitized responses.

Detection: X-SCRIMED-Middleware-Bypass-Header response value shows stripped.
Escalation: Inspect upstream proxy, add WAF rule, verify protected routes still fail closed.
Residual risk: External WAF enforcement and platform log correlation remain deployment tasks.

high

API abuse and model-cost spike

Public APIs, synthetic evaluation routes, model-router scaffolds, future provider calls.

Provider calls disabled by default, cost guardrails, rate-limit metadata, no-secret tests.

Detection: Request-count threshold, cost threshold, repeated synthetic-eval calls, provider-call kill switch.
Escalation: Return safe error, enable stricter route limits, review source IP and tenant context.
Residual risk: Production-grade WAF, bot rules, and billing anomaly alerts require hosting-provider configuration.

moderate

Supply-chain and generated artifact drift

Node dependencies, generated .next artifacts, build cache, package overrides.

Generated integrity check, lockfile discipline, audit-ready dependency structure, no generated duplicate suffixes.

Detection: Integrity check failure, npm audit finding, unexpected generated duplicate.
Escalation: Clean generated cache, review dependency advisory, patch or pin safely.
Residual risk: SBOM signing and external dependency scanning should be added before regulated deployment.

Security Assurance Pipeline

No-secret security checks convert cyber readiness into repeatable release evidence.

automated

Static token and secret leak scan

Source files must not contain live bearer tokens, JWTs, API keys, private keys, or cloud credentials.

scripts/scrimed-security-assurance-contract-check.mjs scans curated source surfaces for high-risk token patterns.

Category: secret_detection
Evidence: npm run security:assurance, npm run test:nonsecret
Failure mode: A live credential-like value enters source, docs, scripts, or route code.
Manual follow-up: Rotate the affected credential, remove the value, preserve only a fingerprint, and rerun nonsecret checks.

automated

Global security header regression gate

CSP, HSTS, frame denial, no-sniff, CORP, COOP, Origin-Agent-Cluster, and restricted Permissions-Policy must remain present.

Security assurance checks next.config.js for required hardening headers and no-authority SCRIMED headers.

Category: browser_hardening
Evidence: next.config.js, npm run smoke:scrimed-cyber-defense
Failure mode: A route or refactor removes browser hardening or no-authority response headers.
Manual follow-up: Review CSP compatibility and add report-only monitoring before stricter enforcement.

automated

Middleware/proxy bypass header guard

Inbound x-middleware-subrequest and token-like debug headers must be stripped before route handlers receive requests.

Security assurance requires proxy.ts to strip bypass and debug headers and emit scrub status headers.

Category: request_sanitization
Evidence: proxy.ts, live header smoke
Failure mode: A crafted header influences route behavior or bypass assumptions.
Manual follow-up: Mirror the stripping rule at WAF or reverse-proxy level before regulated deployment.

automated

Protected API fail-closed evidence

Protected AAL2 and buyer-evidence routes must fail closed without valid authorized context.

Public smoke and nonsecret suite verify protected routes return safe fail-closed states without bearer tokens.

Category: protected_access
Evidence: npm run smoke:public, scripts/aal2-token-policy-selftest.mjs
Failure mode: Unauthenticated access succeeds or exposes protected evidence.
Manual follow-up: Inspect auth/RBAC/RLS policy, revoke suspect sessions, and rerun strict AAL2 smoke with a valid short-lived token.

automated

No-PHI and clinical authority boundary

No route, page, doc, or script may claim live PHI authority, autonomous clinical care, payer submission, or EHR writeback.

Security assurance scans critical surfaces for forbidden authority phrases and unsafe clinical/compliance claims.

Category: phi_boundary
Evidence: app/lib/scrimedSafetyGovernance.ts, docs/NO_PHI_POLICY.md
Failure mode: Marketing or operator copy implies authority that SCRIMED does not have.
Manual follow-up: Route copy through legal, privacy, clinical governance, and qualified external reviewers before release.

automated

Generated integrity and dependency posture

Generated duplicate files must not corrupt resolution, and dependency/security evidence must be ready for external review.

Security assurance requires generated-integrity checks and documents the SBOM/dependency-scan follow-up.

Category: supply_chain
Evidence: scripts/check-generated-integrity.mjs, package.json
Failure mode: Generated duplicates, unsafe dependency drift, or missing dependency evidence blocks clean deploy review.
Manual follow-up: Add SBOM generation, dependency scan, provenance signing, and external vulnerability review before PHI authority.

automated

Framework dependency security floor

Next.js, React, React DOM, eslint-config-next, and postcss must remain pinned to reviewed versions at or above the current SCRIMED security floor.

scripts/dependency-security-floor-contract-check.mjs verifies package.json and package-lock.json agree on exact patched framework versions.

Category: dependency_floor
Evidence: npm run security:dependency-floor, package.json, package-lock.json
Failure mode: A dependency update or merge conflict silently downgrades SCRIMED below the approved framework security floor.
Manual follow-up: Review Next.js and React security advisories, update lockfile through npm ci/install in a trusted environment, then rerun nonsecret checks.

automated

CI workflow governance contract

Main CI and protected smoke workflows must retain no-secret validation, read-only permissions, deterministic Node setup, explicit timeouts, and no fail-open patterns.

scripts/ci-workflow-contract-check.mjs verifies GitHub Actions workflow controls for normal CI and protected AAL2 smoke paths.

Category: ci_workflow
Evidence: npm run contract:ci-workflows, .github/workflows/ci.yml, .github/workflows/*smoke.yml
Failure mode: A workflow edit removes safety checks, disables failure, prints secrets, or leaves protected smokes without timeouts.
Manual follow-up: Review GitHub Actions logs, rotate any exposed credential, restore fail-closed checks, and rerun the nonsecret suite before merge.

manual_required

Incident response lane coverage

Token leakage, PHI exposure risk, protected-route abuse, and public-route abuse must have owners and first-response steps.

Security assurance requires incident readiness metadata to remain visible in the cyber defense command center.

Category: incident_response
Evidence: app/lib/scrimedCyberDefenseCommandCenter.ts, docs/scrimed-cyber-defense.md
Failure mode: Security event response depends on ad hoc operator memory instead of a defined lane.
Manual follow-up: Run a tabletop exercise and attach SIEM/log-drain evidence before production PHI workflows.

planned_external_dependency

External deployment assurance boundary

WAF, bot controls, SIEM/log drains, customer-specific threat models, penetration testing, and legal/security review are required before authority expands.

Security assurance verifies this dependency remains documented and not misrepresented as complete.

Category: deployment_boundary
Evidence: docs/scrimed-cyber-defense.md, /scrimed-cyber-defense
Failure mode: The platform claims regulated or customer go-live readiness before external controls exist.
Manual follow-up: Complete external WAF, SIEM, SBOM, penetration test, incident tabletop, and customer-specific approvals.

Security Release Readiness Gate

SCRIMED can show synthetic and buyer-diligence readiness while keeping PHI and go-live authority blocked.

buyer_diligence_ready

Current authority: synthetic-and-no-phi-only

SCRIMED Security Release Readiness is a no-PHI go/no-go control ladder. It permits synthetic demos, buyer diligence, and protected no-PHI pilot preparation only. It does not approve live PHI, customer go-live, production connectors, payer submission, EHR writeback, regulated clinical use, security certification, SOC 2, HITRUST, HIPAA, FDA, penetration-test completion, or breach immunity.

scrimed-intel-e3e60c03

PHI authority: not-authorized-production-phi
Customer go-live: not-customer-go-live-approved
Go gates: 3
Conditional gates: 1
No-go gates: 2

Synthetic demo ready

Allowed: Public no-PHI demos, Synthetic evaluation, Investor/buyer proof review, Internal security assurance

Blocked: Live PHI, Customer go-live, EHR writeback, Payer submission

Required evidence: No-secret assurance, Browser hardening, Proxy sanitizer, Safety governance
Exit criteria: Public smoke passes, Security assurance passes, No unsafe claims detected

Buyer diligence ready

Allowed: Buyer diligence packets, Protected evidence review, No-PHI pilot scoping

Blocked: Production credentials, Raw connector payloads, Customer deployment claims

Required evidence: Protected routes fail closed, No-authority headers, Risk and boundary docs
Exit criteria: Protected route smoke passes, Security release ladder remains no-PHI

conditional_go

Protected no-PHI pilot ready

Allowed: AAL2 protected no-PHI pilot workspace, Short-lived authorized operator smoke, Evidence binding

Blocked: Live records, Production connector writes, Autonomous clinical action

Required evidence: Strict AAL2 happy path, Tenant role verification, Durable-store feature flag enabled
Exit criteria: Authorized operator smoke passes, Reviewer roles verified, No PHI fixtures used

no_go

PHI preproduction blocked

Allowed: Planning only, External evidence collection, Customer security questionnaire preparation

Blocked: PHI ingestion, PHI model routing, Production connector approval

Required evidence: WAF, SIEM, SBOM, external security review, privacy risk assessment
Exit criteria: Qualified external reviewers approve evidence, Legal/privacy owners sign release packet

no_go

Live PHI production blocked

Allowed: No code-only activation

Blocked: Customer go-live, Live PHI processing, EHR writeback, Payer submission, Clinical production claims

Required evidence: BAA/customer agreement, incident tabletop, retention policy, customer-specific threat model
Exit criteria: Formal customer, legal, privacy, security, and clinical governance approvals exist outside code

Security Diligence Evidence Packet

Buyer and investor evidence is packaged as redacted metadata with explicit blocked-authority boundaries.

scrimed-security-diligence-evidence-packet-active-no-phi

Buyer diligence share-ready: true

SCRIMED Security Diligence Evidence Packet is a metadata-only buyer and investor review layer. It packages no-PHI security evidence, retained boundaries, manual operator actions, and external evidence requirements without exposing secrets, tokens, raw logs, PHI, production connector payloads, or customer go-live authority.

scrimed-intel-267453bf

Ready to share: 4
Manual operator required: 1
External evidence required: 2
Blocked until approved: 1
Questionnaire responses: 8
Questionnaire ready: 3
Questionnaire external evidence gaps: 4
PHI production share-ready: false
Customer go-live share-ready: false

ready_to_share

Browser and proxy hardening packet

How does SCRIMED reduce public-route browser and middleware-bypass risk?

next.config.js, proxy.ts, npm run smoke:scrimed-cyber-defense

Category: platform_controls
Owner: Platform security
Release stage: buyer_diligence_ready
Redaction rule: Share header names, contract status, and control rationale only; do not share raw request logs.
Boundary: No live PHI, no production credentials, no raw log exports, no raw connector payloads, no payer submission, no EHR writeback, no production connector approval, no security certification claim, and no customer go-live approval.

ready_to_share

No-secret regression packet

How does SCRIMED prevent credentials, tokens, or unsafe claims from entering review artifacts?

npm run security:assurance, npm run test:nonsecret

Category: platform_controls
Owner: Platform security
Release stage: buyer_diligence_ready
Redaction rule: Share command names, pass/fail status, and fingerprints only; never share full bearer tokens or keys.
Boundary: No live PHI, no production credentials, no raw log exports, no raw connector payloads, no payer submission, no EHR writeback, no production connector approval, no security certification claim, and no customer go-live approval.

ready_to_share

Protected route fail-closed packet

Do protected buyer, operator, evidence, and AAL2 routes fail closed without authorized context?

npm run smoke:public, scripts/aal2-token-policy-selftest.mjs

Category: access_control
Owner: Platform reliability + security
Release stage: buyer_diligence_ready
Redaction rule: Share status codes and route categories only; do not share tenant records, tokens, or request payloads.
Boundary: No live PHI, no production credentials, no raw log exports, no raw connector payloads, no payer submission, no EHR writeback, no production connector approval, no security certification claim, and no customer go-live approval.

manual_operator_required

AAL2 operator validation packet

Can an authorized short-lived AAL2 session perform protected no-PHI evidence writes?

npm run smoke:aal2:token, npm run smoke:aal2:durable-store:strict

Category: access_control
Owner: Authorized tenant admin, pilot lead, or reviewer
Release stage: protected_no_phi_pilot_ready
Redaction rule: Share token fingerprint and command result only; never share the token, session JSON, or local .env.local.
Boundary: No live PHI, no production credentials, no raw log exports, no raw connector payloads, no payer submission, no EHR writeback, no production connector approval, no security certification claim, and no customer go-live approval.

ready_to_share

Release readiness ladder packet

What is SCRIMED allowed to show today, and what remains blocked before PHI or customer activation?

app/lib/scrimedSecurityReleaseReadiness.ts, /scrimed-cyber-defense

Category: release_governance
Owner: Release steward
Release stage: buyer_diligence_ready
Redaction rule: Share release stages, retained boundaries, and required evidence; do not imply expanded authority.
Boundary: No live PHI, no production credentials, no raw log exports, no raw connector payloads, no payer submission, no EHR writeback, no production connector approval, no security certification claim, and no customer go-live approval.

external_evidence_required

Incident tabletop packet

Has SCRIMED rehearsed incident handling for token leakage, PHI exposure risk, and protected-route abuse?

incident response runbook, tabletop notes, SIEM/log-drain proof

Category: incident_readiness
Owner: Security lead + privacy lead + clinical governance
Release stage: phi_preproduction_blocked
Redaction rule: Share tabletop summary and role matrix only after redaction; do not share raw incident logs.
Boundary: No live PHI, no production credentials, no raw log exports, no raw connector payloads, no payer submission, no EHR writeback, no production connector approval, no security certification claim, and no customer go-live approval.

external_evidence_required

WAF, SIEM, SBOM, and external review packet

Which deployment-specific security controls must exist before regulated or PHI-bearing review?

WAF policy, bot controls, redacted SIEM proof, SBOM artifact, external review letter

Category: external_dependency
Owner: Security lead + qualified external reviewers
Release stage: phi_preproduction_blocked
Redaction rule: Share external attestation metadata and redacted summaries only; do not share sensitive infrastructure details.
Boundary: No live PHI, no production credentials, no raw log exports, no raw connector payloads, no payer submission, no EHR writeback, no production connector approval, no security certification claim, and no customer go-live approval.

blocked_until_approved

BAA, privacy, retention, and customer approval packet

What evidence is required before live PHI, production connectors, or customer activation can be considered?

BAA/customer agreement, privacy risk assessment, retention policy, customer-specific threat model

Category: retained_boundary
Owner: Legal + privacy + customer security
Release stage: live_phi_production_blocked
Redaction rule: Share existence and approval status only through authorized legal/security review channels.
Boundary: No live PHI, no production credentials, no raw log exports, no raw connector payloads, no payer submission, no EHR writeback, no production connector approval, no security certification claim, and no customer go-live approval.

Security Questionnaire Response Library: requires_manual_review

access_control

Describe SCRIMED access control, privileged access, and protected workflow safeguards.

SCRIMED separates public metadata surfaces from protected AAL2 workflows. Protected no-PHI evidence writes require authorized operator context, role checks, and fail-closed route behavior; public diligence surfaces expose metadata only.

Evidence artifacts: protected-route-fail-closed-packet, aal2-operator-validation-packet
Supporting routes: /qa-aal2-run-evidence, /api/scrimed-cyber-defense/evidence-packet
Owner: Security lead + tenant administrator
External evidence needed: Customer SSO/RBAC evidence, access-review cadence, and emergency-access procedure.
Human review required: true
Redaction boundary: Share status codes, route classes, role labels, and token fingerprints only.
Cannot say: Do not claim customer SSO acceptance before customer identity-provider review. Do not share bearer tokens, session JSON, tenant records, or local .env.local values. Do not imply production PHI authority from no-PHI AAL2 smoke evidence.
Audit hash: scrimed-intel-0bc66527

Security Questionnaire Response Library: buyer_safe_answer_ready

data_protection

How does SCRIMED protect sensitive data during current demos and diligence?

Current SCRIMED diligence paths are synthetic/no-PHI and metadata-only. They avoid raw connector payload logging, production credentials, customer records, raw logs, and clinical authority claims while mapping the evidence required before PHI-bearing work.

Evidence artifacts: release-readiness-ladder-packet, baa-privacy-retention-packet
Supporting routes: /scrimed-cyber-defense, /approvals-readiness, /global-certification-readiness
Owner: Privacy lead + legal owner
External evidence needed: BAA/DPA path, privacy risk assessment, retention policy, and customer-specific data-flow approval.
Human review required: true
Redaction boundary: Share boundary language, evidence classes, owners, and approval gaps only.
Cannot say: Do not claim live PHI processing authority. Do not claim BAA/DPA execution unless a specific agreement exists outside this packet. Do not paste or store customer data in this public evidence packet.
Audit hash: scrimed-intel-29c5d7b3

Security Questionnaire Response Library: buyer_safe_answer_ready

application_security

What application security controls are visible in the current SCRIMED app?

SCRIMED exposes browser hardening headers, proxy request-header sanitization, forbidden-claim checks, generated-integrity checks, and no-secret regression gates as reviewable metadata controls.

Evidence artifacts: browser-proxy-hardening-packet, no-secret-regression-packet
Supporting routes: /api/scrimed-cyber-defense, next.config.js, proxy.ts
Owner: Platform security
External evidence needed: WAF policy, bot-management rule summary, CSP report workflow, and external application security review.
Human review required: true
Redaction boundary: Share control names, contract checks, and no-authority headers; keep raw traffic details outside the packet.
Cannot say: Do not represent header hardening as independent penetration-test completion. Do not share raw request logs or infrastructure-sensitive details. Do not imply WAF or bot-management configuration unless deployment evidence exists.
Audit hash: scrimed-intel-512a0e38

Security Questionnaire Response Library: requires_external_artifact

incident_response

How would SCRIMED respond to credential leakage, protected route abuse, or PHI exposure risk?

SCRIMED defines incident lanes for token leakage, protected-route abuse, PHI exposure risk, and public-route abuse. Each lane names a trigger, owner, first response, evidence route, and retained boundary.

Evidence artifacts: incident-tabletop-packet
Supporting routes: /scrimed-cyber-defense, docs/scrimed-cyber-defense.md
Owner: Security lead + privacy lead
External evidence needed: Tabletop notes, escalation roster, SIEM/log-drain proof, and customer notification template.
Human review required: true
Redaction boundary: Share role matrix, triggers, and redacted tabletop summaries only.
Cannot say: Do not claim completed tabletop exercise until external notes exist. Do not share raw incident logs. Do not promise incident outcomes or breach immunity.
Audit hash: scrimed-intel-63565aba

Security Questionnaire Response Library: requires_external_artifact

infrastructure

Which infrastructure controls are required before regulated deployment?

SCRIMED identifies WAF/bot controls, SIEM/log drains, SBOM generation, dependency scanning, external security review, deployment threat modeling, and backup/rollback evidence as required before PHI-bearing or customer production authority.

Evidence artifacts: waf-siem-sbom-review-packet
Supporting routes: /release-continuity, /production-architecture, /scrimed-cyber-defense
Owner: Infrastructure owner + security lead
External evidence needed: Deployment WAF policy, log-drain proof, SBOM artifact, vulnerability scan summary, and rollback evidence.
Human review required: true
Redaction boundary: Share evidence checklist and redacted control summaries only.
Cannot say: Do not claim deployment-specific WAF/SIEM/SBOM completion from code metadata alone. Do not expose infrastructure topology, credentials, IP addresses, or vulnerability details. Do not equate internal readiness with customer security acceptance.
Audit hash: scrimed-intel-e17388d4

Security Questionnaire Response Library: buyer_safe_answer_ready

ai_governance

How does SCRIMED govern AI outputs and prevent autonomous clinical authority?

SCRIMED frames AI outputs as governed decision support, synthetic evaluation, workflow preparation, or metadata review. High-risk clinical, payer, EHR, outreach, prescribing, diagnosis, treatment, and final imaging actions remain blocked or require human review.

Evidence artifacts: release-readiness-ladder-packet
Supporting routes: /scrimed-intelligence-safety-stack, /scrimed-agent-governance, /clinical-authority-readiness
Owner: Clinical governance + AI safety owner
External evidence needed: Clinical validation plan, external clinical governance approval, intended-use review, and customer workflow signoff.
Human review required: true
Redaction boundary: Share safety boundaries, governance gates, audit metadata, and review requirements.
Cannot say: Do not claim independent diagnostic, treatment, prescribing, payer-submission, patient-outreach, EHR-writeback, or final imaging authority. Do not claim clinical validation or regulatory clearance from synthetic evaluation. Do not remove human-review language from buyer-facing answers.
Audit hash: scrimed-intel-67785f87

Security Questionnaire Response Library: requires_external_artifact

business_continuity

What is SCRIMED's current resilience, rollback, and continuity posture?

SCRIMED maintains release-continuity metadata, rollback planning surfaces, generated-integrity checks, and workspace hygiene checks. Regulated continuity claims still require deployment-specific backup, restore, monitoring, and disaster-recovery evidence.

Evidence artifacts: release-readiness-ladder-packet
Supporting routes: /release-continuity, /service-reliability, /production-architecture
Owner: Platform reliability + release steward
External evidence needed: Backup/restore test, monitoring dashboards, incident SLA policy, and customer-approved continuity plan.
Human review required: true
Redaction boundary: Share release lanes, rollback checklists, and recovery evidence requirements only.
Cannot say: Do not claim disaster-recovery completion without tested restore evidence. Do not claim uptime commitments without signed service terms. Do not imply production customer go-live authority.
Audit hash: scrimed-intel-b35a4d1d

Security Questionnaire Response Library: blocked_until_formal_approval

vendor_risk

What evidence is required before SCRIMED integrates with EHR, payer, imaging, or device systems?

SCRIMED keeps production connector approval blocked until customer-specific threat models, least-privilege scopes, rollback plans, legal/privacy review, vendor security review, and formal customer authorization exist.

Evidence artifacts: baa-privacy-retention-packet, waf-siem-sbom-review-packet
Supporting routes: /boundary-release-approvals, /approvals-readiness, /health-records
Owner: Integration owner + customer security
External evidence needed: Connector threat model, scope map, sandbox evidence, credential-management procedure, rollback drill, and customer signoff.
Human review required: true
Redaction boundary: Share connector readiness stages, blocked actions, and external evidence requirements only.
Cannot say: Do not claim production connector approval. Do not claim EHR writeback, payer submission, or device integration authority. Do not share connector credentials, endpoints, raw payloads, or customer topology.
Audit hash: scrimed-intel-d402c1aa

Buyer question

What can SCRIMED safely demonstrate today?

SCRIMED can demonstrate synthetic/no-PHI cyber controls, browser and proxy hardening, no-secret checks, protected route fail-closed behavior, and buyer diligence evidence boundaries.

Evidence artifacts: browser-proxy-hardening-packet, no-secret-regression-packet, protected-route-fail-closed-packet, release-readiness-ladder-packet

Remaining manual step: Keep evidence packet redacted and metadata-only for all public or buyer-facing review.

Buyer question

What blocks protected no-PHI pilot expansion?

The protected no-PHI lane needs an authorized short-lived AAL2 operator happy-path run and reviewer-role verification before broader protected evidence workflows are treated as ready.

Evidence artifacts: aal2-operator-validation-packet

Remaining manual step: Run strict AAL2 durable-store smoke with a valid tenant-admin, pilot-lead, or reviewer session.

Buyer question

What blocks PHI preproduction or live production?

PHI-bearing or regulated production use remains blocked until external WAF/bot controls, SIEM/log drains, SBOM/dependency evidence, external security review, legal/privacy artifacts, and customer-specific approvals exist.

Evidence artifacts: incident-tabletop-packet, waf-siem-sbom-review-packet, baa-privacy-retention-packet

Remaining manual step: Collect deployment-specific external evidence and route it through legal, privacy, security, and clinical governance.

Incident Readiness

Response lanes keep token, protected-route, PHI-risk, and public-route events operator-owned.

Security lead + Workspace owner

Token leakage response

Credential-like value appears in output, logs, screenshot, chat, or forwarded header.

Revoke session, rotate credential, clear clipboard, verify .env.local permissions, rerun token policy self-test.

Evidence route: /qa-aal2-run-evidence
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

Platform reliability + Security lead

Protected route abuse response

Protected AAL2 route receives unauthenticated, malformed, repeated, or unexpected-role access.

Confirm fail-closed status, inspect sanitized headers, rate-limit source, preserve metadata-only evidence.

Evidence route: /workflows/execution-attempts
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

Privacy lead + Clinical governance

PHI exposure risk response

Potential live PHI or raw connector payload is submitted to a no-PHI SCRIMED path.

Block workflow, avoid echoing content, preserve only metadata, start privacy review, update regression case.

Evidence route: /health-records
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

Release steward + Infrastructure owner

Public route abuse response

Traffic spike, abnormal request pattern, bot pressure, or suspicious security header probe.

Enable stricter public rate limits, add WAF rule, run public smoke, verify headers and no-go boundaries.

Evidence route: /release-continuity
Boundary: No PHI, no production credentials, no raw connector payloads, no customer go-live approval, no production connector approval, no security certification claim, and no breach guarantee.

Next Hardening Moves

These are the next steps before regulated data, connectors, customer activation, or broader release authority.

Move 1

Attach production WAF and bot-management rules to public and protected routes.

No live PHI or ePHI authority. No production credentials or raw connector payloads in logs, UI, tests, or chat output. No final clinical diagnosis authority, treatment, prescribing, payer submission, patient outreach, EHR writeback, or final imaging interpretation. No production connector approval, customer go-live approval, security certification claim, or breach guarantee.