← Insights

AI-Driven Security Operations Platform — Interactive Target Architecture v1.0

A production reference architecture for bounded agentic AI in enterprise and MSSP security operations: a deterministic, policy-controlled platform surrounding probabilistic AI components. Gap analysis, twenty architecture layers, twelve bounded agents, response controls, evaluation gates, roadmap, runbooks, decisions, and risks.
HUM human actor SVC deterministic service AI probabilistic AI (bounded) DATA data store EXT external integration CTRL security control TRUST ZONE boundary read-only flow side-effecting flow approval gate

Executive summary

The verdict, the architectural correction, and the design position — before any component detail.

Architecture position Deterministic, policy-controlled security workflows must surround every probabilistic AI component. Multi-agent orchestration is an implementation technique — not the trust boundary, the authorization system, or the evidence model. Agents sit inside the security and assurance foundation. They do not sit above it.

Overall finding

The originating requirements are strong as a senior hybrid-engineer profile but materially incomplete as a platform specification. Building directly from them would likely produce an impressive SOC copilot or demonstration — not a defensible enterprise or MSSP security operations product.

The requirements describe what agents should do, but not:

  • What each agent and workflow is authorized to do.
  • Which evidence is trusted, untrusted, sufficient, incomplete, or contradictory.
  • How tenant isolation is enforced at every service and data boundary.
  • How findings are proven, replayed, audited, and challenged.
  • How response actions are approved, verified, reversed, and stopped.
  • What happens when models, connectors, data sources, queues, or policy services fail.
  • How quality is measured against security outcomes rather than stylistic output.
  • How the platform itself is secured, monitored, investigated, and recovered.

Core architectural correction

Do not treat “multi-agent” as the architecture. The correct pattern is a deterministic, policy-controlled security platform containing bounded probabilistic components.

Identity, policy, evidence, workflow state, approval, execution, verification, and audit remain outside model discretion — implemented by deterministic services that bind every operation to a tenant, case, workflow, and actor identity.

Deterministic shell surrounding a bounded probabilistic AI core DETERMINISTIC PLATFORM — system of authority identity · tenancy · policy · durable workflow · evidence · approval · execution · verification · audit TYPED, POLICY-ENFORCED SERVICE BOUNDARY tool broker · model gateway · context builder · schema validation · budgets · audit binding BOUNDED PROBABILISTIC CORE classify · summarize · correlate · rank · plan draft queries · draft detections · draft response plans never: authn · authz · tenancy · secrets · approval · execution · audit Untrusted evidence enters only through the context builder; actions leave only through the response plane.
20
architecture layers, L01–L20, across control, data, AI, and assurance planes
66
cataloged components — every node in the Overview is clickable
12
bounded specialized agents; none holds execution authority
R0–R5
action-risk classes, from observe-only to prohibited destructive actions
12
trust zones with enforced boundaries and separate identities
5
tenancy tiers — Standard, Enhanced, Regulated, Sovereign, Customer-managed
7
delivery phases (0–6); response authority ships last, one action at a time
0
tolerance for cross-tenant disclosure, unauthorized execution, unattributed actions

Ten mandatory design principles

P01 · Evidence before narrative

Every material conclusion identifies supporting and contradicting evidence. A persuasive summary is not a finding.

P02 · Least agency

Each agent gets the minimum tools, data, duration, context, and authority for one bounded responsibility. No universal SOC agent.

P03 · Read-only by default

Side effects require a separate deterministic response workflow. Read-only still gets scopes, quotas, masking, and full audit.

P04 · Human accountability

Sensitive actions require identifiable human approval and complete attribution — who, what evidence, which policy, what result.

P05 · Decision traces, not chain-of-thought

Store structured traces: evidence, queries, tools, versions, approvals, decisions. Hidden model reasoning is not the audit artifact.

P06 · Replayable by design

Evidence, mappings, models, prompts, agents, tools, policies, and workflows are versioned so any decision can be reconstructed.

P07 · Tenant isolation by design

Isolation exists at routing, identity, data, encryption, queue, storage, policy, observability, and operational layers — never a tenant_id column alone.

P08 · Fail closed for authority, open for visibility

Policy, approval, or audit failure stops high-risk action. Alert receipt, raw evidence, and analyst visibility continue when safely possible.

P09 · Deterministic where possible

No AI agent for work that code, rules, parsers, or queries do more reliably. Each agent must beat the deterministic baseline in benchmarks.

P10 · No unsupported claims

Nothing is called compliant or secure without stating how it is implemented and verified. Assumptions are labeled ASSUMPTION.

How to read this artifact

Press / to search. Click any node in the Overview for its contract: purpose, data, controls, failure modes, owner, open items. Amber chips mark trust-zone boundaries; violet marks probabilistic components.

Honesty register

Severity ratings, owners, runbook procedures, and ADR positions are recommended baselines from this review — not customer commitments or compliance claims. Open items are listed under Open Decisions.

Critical gap analysis

What the original requirements get right, and precisely where they are incomplete. Severity is assessed in this review: Critical = unsafe to ship without it · High = unsafe to scale without it · Medium = quality/operability debt.

12
Critical gaps — authority, isolation, evidence, injection, audit, kill switch
19
High gaps — durability, gateway, secrets, supply chain, degraded modes
11
Medium gaps — lifecycle models, memory, fine-tuning governance, SLIs
1
structural finding — one hybrid role cannot own the full platform safely
Severity

1 · Product and requirements definition

AreaCurrent positionCritical gapRequired correctionSeverity
Product boundary“AI-driven security operations platform”Supported and unsupported use cases are undefined.Define precise initial workflows: triage, enrichment, investigation assistance, response recommendation, gated execution.Critical
Operating modelEnterprise and managed environmentsEnterprise SOC vs MSSP delegation, reporting, data access, and authorization differ materially.Define enterprise, MSSP, regulated, sovereign, and customer-managed operating modes.High
Autonomy“Read-only by default”No formal autonomy or action-risk model.Define risk classes R0–R5, approval requirements, prohibited actions, kill-switch behavior.Critical
Multi-agent designAssumed foundationalNo proof that multiple agents outperform simpler workflows.Require benchmark evidence before introducing each agent or handoff.High
Investigation qualityGeneral methodologyNo completeness criteria, pivot requirements, or stop conditions.Define mandatory pivots, evidence standards, collection gaps, escalation, closure criteria.High
EvidenceMentioned indirectlyNo canonical evidence, provenance, or chain-of-custody model.Immutable evidence objects with source, query, timestamps, hashes, transformations, retention.Critical
Confidence“Escalate when confidence is low”Model self-confidence is uncalibrated and misleading.Use calibrated workflow risk scores and explicit evidence sufficiency — not model-reported confidence.High
Human accountabilityApproval gates mentionedNo responsibility or authority matrix.Define analyst, investigator, incident commander, tenant admin, approver, responder, auditor roles.Critical
Case managementNot specifiedAlerts, investigations, incidents, findings, actions lack a lifecycle model.Formal case state machine with ITSM/SOAR integration.Medium
Service objectives“Balance quality, latency, cost”No measurable SLOs or release thresholds.Workflow-specific security, reliability, performance, safety, and cost targets.High
Failure handling“Fail-safe escalation”No failure taxonomy or degraded-mode behavior.Define handling for model, connector, data, policy, audit, workflow, approval, execution failures.High
Deployment model“Major cloud platforms”SaaS, hybrid, sovereign, private, air-gapped boundaries unclear.Declare supported deployment patterns and shared-responsibility boundaries.Medium
OwnershipOne hybrid roleRole implicitly owns AI, SOC, integration, tenancy, SRE, platform security, and product controls.Cross-functional product and platform team with explicit design authority.Critical

2 · AI and agent architecture

AreaMissing requirementConsequenceRequired designSeverity
Agent contractsNo typed I/O, permissions, budgets, ownership, or stop conditions.Inconsistent behavior; untestable permissions.Versioned contracts: schemas, allowed tools, data classes, budgets, pre/postconditions, evaluation suites.Critical
Durable orchestrationFramework-level agent state implied.Work lost, duplicated, or resumed incorrectly.Durable workflow engine: checkpoints, idempotency, retries, compensation, replay, human waits.High
Agent identityNot defined.Actions cannot be strongly attributed or authorized.Unique workload identity per agent, workflow, worker, connector, actuator.Critical
MemoryNo working / case / tenant / long-term memory model.Cross-tenant leakage, poisoning, stale context, cost growth.Separate ephemeral working context from authoritative case evidence and governed long-term memory.Medium
Context constructionNot defined.Irrelevant or hostile evidence may control model behavior.Context builder that classifies trust, redacts, delimits, minimizes, binds objects to tenant and case.Critical
Prompt hierarchyNot defined.System, tenant, playbook, user, evidence instructions may conflict.Enforceable precedence outside prompts; evidence never alters system policy.High
Tool security“General tool integration.”Over-permissioned, arbitrary, unsafe calls.Typed tool broker: schema + semantic validation, scopes, quotas, egress control, audit.Critical
Model gatewayMultiple providers expected, no control point.Lock-in, inconsistent policy, uncontrolled data transfer, opaque changes.Centralized routing, policy, redaction, version pinning, fallback, budgets, provider assurance.High
Open-weight modelsMentioned for privacy/latency only.No artifact integrity, GPU isolation, patching, capacity, or acceptance criteria.Approved model lifecycle: hashes, provenance, vulnerability process, benchmarks, isolation.High
Structured outputsRequired but conflated with correctness.Syntactic validity mistaken for factual or semantic correctness.Validate syntax, semantics, evidence references, authorization, and workflow conditions separately.Medium
Hallucination controlNot specified.Plausible but unsupported findings accepted.Evidence-linked claims; visibly mark unsupported or assumed statements.High
Prompt injectionAcknowledged only generally.No untrusted-evidence security architecture.Treat logs, emails, tickets, filenames, URLs, tool responses as hostile by default.Critical
Model fallbackNot defined.Provider failure can halt incident response.Deterministic degraded mode and tenant-approved model fallback.High
Model changesNot defined.Silent provider updates alter decisions.Pin versions where possible; regression, shadow, canary, rollback, inventory updates.Medium
Fine-tuningListed as preferred experience.No tenant consent, data provenance, contamination, isolation, or rollback rules.Permit only under explicit governance and dataset lineage controls.Medium
Kill switchNot specified.Unsafe workflows cannot be rapidly stopped.Global, cell, tenant, workflow, agent, tool, model, and action-level disable controls.Critical
ReplayabilityNot specified.Decisions cannot be reconstructed after changes.Persist versioned evidence, mappings, models, prompts, tools, policies, workflow definitions.High

3 · Platform security and tenant isolation

LayerMissing capabilityRiskRequired designSeverity
Threat modelNo platform threat model.Critical trust-boundary and abuse paths remain invisible.Threat-model tenant data, identities, agents, tools, models, connectors, actions, audit, CI/CD, admin access.Critical
Human identityNo SSO, federation, MFA, provisioning, or privilege model.Weak access governance.OIDC/SAML, SCIM, phishing-resistant MFA, conditional access, JIT privilege, break-glass, session control.High
Authorization“Strong tenant isolation” is too general.Authorization devolves to app code and tenant filters.RBAC plus ABAC and policy-as-code at every service and resource boundary.Critical
Workload identityNot specified.Shared credentials; poor attribution.Short-lived workload identity with tenant-bound claims; mTLS everywhere.High
SecretsNot specified.Credentials leak into prompts, traces, logs, or code.Central secret manager, secret references, short-lived tokens, rotation, separate read/action credentials.Critical
EncryptionNot specified.Inconsistent protection; weak tenant separation.TLS/mTLS, envelope encryption, tenant DEKs, KMS/HSM roots, customer-managed keys where required.High
Network segmentationNot specified.Compromise traverses ingestion, inference, data, and execution planes.Separate the twelve trust zones with private endpoints and egress control.High
DLP and privacyNot specified.Sensitive data exposed to models, users, logs, providers.Pre-inference classification/redaction, egress policy, output filtering, minimization, residency enforcement.High
Connector compromiseNot specified.A compromised connector becomes a privileged path.Per-connector capabilities, scopes, credentials, signing, health, egress allowlists, limits, isolation.High
Insider & support riskNot specified.Privileged staff access customer security data without adequate controls.Separation of duties, customer-approved support access, recording, time limits, immutable audit.Medium
Platform incident responseNot specified.The security product itself is unprepared for compromise.Platform-specific IR plans, forensic logging, credential revocation, tenant notification, recovery exercises.High
Vulnerability managementNot specified.Unknown exposure in code, images, infrastructure, dependencies.SAST, DAST, SCA, IaC, secrets, container, runtime scanning; patch SLAs.Medium
Supply chainNot specified.Compromised models, packages, prompts, datasets, or builds.Signed builds, provenance, SBOM, AI/ML-BOM, dependency pinning, artifact admission, revocation.High
Audit protectionAuditability stated as a principle only.Incomplete or alterable records fail customer, legal, forensic scrutiny.Append-only, tamper-evident, independently retained audit and decision records.Critical
Deletion & offboardingNot specified.Residual customer data and credentials persist.Cryptographic erasure, backup expiry, legal-hold handling, access revocation, verified offboarding reports.High
Business continuityNot specified.Security operations stop during provider or regional failure.Backup, restore, regional resilience, provider fallback, queue protection, tested degraded modes.Medium

4 · SOC and detection-engineering definitions still required

  • Deduplication, correlation, suppression, maintenance windows, incident promotion.
  • Severity normalization using asset criticality, identity risk, business context, threat confidence.
  • Evidence sufficiency, contradictions, incomplete telemetry, collection-gap handling.
  • Pivot strategies, time-window expansion, query limits, source-specific validation.
  • TI source reliability, confidence, expiration, revocation, deconfliction.
  • Detection ownership, tuning, exceptions, staged deployment, rollback, health.
  • Disposition taxonomy, analyst disagreement, reopen criteria, post-incident learning.
  • MSSP customer-specific policy, delegated administration, reporting, response authorization.
  • ATT&CK mapping governance — coverage is not proof of detection effectiveness.

Where the gaps land in this architecture

Every gap above maps to a component or control in this artifact:

  • Authority and autonomy → Response Controls (R0–R5, approvals, kill switch).
  • Evidence and provenance → Data & Evidence (immutable store, schemas, lineage).
  • Injection and context → Security (context builder, untrusted-evidence controls).
  • Isolation → Multi-Tenancy (cells, tiers, keys, negative testing).
  • Quality → Evaluation (five levels, zero-tolerance metrics, release gates).
  • Reconstruction → Observability (telemetry vs audit vs decision ledger).
  • Sequencing → Roadmap (authority ships last, behind exit criteria).

The structural finding — a single hybrid role owning everything — is addressed in Roadmap → Team & ownership.

Architecture overview — the whole platform on one canvas

Click any node for its full contract: purpose, data handled, trust zone, dependencies, controls, failure modes, owner, and open decisions. Use the path toggles to trace how an alert, a response, or hostile evidence moves through the system.

Architecture rule A model or agent never receives direct database credentials, provider secrets, unrestricted vendor APIs, or standing response authority. It interacts through versioned, typed, policy-enforced services that bind every operation to a tenant, case, workflow, and actor identity.
Trace Read-only investigation and side-effecting response never share a runtime identity.
PeopleHUMAN ACCOUNTABILITY
Experience & access — L01Z1 · PUBLIC EDGE
Edge & API security — L02Z1 · PUBLIC EDGE
Identity & policy — L03 / L04Z2 · IDENTITY & ADMIN
Control plane — L05 / L06Z3 · CONTROL PLANE
Tenant data cell — bounded blast radiusZ4 · INGESTIONZ5 · DATA & EVIDENCE
Orchestration & AI — L11–L14Z6 · ORCHESTRATIONZ7 · INFERENCE
Tools & response — L15 / L16Z8 · CONNECTOR EGRESSZ9 · RESPONSE EXECUTION
Assurance backplane — L17–L20Z10 · OBSERVABILITY & AUDITZ11 · CI/CDZ12 · PRIVILEGED MGMT
External systems — customer security estate & providersOUTSIDE PLATFORM TRUST

Primary pipeline — ingestion to disposition

Receive Authenticate source Validate schema Bind tenant + source Persist raw Deduplicate Normalize (OCSF) Enrich Classify Correlate Route
Deterministic triage AI triage when justified Plan investigation Bounded queries via broker Validate evidence Structured findings Evidence sufficiency Recommend disposition Human review / policy closure Response plane

Control plane and data plane are separated; read-only investigation (Z8) and side-effecting response (Z9) run with different identities, credentials, and network paths.

Architecture layers L01–L20

The complete layer catalog. Expand a layer for its components — chips open the same detail panel as the Overview. Components without a dedicated catalog entry are listed as capabilities of their parent service.

L01 · User experience & analyst workspace Z1

Consoles for analysts, incident commanders, MSSP operators, tenant admins, detection engineers, evaluators, and auditors. The case timeline visibly separates source evidence, normalized evidence, AI inference, deterministic rules, assumptions, recommendations, and executed actions. AI text never renders as verified fact unless linked to validated evidence.

Analyst & IC consoleMSSP workspaceTenant admin portalDetection workspaceEvaluation dashboardAudit portalAPI & CLI

L02 · Edge & API security Z1

DNS, CDN where appropriate, DDoS protection, WAF, API gateway, rate limiting, bot/abuse protection, request validation.

Edge securityAPI gateway

L03 · Human identity & access management Z2

OIDC/SAML federation, SCIM provisioning, phishing-resistant MFA, RBAC + ABAC, conditional access, PAM, JIT elevation, break-glass with recording and separate audit.

Identity federationPAM & break-glassPolicy decision point

L04 · Workload & agent identity Z2

Unique identity for every service, agent, workflow, connector, and actuator; short-lived credentials, mTLS, workload identity federation. Authorization context includes subject, tenant, region, environment, severity, action-risk class, approval state, and data classification.

Workload & agent identity

L05 · Control plane Z3

Tenant lifecycle, entitlements, regional placement, and the registries that make every AI artifact governed: connectors, agents, prompts/playbooks (signed), models (with provider policy), policies, flags, evaluation, and the governance inventory (SBOM, AI/ML-BOM, attestations).

Tenant managementConnector registryAgent registryPrompt registryModel registryPolicy administrationFlags & entitlementsEvaluation controlGovernance inventory

L06 · Tenant & cell architecture Z3→Z5

Tenant router, regional data cells, tenant data-plane boundary, tenant encryption keys, tenant-scoped queues, storage, credentials, and quotas. See Multi-Tenancy for tiers and isolation mechanics.

Cell routerTenant managementTenant keys (KMS)

L07 · Integration & connector layer Z4/Z8

SIEM, EDR/XDR, identity, cloud, email, network, TI, ITSM/SOAR, CMDB/asset, and collaboration connectors — each with versioned capabilities, scopes, per-tenant credentials, health, limits, and side-effect classification. See Integrations.

Connector runtimeConnector registry

L08 · Ingestion & messaging Z4

Webhook receiver, polling workers, streaming ingestion, durable event bus, dead-letter queues, schema validation, deduplication, and backpressure management.

Webhook & pollersEvent busValidation & dedup

L09 · Normalization & enrichment Z4

OCSF normalization with vendor-field preservation, schema-mapping registry, entity extraction, asset/identity/TI enrichment, and data-quality scoring. A common schema improves correlation; it does not erase vendor semantics.

OCSF normalizerEnrichment

L10 · Data & evidence Z5

Immutable raw evidence store, hot search/event store, PostgreSQL transactional store (cases, findings, approvals), object storage, cache, optional vector index, lineage, retention, and legal hold. See Data & Evidence.

Raw evidence storeSearch storeCase DBCacheVector indexLineage & retention

L11 · Durable workflow orchestration Z6

Persisted state, checkpointing, retries, timeouts, compensation, human-approval waits, cancellation, replay, and workflow versioning. The engine — not an agent framework loop — owns investigation state.

Workflow engine

L12 · Context & evidence construction Z6

Context builder, evidence trust classifier, prompt-injection detection, content sanitization, redaction, retrieval policy, context limits, and evidence references. The only doorway between untrusted evidence and any model.

Context & evidence builderDLP & classification

L13 · Agent services AI Z6

Twelve specialized bounded agents — triage, correlation, investigation planning, evidence collection, identity, endpoint, cloud, threat-intel, detection drafting, response planning, reporting, quality review. See Agents.

Specialized agent services

L14 · Model gateway & inference Z7

Provider abstraction, model routing, data-classification and residency enforcement, allowlisting, version pinning, token/cost budgets, fallbacks, circuit breakers, private endpoints, open-weight serving.

Model gatewayInference endpointsModel registry

L15 · Policy enforcement & tool broker Z8

PDP/PEP, typed tool registry, JSON-Schema + semantic validation, tenant binding, credential resolution at call time, query guardrails, rate limits, result filtering, and audit on every call.

Typed tool brokerQuery compilerPDP

L16 · Response planning & execution Z9

Action-risk classification, response planner, target validation, dry-run, approval workflow, dual control, short-lived execution tokens, actuator, post-action verification, rollback, kill switch. See Response Controls.

Response plannerApprovalsActuatorVerificationKill switch

L17 · Observability & audit Z10

OpenTelemetry traces/metrics/logs, token and cost accounting, agent failure analysis, security audit log, case decision ledger, tamper-evident storage. See Observability.

ObservabilityAudit logDecision ledger

L18 · Security platform services Z10/Z12

KMS/HSM, secrets management, certificate management, DLP, malware scanning, runtime security, vulnerability management, privileged-access monitoring.

KMS & secretsDLP & egress

L19 · DevSecOps & supply chain Z11

Source control, CI/CD, IaC, policy-as-code, detection-as-code, SAST/DAST/SCA, secret scanning, container and IaC scanning, SBOM, AI/ML-BOM, signed artifacts, build provenance (SLSA).

DevSecOps & supply chainGovernance inventory

L20 · Reliability & operations Z10

SLO management, capacity, autoscaling, backup, disaster recovery, restore testing, chaos testing, incident management, degraded operating modes, cost management.

Reliability operations

Agent architecture — a small set of bounded specialists

More agents are not better. An agent exists only where it measurably beats deterministic code or a simpler model call — proven in benchmarks before release. No agent holds execution authority, standing credentials, or cross-tenant scope.

Two prohibitions There is no general-purpose SOC agent with unrestricted access across SIEM, EDR, cloud, identity, email, network, ticketing, and response systems. And natural-language agent-to-agent conversation is not the control protocol — every handoff is a typed, authenticated, expiring message.

Triage agent read-only

Classifies and prioritizes alerts against tenant context.

Why an agent: deterministic dedup and rules run first; the agent adds value only on ambiguous residue.

Budgets: ≤10 steps · ≤60 s · capped tokens/cost per alert.

Escalates: insufficient evidence, contradiction, policy uncertainty.

Correlation agent read-only

Finds related alerts, entities, and cases across the tenant scope.

Why an agent: entity-graph joins are deterministic; semantic similarity across noisy alerts is not.

Budgets: bounded lookback and result counts via broker.

Escalates: cross-case links touching other tenants → refused by design.

Investigation planner no direct tool execution

Selects evidence-collection steps and pivots for the case type.

Why an agent: playbook selection under partial information benefits from reasoning; execution stays deterministic.

Output: typed evidence plan; every step re-authorized at the broker.

Evidence collection agent read-only · strictly scoped

Executes approved, bounded queries from the plan via canonical query intent.

Never: raw vendor query strings to production systems.

Budgets: per-query lookback, row, volume, and cost caps.

Identity investigation agent read-only

Analyzes authentication, privilege, session, and directory activity.

Data limits: identity telemetry for the bound tenant only; PII minimized before inference.

Endpoint investigation agent read-only

Analyzes process, file, network, and endpoint telemetry from EDR/XDR.

Injection note: command lines and filenames are hostile input — delimited and never interpreted as instructions.

Cloud investigation agent read-only

Analyzes control-plane, workload, storage, and identity activity in Azure, AWS, GCP, Kubernetes.

Threat-intelligence agent read-only

Contextualizes indicators with source reliability, confidence, and expiry.

Rule: external intelligence informs; it never becomes automatically authoritative.

Detection agent draft only

Drafts or translates detection content (Sigma → KQL/SPL/CQL).

Rule: conversion is not semantic proof — drafts enter the detection-as-code pipeline, never production directly.

Response planning agent no execution

Creates action, impact, verification, and rollback plans.

Boundary: the plan is inert until the deterministic response plane validates, approves, and executes it.

Reporting agent read-only

Creates evidence-linked summaries and timelines.

Rule: every claim carries evidence references; unsupported statements are visibly marked.

Quality review agent read-only

Checks findings for evidence coverage, contradictions, and policy compliance before human review.

Rule: one signal among several — never the release authority.

Mandatory agent contract (versioned, in the agent registry)

agent_id · version · owner · purpose
approved_workflows
input_schema · output_schema
allowed_tools · denied_tools
allowed_data_classifications · tenant_scope
maximum_steps · maximum_runtime
maximum_input_tokens · maximum_output_tokens · maximum_cost
required_evidence · preconditions · postconditions
stop_conditions · escalation_conditions
failure_behavior · side_effect_class
evaluation_suite · release_state

An agent version cannot reach production without its contract, its evaluation suite passing, and a named owner.

Typed inter-agent handoff (the only agent-to-agent channel)

tenant_id · case_id · workflow_id · task_id
requested_result
evidence_refs[]        — references, never raw blobs
tool_permissions[]     — subset of sender's own grants
constraints            — lookback · volume · cost · time
expiration             — every handoff expires
required_output_schema
sender_identity        — authenticated workload identity

Handoffs are events on the durable workflow engine — checkpointed, replayable, and attributable. A receiving agent can hold fewer permissions than the sender, never more.

Where AI is deliberately not used Deduplication, schema validation, tenant binding, authorization, budget enforcement, approval routing, idempotent execution, verification checks, and audit writes are deterministic code. Each is cheaper, faster, and testable — and none may ever depend on a model's judgment (P09).

Integration architecture

Every integration is a governed connector with versioned capabilities, per-tenant credentials, and a declared side-effect class. Not every named platform ships in the first release — the matrix below is the priority order.

Connector rule Agents never call vendor APIs directly. All traffic passes the typed tool broker: tenant binding → schema and semantic validation → policy decision → credential resolution at call time → rate/volume/cost limits → result filtering → evidence registration → audit. Side-effecting connector operations additionally require an approval token from the response plane.
Phase
CategoryRepresentative platformsPatternCapability classPhaseNotes
SIEMMicrosoft Sentinel · Splunk · Elastic · Google SecOpsPoll + webhook + streamingR0 read · R2 case syncP1Start with one SIEM end-to-end; certify others against the same connector conformance suite.
EDR / XDRDefender XDR · CrowdStrike · SentinelOne · CortexStreaming + APIR0 read · R3 isolate (P5)P1One EDR in MVP. Response verbs stay disabled until Phase 5 gates pass.
CollaborationSlack · Teams · email notificationsOutbound webhookR1 notify/draftP1Draft-only notifications; no inbound instructions accepted from chat.
IdentityEntra ID · Okta · Active DirectoryAPI + event streamR0 read · R3/R4 revoke (P5)P2Session revocation and account disable are R3/R4 verbs behind approvals.
Cloud platformsAzure · AWS · GCP · Kubernetes auditNative audit feedsR0 readP2Control-plane, workload, storage, identity activity for cloud pivots.
Threat intelligenceSTIX 2.1 / TAXII 2.1 feeds · commercial TITAXII pollR0 readP2Reliability, confidence, expiry, revocation, and decay preserved per source.
ITSM / SOARServiceNow · Jira · existing SOARAPI, bidirectionalR1 draft · R2 ticket opsP2Case lifecycle sync; the platform complements rather than replaces existing SOAR early on.
CMDB / assetCMDB · asset inventory · DSPMScheduled syncR0 readP2–3Asset criticality and ownership feed severity normalization.
Email securityEmail security gatewaysAPIR0 read · R3 quarantine (P5+)P4+Email bodies are prime injection carriers — sanitization is mandatory before context.
Network securityFirewall · DNS · proxy/SASE · NDRAPI + streamingR0 read · R3 temp block (P5+)P4+Firewall-wide changes are R4 dual-control; broad blocks never automated.
Vulnerability mgmtVM platformsScheduled syncR0 read · R2 request scanP4+Context for exposure-aware prioritization.
SaaS auditSaaS audit-log sourcesPollR0 readP4+Extends coverage to business-critical SaaS.

Connector capability model

  • Declared capabilities — each verb registered with schema, side-effect class (R0–R5), and cost profile.
  • Authentication — per-tenant credentials from the secret manager, resolved at call time; OAuth client credentials or scoped API keys; separate read vs action credentials.
  • Health model — liveness, auth validity, quota headroom, schema drift, and lag surfaced per tenant.
  • Rate & retry strategy — vendor-aware budgets, exponential backoff with jitter, circuit breakers, dead-letter queues; ingestion favors at-least-once with deterministic dedup.
  • Versioning & certification — connectors are versioned artifacts passing a conformance suite (schema, limits, failure injection, cross-tenant negative tests) before release.

Safe query generation — the only path for model-generated queries

NL investigation intent Typed canonical query intent Vendor compiler Parser / AST validation Tenant + policy checks Time-range + cost estimate Read-only execution Result validation Evidence registration

Guardrails: operation allowlists, prohibited statements, maximum lookback, result-count and volume caps, timeouts, cancellation, per-tenant rate limits, query-cost monitoring, and full attribution. Unrestricted LLM-generated KQL/SPL/ CQL/SQL never reaches a production system.

Data & evidence model

Evidence is a first-class domain object with provenance and chain of custody — not text in a prompt. The original event is always preserved, even after normalization.

StoreHoldsExplicitly not forNotes
PostgreSQL (system of record)Cases, findings, workflow metadata, configuration, approvals, transactional stateBulk event storageRow-level security + tenant keys; the authority for case truth.
Immutable object storeRaw evidence, reports, large query results, replay packages, legal holdsMutation of any kindWORM/object-lock, cryptographic hashes, strongest durability tier.
Hot search / event storeInteractive investigation queries over normalized eventsBeing the system of recordRebuildable projection; tenant-isolated indexes where justified.
RedisCache, rate control, short-lived coordinationDurable investigation stateLoss is a performance event, never a correctness event.
Durable message busAsynchronous processing, connector decoupling, DLQsLong-term retentionTenant-partitioned topics; backpressure managed.
Vector index optionalApproved approximate-retrieval use cases onlyAuthoritative evidence storageEvery hit re-validated against the source of record before use.

Minimum evidence record

evidence_id · tenant_id · case_id · alert_id
source_system · source_record_id
source_query · query_parameters
event_time · observed_time · ingestion_time
raw_object_uri · raw_hash
normalized_schema · normalization_version
content_classification · provenance · transformations
collection_status · retention_policy · legal_hold
redaction_status · access_policy

Three timestamps are deliberate: when it happened, when the source saw it, when the platform received it. Divergence between them is itself investigative signal.

Minimum finding record

finding_id · claim · finding_type · status
supporting_evidence_ids[] · contradictory_evidence_ids[]
assumptions · collection_gaps · attack_mappings[]
severity · calibrated_risk_score · recommended_action
agent_version · model_version · prompt_version · policy_version
review_status · reviewer · disposition

Contradictory evidence and collection gaps are mandatory fields — their absence must be asserted, not implied. Model-reported confidence is never the risk score; calibration is empirical per workflow.

Evidence provenance flow

Source event Raw persisted + hashed OCSF normalized mapping version pinned Enriched provenance appended Context builder trust-classed · redacted Model inference Finding cites evidence_ids Decision ledger entry

Every transformation appends to lineage. Replay pins the evidence set, mapping version, model, prompt, agent, tool, policy, and workflow definition — so any past decision can be reconstructed exactly (P06).

Retention, legal hold, and deletion Retention policy is tenant-configurable per data class; legal hold overrides expiry and blocks cryptographic erasure until released. Offboarding uses key destruction for bulk erasure plus verified deletion of extant replicas and backup expiry — with a written offboarding report. See Multi-Tenancy.

Platform security architecture

The platform guards other people's security data — it is a high-value target itself. Twelve trust zones, a hostile-evidence pipeline, engineered secret handling, and an explicit threat model.

Trust zones — separate identities, networks, and credentials

Z1Public edgeDNS, CDN, DDoS, WAF, gateway. Only zone facing the internet.
Z2Identity & administrationHuman IdP federation, PAM, workload identity issuance.
Z3Control planeTenant mgmt and registries. No bulk tenant security data.
Z4Tenant ingestionConnectors, queues, validation, normalization.
Z5Tenant data & evidenceRaw store, search, case DB — tenant keys, RLS.
Z6Agent orchestrationWorkflow engine, context builder, agent runtimes.
Z7Model inferenceGateway plus hosted/private/local endpoints.
Z8Connector egressRead-path calls to customer systems; egress allowlists.
Z9Response executionActuator only. Separate identity, network, credentials.
Z10Observability & auditTelemetry, tamper-evident audit, decision ledger.
Z11CI/CD & buildPipelines, signing, artifact admission.
Z12Privileged managementBreak-glass and support access — recorded, time-boxed.
Zone rule A response actuator (Z9) must not share broad credentials, network paths, or runtime identity with a read-only investigation worker (Z8). Zone crossings are explicit, mutually authenticated, and audited.

Untrusted evidence & prompt-injection containment

All external and customer-generated content is potentially hostile: logs, emails, tickets, filenames, URLs, command lines, process arguments, threat reports, endpoint artifacts, cloud metadata, tool responses, retrieved documents.

  • Four instruction classes with enforced precedence outside prompts: system policy → tenant configuration → analyst instruction → untrusted evidence. Evidence can never modify policy, authority, tenancy, tool permissions, or approval requirements.
  • Structural delimiting; active content stripped; embedded-instruction and suspicious-encoding detection.
  • Length, nesting, recursion, and retrieval-volume limits (context flooding).
  • Credential and unnecessary-PII removal before inference; source, hash, provenance preserved.
  • Every retrieved object bound to the current tenant, case, and authorization decision.
  • Output validation on the way back: schema, evidence references, authorization, unsupported-claim marking.
Prompt-injection reality A system prompt that says “ignore malicious instructions” is not a control. Injection defense = trust separation + context construction + least privilege + typed tools + deterministic policy + output validation + monitoring + adversarial testing.

Secrets, keys, and supply chain

Secrets & keys

  • Secret references in configuration — never values; resolution immediately before use; short-lived credentials preferred.
  • No credentials in prompts, model context, traces, or case narratives — enforced by scanning, not convention.
  • Separate read-only vs action credentials; per-tenant DEKs under KMS/HSM roots; customer-managed keys where contracted.
  • Automatic rotation; revocation on offboarding or compromise; independent secret-access audit with anomaly alerts.

Software & AI supply chain

  • SBOM for application, container, and infrastructure dependencies; AI/ML-BOM for models, adapters, datasets, evaluation sets, prompts, and agent configurations.
  • Signed images, prompt/playbook packages, and model artifacts; admission verifies provenance (SLSA-aligned).
  • Pinned model artifact hashes for local serving; approved provider/model inventory; license and vulnerability process; revocation and emergency rollback paths.

Threat model — minimum scenario set

ScenarioPrimary architectural counterFramework lens
Compromised tenant administrator or analystABAC + conditional access, JIT elevation, dual control on R4, session risk, behavioral auditATT&CK (valid accounts)
Malicious insider or support engineerSeparation of duties, customer-approved support access, recording, time limits, immutable audit (Z12)NIST 800-53 AC/AU
Compromised connector credentialsPer-tenant scoped credentials, egress allowlists, health anomaly detection, rapid revocation runbookATT&CK (trusted relationship)
Malicious alert / log / ticket / email / filename / URL / tool responseContext builder trust classing, sanitization, delimiting, output validation, injection metricsOWASP LLM01 · ATLAS
Cross-tenant attacker / confused deputyCell isolation, tenant-bound claims on every call, RLS, per-tenant keys, continuous negative testsOWASP LLM · multi-tenancy
Compromised model provider or poisoned open-weight modelGateway allowlist + pinning, artifact hashes, shadow comparison, provider health, emergency model disableATLAS (supply chain)
Poisoned embedding corpus or evaluation datasetDataset inventory + lineage in AI/ML-BOM, non-authoritative vector index, eval-set isolationATLAS (data poisoning)
Malicious agent tool / prompt / playbook / package updateSigned registries, review + release gates, admission control, revocationSLSA · OWASP agentic
Compromised CI/CD pipeline or artifact registryBuild provenance, signed artifacts, admission verification, separated build zone (Z11)SLSA
Audit suppression or tamperingAppend-only tamper-evident store, independent retention, audit-write failure blocks R3+ executionNIST 800-53 AU-9
Response-target substitution / actuator abuseTarget re-validation at execution time, short-lived single-action tokens, post-action verification, kill switchOWASP agentic (excessive agency)
Token / API / GPU / queue / cost exhaustionBudgets at gateway and broker, quotas per tenant, circuit breakers, backpressure, cost alertingOWASP LLM10
Control-plane compromise / data-residency violationControl-data plane separation, regional cell routing, residency enforcement at gateway, no bulk data in Z3NIST 800-207

Multi-tenant & cell architecture

Cell-based tenancy bounds the blast radius of any failure or compromise, simplifies regional isolation, and creates a credible path from shared SaaS to sovereign and customer-managed deployments.

Not negotiable A tenant_id column is not an isolation architecture. Isolation is enforced at routing, identity claims, API authorization, queue partitioning, database row-level security, object prefixes and policies, search indexes, encryption keys, connector credentials, model policy, response policy, quotas, and audit — and proven by cross-tenant negative tests on every release plus continuous runtime canaries.
TierDesignTypical driver
StandardShared cell, strong logical isolation: tenant-aware routing, RLS, quotas, tenant-specific encryption keysEnterprise and MSSP customers at standard assurance
EnhancedDedicated databases, indexes, queue partitions, worker pools inside a shared cloud boundaryNoise isolation, elevated assurance, MSSP anchor tenants
RegulatedDedicated cell and dedicated cloud account/subscription/project, private network, approved inference endpointsRegulated industries; auditor-driven separation
SovereignRegion-restricted dedicated cell, controlled support access, locally approved providers and modelsData-residency and sovereignty mandates
Customer-managedData plane in the customer environment; narrow, authenticated management channel; no standing platform access to dataMaximum-control customers; air-gap-adjacent needs

Defense-in-depth checklist (all tiers)

  • Tenant-aware routing and tenant-bound workload claims on every internal call.
  • Separate queue topics/partitions and per-tenant quotas — no noisy-neighbor cascade.
  • Database row-level security plus resource-level authorization at the API.
  • Tenant object prefixes, per-tenant search isolation where justified, tenant DEKs.
  • Per-tenant connector credentials; per-tenant model and response policy.
  • Cell-level capacity and failure containment; cells fail independently.
  • Cross-tenant negative tests in CI and continuously in production.

Offboarding & verified deletion

Offboard request Legal-hold check Access + credential revocation Tenant key destruction cryptographic erasure Replica + index purge Backup expiry schedule Verified offboarding report

MSSP mode adds delegated administration: the MSSP operates many customer tenants under explicit per-customer context, per-customer policy, response authorization, and reporting — with every delegated act attributed to both the operator identity and the customer tenant.

Response controls — the action-risk model

Response planning and response execution are separate services with separate identities. The planner has no execution authority; the actuator cannot reinterpret the plan.

ClassExamplesDefault authorityAdditional requirements
R0 · ObserveQuery SIEM, EDR, identity, cloud, case dataAutomatic under read policyScopes, quotas, masking, full audit — read-only is not a free pass
R1 · RecommendDraft ticket, summary, containment plan, analyst messageAutomatic draftDrafts visibly marked; evidence-linked claims required
R2 · Administrative reversibleTag case, assign ticket, request approved scan, enrich assetTenant policy-controlledTenant authorization policy; full attribution
R3 · Security reversibleTemporarily isolate endpoint, revoke session, temporary indicator blockExplicit human approvalTarget validation, TTL, verification query, rollback plan
R4 · High impactDisable account, broad credential revocation, org-wide or firewall-wide changeTwo-person approvalDual control, dry-run where supported, expected business impact recorded
R5 · Destructive / irreversibleWipe device, delete asset, permanent data deletionProhibited in-platformExternal emergency procedure only, outside platform authority

Required control sequence for every side-effecting action

Plan response planner — inert Risk-classify (deterministic) Evidence sufficiency Target re-validation Dry-run / simulate where supported Approve human · dual for R4 Short-lived execution token Execute idempotent · actuator only Verify intended state Rollback if verification fails Audit + ledger

Every action defines: required evidence, required role, approval level, dual-control requirement, target validation, expected business impact, TTL, dry-run support, execution method, idempotency behavior, success criteria, verification query, rollback method, and escalation path — before it is ever enabled for a tenant.

Non-negotiable action rule An audit-write failure, authorization ambiguity, stale target identity, missing rollback method, or failed post-action verification blocks or immediately escalates high-risk execution. The platform fails closed for authority.

Kill-switch hierarchy

Independent disable controls, each tested in release gates and exercised in drills:

Global Cell Tenant Workflow Agent Tool Model Action type

Activation is deterministic, requires no model availability, and is itself audited. Investigations queue; raw alert receipt continues.

Detection engineering — detection as code

Detections are versioned, tested, staged artifacts with owners and health monitoring — not strings pasted into a SIEM console.

Pipeline

Threat requirement Detection hypothesis Telemetry check Rule authoring AI drafts allowed Static validation Query compilation Unit tests Historical replay
Performance tests False-positive tests ATT&CK mapping Peer approval Shadow Canary General release Health monitoring Tuning or rollback

Repository structure

/detections
  /sigma          /kql          /spl
  /cql            /vendor-native
/tests
  /positive       /negative     /performance
/schema
/mappings
  /attack         /ocsf         /vendor
/playbooks
/exceptions

Code ownership per directory; review requirements enforced in CI; exceptions are versioned artifacts with expiry — not tribal knowledge.

Threat-intelligence integration

  • STIX 2.1 for structured threat and observable information; TAXII 2.1 for exchange where supported.
  • Source reliability, confidence, expiration, revocation, relationships, provenance, and tenant policy preserved.
  • Indicator decay, deconfliction, and suppression of stale or low-quality intelligence.
  • External intelligence informs investigations — it never becomes automatically authoritative.

Detection health

Per-rule: fire rate, FP feedback loop from case dispositions, telemetry-gap tracking, latency and cost, staleness, deprecation workflow, and one-click rollback to the prior version.

Blunt caution — conversion is not semantic proof A Sigma rule converted to KQL, SPL, CQL, or a vendor-native query must be re-validated against destination field mappings, data availability, time semantics, aggregation behavior, and query performance. Portable formats travel; meaning does not travel for free.

Evaluation, testing & release governance

Golden datasets and LLM-as-judge are necessary but not sufficient. Evaluation spans components, workflows, adversarial conditions, shadow production, and controlled release — measured against security outcomes, not stylistic quality.

LevelRequired coverage
1 · ComponentPrompts, structured-output schemas, policy checks, query compilation, tool selection, evidence citation, normalization mappings, connector authorization.
2 · WorkflowTrue/false positives, duplicates, missing logs, conflicting evidence, stale intelligence, source outage, provider outage, rate limits, partial completion, human disagreement.
3 · AdversarialDirect/indirect injection, poisoned retrieval, malicious tool responses, tool-schema manipulation, cross-tenant requests, context flooding, cost exhaustion, loop induction, credential extraction, audit suppression, response-target substitution.
4 · Shadow productionRun against live alerts without changing dispositions or executing actions; compare with analyst outcomes at scale.
5 · Controlled productionEnable one narrowly defined outcome at a time, after workflow-specific evidence and rollback validation.

Metric domains

  • Security outcomes — true/false-positive triage rates, critical false negatives, escalation accuracy, evidence completeness, pivot coverage, time to correct disposition, time to validated containment, override rate, reopen rate.
  • Agent performance — correct tool choice, query validity/relevance, structured-output validity, grounded-claim rate, citation correctness, unsupported claims, completion, loops, retries, latency, cost per case.
  • Safety — unauthorized tool attempts, cross-tenant retrieval, sensitive-data disclosure, injection success rate, policy bypass, unsafe recommendations, unauthorized actions, rollback failures, audit completeness, unattributed actions.
  • Reliability — connector availability, queue delay, workflow completion, model-fallback rate, degraded-mode success, restore results, cell saturation, P50/P95/P99 latency.
Zero-tolerance measures Cross-tenant disclosure, unauthorized response execution, and unattributed security actions have a target of zero. A low percentage is not an acceptable mature-state objective for these control failures.

LLM-as-judge restrictions

  • One evaluation signal — never the sole production release authority.
  • Explicit rubrics, deterministic checks, human adjudication, judge-to-human agreement measured.
  • Candidate identity hidden; ordering randomized to reduce preference bias.
  • Judges versioned and revalidated after any change.
  • Development, regression, adversarial, and final holdout sets kept separate; no training or prompt leakage from evaluation data.

Release gate — all ten, every release

  • 1 · Regression and adversarial suites pass.
  • 2 · No critical safety regression.
  • 3 · Schema and connector compatibility pass.
  • 4 · Permission changes separately reviewed and approved.
  • 5 · Cost and latency within budget.
  • 6 · Human-review agreement meets the workflow threshold.
  • 7 · Rollback and kill switches tested.
  • 8 · SBOM and AI/ML-BOM inventories updated.
  • 9 · Audit and observability fields complete.
  • 10 · A named owner accepts the release.

Observability, audit & decision reconstruction

Three record categories with different consumers, retention, and integrity requirements — deliberately not one log.

Record categoryPurposeIntegrity requirement
Operational telemetrySRE and engineering: latency, failures, resources, cost, reliabilityBest-effort; sampled; short retention
Security audit logWho or what accessed, changed, approved, or executed a protected operationAppend-only, tamper-evident, independently retained
Case decision ledgerReconstruct evidence, findings, policies, approvals, actions, verification, disposition per caseComplete per case; replay-grade; legal-hold aware

OpenTelemetry instrumentation

Spans for workflow, agent, model call, tool call, connector, policy decision, approval, response execution, verification, and rollback — with token, cost, retry, and failure-category attributes.

tenant.id | case.id | alert.id
workflow.id | workflow.version
agent.id | agent.version
model.provider | model.name | model.version
prompt.version | tool.name | tool.version | policy.version
action.risk_class | approval.id | evidence.count
token.input | token.output | cost.estimated
result.status | failure.category

Default: sensitive prompt and completion content is disabled or redacted in production telemetry.

Minimum security audit event

  • Tenant and actor type/identity — human session, service, workflow, and agent identity.
  • Source, destination, requested operation, protected resource.
  • Policy version, inputs, decision, and any exception applied.
  • Approval identity and scope where applicable.
  • Evidence references and data classification.
  • Tool, connector, model, and prompt versions.
  • Action result, verification result, before/after state.
  • Trusted timestamp, trace identifier, integrity proof.

Audit-write failure is a platform incident — and blocks R3+ execution while it lasts.

Standards & framework baseline

Standards are used pragmatically, mapped to concrete components with a validation method. Referencing a framework is not compliance; no certification is claimed here.

DomainBaselineUsed forLands inValidated by
AI governanceNIST AI RMF + GenAI ProfileGovern, map, measure, manage AI risk across the lifecycleGovernance inventory, evaluation controlGovernance review
Cyber governanceNIST CSF 2.0Enterprise cybersecurity outcomes and governanceControl mappings, risk registerControl assessment
Incident responseNIST SP 800-61r3IR lifecycle integrated with CSF 2.0Case lifecycle, platform IR runbooksIR exercises
Security controlsNIST SP 800-53r5Control catalog and assessment basisSecurity-control matrixInternal audit
Zero trustNIST SP 800-207/207AIdentity-, workload-, resource-centric accessPDP/PEP, workload identity, zonesAccess reviews, pen tests
Secure developmentNIST SSDF 800-218/218ASecure SDLC including AI-specific practicesDevSecOps pipelinePipeline evidence
Adversary behaviorMITRE ATT&CKTactics, techniques, detection mapping — not proof of coverageFindings, detection mappingsMapping governance
AI threatsMITRE ATLASThreats and attack paths for AI/agentic systemsThreat model, adversarial suitesRed-team exercises
LLM / agent securityOWASP GenAI & AgenticInjection, excessive agency, tools, memory, supply chainContext builder, broker, evaluationAdversarial testing
Event schemaOCSFVendor-neutral normalized security eventsNormalization layerMapping-quality scoring
Threat intelSTIX 2.1 / TAXII 2.1Structured threat information and exchangeTI connectors and storeFeed conformance
Detection portabilitySigmaPortable detection source requiring destination validationDetection-as-code repoReplay + FP testing
PlaybooksCACAO 2.0Structured security playbooksPlaybook registrySchema validation
Response commandsOpenC2Standard command semantics where integrations support themResponse actuator verbsConnector certification
ObservabilityOpenTelemetryVendor-neutral traces, metrics, logs, GenAI telemetryObservability layerTrace completeness checks
Supply chainSLSA · CycloneDX SBOM / AI-BOMBuild provenance; software, model, dataset inventoryCI/CD, governance inventoryAdmission verification
Honesty note “Lands in” and “Validated by” describe the intended implementation and its evidence of operation — they become true when built and verified, not by appearing in this table. Formal certifications (e.g., SOC 2, ISO 27001) are a separate, later program with their own evidence trail.

Implementation roadmap — authority ships last

Each phase ships value and retires a named risk. No autonomous response exists until Phase 5, and then only one narrowly defined action at a time. Effort bands are relative for a properly staffed team: M ≈ 1–2 months · L ≈ one quarter · XL ≈ multi-quarter.

#PhasePrincipal deliverablesRetires riskExit conditionEffort
0Architecture & safety foundationProduct boundary; threat model; data classification; tenant-isolation design; evidence/case schemas; action-risk model; agent contracts; provider policy; evaluation strategy; SLOs; responsibility matrixBuilding the wrong system safelyNo coding on autonomous response until authority and evidence models are approvedM
1Read-only single-workflow MVPOne SIEM + one EDR; alert ingestion; OCSF normalization; deterministic dedup; triage; evidence-linked summary; analyst feedback; telemetry; audit; shadow modeUnproven value; ungrounded summariesDemonstrated improvement over analyst baseline without unacceptable safety or evidence defectsL
2Investigation platformDurable workflows; specialized agents; typed query layer; case management; TI; identity/cloud pivots; evidence sufficiency; replay; model gateway; adversarial setsLost state; unbounded queries; provider lock-inRepeatable quality across supported scenarios and data-failure conditionsL
3Multi-tenant & MSSP controlsCell architecture; regions; tenant keys; MSSP delegation; per-tenant policy/models/quotas; dedicated tier; cross-tenant testing; offboardingCross-tenant exposure; unbounded blast radiusIndependent validation of isolation and blast-radius controlsXL
4Detection engineeringDetection-as-code; Sigma + vendor compilers; historical replay; performance and FP testing; shadow/canary; ATT&CK mapping; telemetry gaps; health monitoringUntested detections; silent rule driftSafe staged release and measurable detection qualityL
5Gated responseTicketing; notifications; temporary block; session revocation; endpoint isolation — with dry-run, approval, idempotency, verification, rollback, kill switchWrong-target and unauthorized actionsNo high-impact autonomous execution; expand one action at a timeL
6Advanced optimizationLocal model serving; ensembles; fine-tuning where justified; low-risk automation; cross-case learning; adaptive routing; MSSP analytics; customer-managed data planesCost/latency ceilings; provider dependenceOnly after safety, isolation, evidence, reliability, and operations are matureXL

Service objectives (starting targets, not contracts)

0
unauthorized cross-tenant access · unauthorized response actions — zero-tolerance
100%
security actions attributed and audited
99.9%
initial core availability; 99.95% after operational maturity
<30s
P95 high-severity alert acceptance; triage initiation <60s
<15s
P95 interactive query, excluding source-system latency
<5m
preferred model-provider failover
Quarterly
restore testing, minimum cadence
Every release
cross-tenant isolation tests, plus continuous runtime canaries

Final SLOs follow customer risk, deployment mode, source-system behavior, and commercial tiers. Degraded modes are defined for: preferred model unavailable, all external models prohibited, SIEM/EDR rate-limited or down, stale TI, cache loss, search degraded, audit unavailable (blocks R3+), and regional cell loss — with deterministic triage, queued investigations, disabled automation, and preserved raw evidence.

Team & ownership — the staffing reality

Structural finding The originating role compresses at least three senior roles — AI/agent engineer, detection/IR engineer, and platform/security architect. One exceptional hybrid engineer can lead the core architecture. One person should not be expected to safely build and operate the full product.
FunctionCore ownership
Product security leadThreat model, isolation, secure design, abuse cases, risk acceptance
AI platform leadModel gateway, agents, evaluation, prompts, model lifecycle, AI supply chain
Security engineering leadSOC methodology, detections, investigations, evidence, response policy
Backend / platform engineersAPIs, workflows, data services, state, connectors, integration SDKs
Data engineerIngestion, schemas, normalization, lineage, event stores, data quality
Detection engineersRules, testing, telemetry requirements, mappings, tuning, health
SRE / platform engineerReliability, capacity, observability, DR, operations, incident management
Cloud security engineerIAM, network, KMS, runtime, workload identity, deployment guardrails
Front-end / product engineerAnalyst experience, evidence presentation, approvals, accessibility
QA / adversarial engineerRegression, red-team, negative, safety, cross-tenant, failure testing
Governance / compliance ownerPrivacy, control mappings, customer assurance, contracts, audit readiness

Separation of duties: release approval, response approval, risk acceptance, and audit review are held by different named owners; no self-approval anywhere in the chain.

Operational runbooks

Condensed operating procedures — recommended baselines to be finalized during Phase 0–1. Every runbook records evidence, requires a named role, and ends with verification.

Lifecycle

Tenant onboarding

Trigger: signed agreement, tier selected. Role: platform operations + tenant admin. Approval: tenant management owner.

  • Create tenant in control plane → assign region/cell per tier and residency → generate tenant DEKs in KMS.
  • Provision queue partitions, storage prefixes, RLS policies, quotas; issue tenant-scoped workload claims.
  • Configure SSO federation, SCIM, roles, and (MSSP) delegation mappings; set model and response policy defaults (restrictive).
  • Onboard first connectors (read-only); run ingestion conformance and cross-tenant negative tests.

Verify: isolation canaries pass; test alert flows end-to-end; audit events present. Rollback: disable tenant, revoke claims. Evidence: onboarding report in ledger.

Tenant offboarding & verified deletion

Trigger: contract end or termination. Role: platform operations. Approval: tenant management owner + governance owner.

  • Legal-hold check → export package if contracted → revoke all tenant access, delegations, and connector credentials.
  • Destroy tenant DEKs (cryptographic erasure) → purge replicas, search indexes, caches → schedule backup expiry.

Verify: sampled reads fail across all stores; credential audit clean. Evidence: signed offboarding report with dates and scope. Escalate: any residual-data finding = platform incident.

Connector onboarding

Trigger: tenant requests integration. Role: tenant admin + platform engineer. Approval: tenant policy; action verbs additionally security lead.

  • Register connector version from registry; confirm declared capabilities and side-effect classes.
  • Store credentials in secret manager (scoped, least-privilege, read-only first); bind egress allowlist.
  • Run conformance suite: schema, limits, failure injection, cross-tenant negatives; enable health monitoring.

Verify: events flowing, dedup working, provenance recorded. Rollback: disable connector; credentials revoked.

Connector credential rotation

Trigger: schedule, staff change, or suspicion. Role: platform operations. Approval: automatic if scheduled; security lead if suspicion.

  • Issue new credential in source system → store new version in secret manager → connectors resolve at next call.
  • Confirm traffic on new credential → revoke old → verify old credential fails.

Verify: health green; zero auth errors after cutover. Evidence: rotation entry in secret-access audit.

Model onboarding / model & prompt rollback

Trigger: new model/version or quality-safety regression. Role: AI platform lead. Approval: release gate (all ten items) for onboarding; any on-call for rollback.

  • Onboard: register model (provider, version, data eligibility, context limits, hashes for local artifacts) → benchmark + safety suites → shadow comparison → canary → allowlist per tenant policy → update AI/ML-BOM.
  • Rollback: pin previous model/prompt version at the gateway/registry (configuration change, immediate) → re-run smoke evaluation → note affected cases in ledger.

Verify: routing shows expected version; no eligibility violations. Escalate: suspected provider compromise → emergency model disable + platform incident.

Agent release / agent disablement

Trigger: new agent version, or safety/quality threshold breach. Role: AI platform lead; disablement by any on-call.

  • Release: contract updated + evaluation suite pass + permission diff separately approved → shadow → canary → general.
  • Disable: flip agent kill switch (tenant- or global-scope) → workflows fall back to deterministic path or queue → notify owners.

Verify: no new spans from disabled version; queued work draining. Evidence: registry state change in audit log.

Detection release / rollback

Trigger: pipeline promotion or health regression. Role: detection engineer; peer approval mandatory.

  • Release: full pipeline (tests, replay, FP, performance, ATT&CK mapping) → shadow → canary → general.
  • Rollback: revert to prior version tag; suppression rules for in-flight noise; document cause in repo.

Verify: fire-rate and FP within band. Evidence: pipeline run + approval in repo history.

Incidents

Prompt-injection incident

Trigger: injection metric alert, analyst report, or anomalous agent behavior. Role: platform security on-call.

  • Contain: disable affected workflow/agent (kill switch); quarantine offending evidence objects (flag, never delete).
  • Investigate from the decision ledger: what context reached which model; which tool calls followed; any policy denials bypassed (should be none).
  • Add the payload to the adversarial suite; update sanitizer/classifier; re-run affected cases via replay; notify affected tenants per contract.

Verify: replayed cases produce clean traces; suite catches payload. Evidence: incident record with evidence hashes.

Compromised connector

Trigger: anomalous connector behavior, vendor advisory, credential leak signal. Role: platform security on-call.

  • Revoke connector credentials immediately; disable connector (tenant- or fleet-scope as indicated).
  • Sweep audit for actions performed via the connector; classify data touched; check egress allowlist violations.
  • Rotate adjacent secrets; re-onboard via conformance suite after vendor remediation; tenant notification per contract.

Verify: old credentials dead; clean re-certification. Evidence: timeline + scope report.

Unauthorized action attempt

Trigger: broker or actuator denies an action outside policy, or detects an attempt to bypass approval. Role: platform security on-call. Severity: always an incident, even when blocked.

  • Preserve the full trace (agent, workflow, model, prompt, tool versions); freeze the requesting agent version.
  • Determine cause: injection, contract error, policy gap, or compromise; fix at the strongest layer.
  • Add to adversarial suite; review the zero-tolerance metric with owners.

Verify: replay shows denial holds. Evidence: audit chain + corrective action record.

Audit pipeline failure

Trigger: audit-write errors or integrity-check failure. Role: SRE on-call + security lead (dual).

  • Automatic: R3+ execution blocks platform-wide while audit integrity is not guaranteed (fail closed).
  • Buffer audit events durably at source; restore audit service; drain buffers in order; run integrity verification.

Verify: hash chain continuous; no gaps in sequence. Escalate: evidence of tampering → platform security incident + forensic hold.

Resilience

Regional cell failover

Trigger: cell health below threshold or regional outage. Role: SRE on-call. Approval: incident commander.

  • Freeze non-essential writes; confirm queue durability; route new ingestion to standby per residency policy (sovereign tenants never leave region — they degrade instead).
  • Promote replicas within residency limits; replay from checkpoints; reconcile dedup.

Verify: RTO/RPO met; no cross-region residency violation; evidence gaps flagged on affected cases. Evidence: failover report.

Backup restoration

Trigger: data-loss event or quarterly restore test. Role: SRE. Approval: data owner; governance owner if tenant data involved.

  • Restore into isolated environment first; verify integrity hashes and tenant-key decryption; validate row counts and lineage continuity; then promote.

Verify: sampled evidence hashes match; RLS intact. Evidence: restore-test record (quarterly minimum).

Emergency kill switch

Trigger: unsafe behavior, suspected compromise, or regulator/customer demand. Role: any on-call may activate; only incident commander deactivates.

  • Choose narrowest effective scope: action type → model → tool → agent → workflow → tenant → cell → global.
  • Activation is deterministic and model-independent; investigations queue; raw alert receipt continues; automated response halts.

Verify: disabled paths produce zero executions; queues draining post-restore. Evidence: activation/deactivation entries with actor identity in audit log.

Architecture decision records

The material decisions, each with the rejected alternative and a revisit trigger. Status Proposed baseline v1.0 — to be ratified in Phase 0. Items no default can settle are in Open Decisions.

ADR-001 · Deterministic shell, probabilistic core

Status: Proposed baseline · shapes: every view

Decision: AI classifies, correlates, plans, and drafts. Identity, tenancy, policy, workflow state, approval, execution, verification, and audit are deterministic services.

Alternative rejected: agent-framework-centric design where orchestration is the architecture — untestable authority, unreplayable state, injection-to-action paths.

Revisit when: never for authority placement; revisit component boundaries as agent capabilities mature.

ADR-002 · Durable workflow engine owns investigation state

Status: Proposed baseline · shapes: Overview, Layers L11

Decision: a durable workflow engine (Temporal-class) with checkpoints, retries, compensation, human waits, versioning, and replay.

Alternative rejected: in-memory agent loops — lost or duplicated work, no replay, no long-running approvals.

Revisit when: engine operational cost dominates, or workflow volume justifies a managed equivalent.

ADR-003 · OCSF internal schema, raw always preserved

Status: Proposed baseline · shapes: Data & Evidence

Decision: normalize to OCSF with mapping versions and vendor-field preservation; the raw event remains the evidentiary source of truth.

Alternative rejected: proprietary schema (portability loss) or normalize-and-discard (destroys evidence).

Revisit when: OCSF coverage gaps force excessive vendor extensions for priority sources.

ADR-004 · Cell-based tenancy with five tiers

Status: Proposed baseline · shapes: Multi-Tenancy

Decision: regional cells bound tenant count and blast radius; tiers from Standard to Customer-managed share one control plane.

Alternative rejected: single pooled data plane with row filters only — one bug from cross-tenant exposure, no sovereignty path.

Revisit when: cell count makes fleet operations the dominant cost — invest in cell automation, not pooling.

ADR-005 · PostgreSQL as system of record; stores by role

Status: Proposed baseline · shapes: Data & Evidence

Decision: PostgreSQL for transactional truth; WORM object storage for raw evidence; search store as rebuildable projection; Redis ephemeral only.

Alternative rejected: search engine or vector DB as system of record — weak transactional guarantees, mutable evidence.

Revisit when: case-store scale exceeds single-cluster comfort per cell — shard by cell first.

ADR-006 · Central model gateway, no direct provider SDKs

Status: Proposed baseline · shapes: Overview, Security

Decision: one gateway enforces allowlists, residency, classification, pinning, budgets, fallback, redaction, accounting, and emergency disable.

Alternative rejected: per-service provider SDKs — policy drift, uncontrolled data transfer, silent version changes.

Revisit when: a workload needs provider features the gateway cannot proxy — extend the gateway, never bypass it.

ADR-007 · Mixed model estate under per-workload policy

Status: Proposed baseline · shapes: Agents, Integrations

Decision: hosted, private-endpoint, and open-weight models selected per workload sensitivity, quality, latency, and cost; open-weight is not assumed safer or cheaper — it carries GPU capacity, patching, integrity, and quality burdens.

Alternative rejected: single-provider simplicity (residency and outage exposure) or local-only purity (quality and operations cost).

Revisit when: benchmarks show a tier consistently dominated on quality-per-cost for its workload.

ADR-008 · Vector index optional and never authoritative

Status: Proposed baseline · shapes: Data & Evidence

Decision: approximate retrieval only for approved use cases; every hit re-validated against the source of record before influencing a finding.

Alternative rejected: RAG-first evidence architecture — poisoning surface, no chain of custody.

Revisit when: a use case demonstrates recall value that outweighs governance cost.

ADR-009 · Policy-as-code with central PDP, PEPs everywhere

Status: Proposed baseline · shapes: Security

Decision: versioned policy (OPA/Cedar-class) evaluated at gateway, broker, workflow, and response plane; RBAC + ABAC with tenant-bound claims.

Alternative rejected: authorization scattered in application code — unauditable, untestable, divergent.

Revisit when: policy latency requires local evaluation with centrally distributed bundles (expected; plan for it).

ADR-010 · Typed tool broker + canonical query compiler

Status: Proposed baseline · shapes: Integrations

Decision: agents call typed tools only; model-generated queries exist as canonical intent compiled and AST-validated per vendor.

Alternative rejected: raw LLM query strings to production SIEMs — injection, cost, and correctness risks.

Revisit when: a vendor exposes a safe, parameterized query API that lets the compiler thin out.

ADR-011 · Action-risk model R0–R5 with dual control at R4

Status: Proposed baseline · shapes: Response Controls

Decision: six classes; R3+ needs explicit approval, R4 two-person, R5 prohibited in-platform; audit-write failure blocks R3+.

Alternative rejected: per-action ad-hoc judgment — inconsistent, unanalyzable, untrainable.

Revisit when: operating evidence supports moving a specific, verified, reversible action down a class — one action at a time.

ADR-012 · Tamper-evident audit separate from telemetry

Status: Proposed baseline · shapes: Observability

Decision: three record categories; the audit log is append-only, hash-chained, independently retained; the decision ledger is replay-grade.

Alternative rejected: one log stream for everything — sampling and retention needs conflict with forensic integrity.

Revisit when: volume forces tiering — tier telemetry, never audit.

ADR-013 · Durable event bus for ingestion decoupling

Status: Proposed baseline · shapes: Overview

Decision: durable, tenant-partitioned messaging (Kafka-class) with DLQs and backpressure between connectors and processing.

Alternative rejected: synchronous ingestion — source spikes become platform outages.

Revisit when: operational burden justifies a managed bus within residency constraints.

ADR-014 · Minimal agent set behind benchmark gates

Status: Proposed baseline · shapes: Agents

Decision: twelve specialized agents maximum at target state; each must beat the deterministic baseline in evaluation before release; deterministic-first for everything else (P09).

Alternative rejected: open-ended agent swarm — cost, latency, attack surface, and unattributable behavior.

Revisit when: a new workflow shows measured value for an additional bounded agent.

Risk register

Inherent risk assumes no controls; residual assumes the architecture as designed and operated. Likelihood/impact are this review's assessment for a typical MSSP-scale deployment. Owners are functions, not names.

Inherent
IDRiskThreat actorLIInherentKey controls (preventive · detective · corrective)ResidualOwner / treatment
R-01Cross-tenant data exposureCross-tenant attacker; defectMedSevereCriticalCells, claims, RLS, tenant keys · continuous canaries, release negative tests · incident + notification runbookLow–MedProduct security · Mitigate
R-02Prompt injection alters investigation or actionMalicious evidence sourceHighSevereCriticalContext builder, typed tools, least agency · injection metrics, output validation · quarantine + replay runbookMediumAI platform · Mitigate
R-03Unauthorized or wrong-target response actionCompromised actuator; defectLow–MedSevereCriticalR0–R5, approvals, target re-validation, tokens · post-action verification · rollback + kill switchLowSecurity eng · Mitigate
R-04Excessive agency / tool abuseInjected or defective agentMedHighCriticalContracts, broker allowlists, budgets · unauthorized-attempt metric (zero-tolerance) · agent disableLowAI platform · Mitigate
R-05Audit gaps or tamperingInsider; attacker covering tracksLowSevereCriticalAppend-only hash-chained store · integrity verification · R3+ execution blocks on failureLowProduct security · Mitigate
R-06Secrets exposure via prompts, logs, or tracesAny; often accidentalMedSevereCriticalSecret references, call-time resolution · scanning in telemetry pipeline · rotation + revocationLowCloud security · Mitigate
R-07Unsupported findings accepted as factModel hallucinationHighHighHighEvidence-linked claims, UX labeling · grounded-claim metrics, quality agent · human review, reopen flowMediumSecurity eng · Mitigate
R-08Model provider outage or silent changeProviderHighMedHighPinning, fallback chain, deterministic degraded mode · shadow comparison, health · failover <5mLowAI platform · Mitigate
R-09Connector outage or rate limitingVendor; misconfigHighMedHighQueues, backpressure, circuit breakers · health model · queued investigations, incomplete-evidence stateLowSRE · Mitigate
R-10Data / retrieval poisoningAdversary seeding contentMedHighHighNon-authoritative vector index, TI decay/deconfliction · lineage, quality scoring · re-validation vs sourceMediumAI platform · Mitigate
R-11Compromised open-weight model artifactSupply-chain attackerLow–MedHighHighArtifact hashes, AI/ML-BOM, admission control · benchmark drift detection · model disable + rollbackLowAI platform · Mitigate
R-12Data-residency violationRouting or policy defectMedHighHighRegional cells, gateway residency enforcement · routing audit · sovereign tenants degrade rather than moveLowGovernance · Mitigate
R-13Cost / token exhaustion; agent loopsAdversary; defectHighMedHighBudgets at gateway + broker, step caps · cost telemetry, loop detection · circuit breakers, quotasLowSRE · Mitigate
R-14Detection translation errors (semantic drift)Tooling; complexityHighMedHighReplay + FP testing per destination · detection health · rollbackMediumDetection eng · Mitigate
R-15Critical false negativesAdversary evading triageMedSevereHighDeterministic floor rules, escalation bias in policy · FN metrics from reopens/hunts · threshold retuningMediumSecurity eng · Mitigate
R-16Analyst overreliance on AI outputOrganizationalHighMedHighEvidence-labeled UX, unsupported-claim marking · override/agreement metrics · training, workflow designMediumProduct · Mitigate
R-17Privileged insider misuseInsider; support engineerLowSevereHighSoD, customer-approved support access, JIT · session recording, secret-access audit · revocation, incidentLow–MedProduct security · Mitigate
R-18CI/CD or artifact supply-chain compromiseSupply-chain attackerLow–MedSevereHighSigned builds, provenance, admission · pipeline integrity monitoring · revocation + rollbackLowCloud security · Mitigate
R-19Recovery failure (backup/restore)OperationalLowHighMediumQuarterly restore tests · integrity hashes · isolated-restore runbookLowSRE · Mitigate
R-20Stale or conflicting threat intelligenceFeed qualityHighLowMediumExpiry, decay, deconfliction · source-quality scoring · suppressionLowSecurity eng · Mitigate
R-21Vendor / model lock-inCommercialMedMedMediumGateway abstraction, OCSF, canonical query intent · portability review per ADR · alternative certificationLow–MedProduct · Accept partially
R-22Fine-tuning contamination or driftProcessLowMedMediumDeferred to Phase 6; dataset lineage, consent, isolation · eval regression · rollbackLowAI platform · Defer
R-23Cell fleet operational sprawlOperationalMedMedMediumCell automation, IaC, fleet observability · capacity reviews · consolidation within tiersLow–MedSRE · Mitigate
R-24Evaluation-set leakage into training/promptsProcessMedMedMediumSet separation, holdout governance · leakage checks · re-baseliningLowAI platform · Mitigate

Open decisions — required before build

Fifteen questions that must be answered and recorded in Phase 0. Each carries a recommended default from this review and the consequence of leaving it unresolved.

1 · What is the product, primarily?

Recommended default: an analyst copilot with gated response — not an autonomous SOC worker, not a SOAR replacement, not a detection platform first. If unresolved: every later scope debate reopens; evaluation targets are unanchorable.

2 · Which actions are categorically prohibited; which need single vs dual approval?

Recommended default: the R0–R5 model — R3 single approval, R4 dual, R5 prohibited in-platform. If unresolved: response scope creeps action by action with no principled line.

3 · What is sufficient evidence for each disposition and response?

Recommended default: per-workflow sufficiency rules defined with detection engineering; contradictions and gaps always block policy-controlled closure. If unresolved: “confidence” silently substitutes for evidence.

4 · What is the authoritative internal event and evidence schema?

Recommended default: OCSF plus the evidence-record envelope (ADR-003), versioned mappings. If unresolved: every connector invents its own semantics; replay is impossible.

5 · What data may leave a tenant region, cloud, or customer environment?

Recommended default: nothing raw leaves the cell by default; model calls follow tenant model policy with classification-based redaction; sovereign tenants degrade rather than transfer. If unresolved: residency violations arrive with the first misconfigured tenant.

6 · How is tenant isolation enforced beyond application filtering?

Recommended default: the Multi-Tenancy defense-in-depth stack: cells, claims, RLS, keys, partitioned queues, per-tenant credentials, negative tests. If unresolved: isolation is one WHERE clause away from an incident.

7 · How are human, service, agent, workflow, connector, and actuator identities issued and authorized?

Recommended default: workload identity federation with short-lived, tenant-bound credentials; humans via OIDC/SAML + phishing-resistant MFA. If unresolved: attribution — the platform's core promise — fails.

8 · Which service owns durable workflow state and replay?

Recommended default: the workflow engine (ADR-002); agents are stateless workers. If unresolved: state fragments across frameworks and queues; nothing is reconstructable.

9 · How are models, prompts, tools, policies, schemas, playbooks, detections versioned and rolled back?

Recommended default: everything through the signed control-plane registries with release gates and one-step rollback. If unresolved: silent drift — a provider update becomes a production behavior change nobody approved.

10 · How is quality calibrated per workflow and security outcome?

Recommended default: the five-level evaluation hierarchy with per-workflow calibrated risk scores and human-agreement thresholds. If unresolved: stylistic quality masquerades as security quality.

11 · How are malicious evidence and tool responses contained?

Recommended default: the context-builder architecture plus typed tools and output validation; adversarial suite grows with every incident. If unresolved: the SOC's own telemetry is the attack vector.

12 · How is every action simulated, approved, verified, and reversed?

Recommended default: the response control sequence — dry-run where supported, verification query mandatory, rollback method required before enablement. If unresolved: the first wrong-target isolation is a customer-trust event.

13 · What does the system do when no LLM is available or permitted?

Recommended default: deterministic triage floor, queued investigations, disabled automation, preserved raw evidence, explicit incomplete-evidence status. If unresolved: a provider outage becomes a security-operations outage.

14 · How is the platform itself investigated and recovered when it becomes the incident?

Recommended default: platform IR runbooks, forensic logging, credential revocation, tenant notification, recovery exercises — treated with the same rigor the platform sells. If unresolved: the incident-response product has no incident response.

15 · What responsibilities remain with customer, MSSP, model provider, integration vendor, and operator?

Recommended default: a published shared-responsibility matrix per deployment tier, contract-referenced. If unresolved: every incident becomes a liability negotiation.