Publication DSL
AI Security Figure Library
115 canonical figures authored in structured Markdown and compiled for responsive pages, publications, slides, and social distribution.
academy
Keep the LMS and marketplace. Add the AI security content layer.
Learning platforms, course marketplaces, and enterprise training programs keep their delivery experience while AI Security LLC supplies structured, standards-ready AI security content.
capability plane
From source material to published learning product.
A reusable learning offer needs editorial review, instructional packaging, standards validation, platform testing, publication controls, and versioned updates.
governed lifecycle
One course. One platform. One distribution pilot.
A bounded pilot packages one representative course, validates it in the target LMS or marketplace workflow, and returns the operational assets needed for launch.
timeline horizon
Course Packaging Engine
Canonical curriculum can compile into partner-ready modules, assessments, lab assets, and reporting packages without duplicating source content.
transformation engine
attack path analysis
Grounded core and explicit inference.
Attack Path Analysis separates what the evidence directly supports from what remains a speculative extension requiring validation or analyst review.
grounding boundary
Interrupt the paths.
Multiple attack paths may converge on a small number of authority, approval, or tool-access weaknesses that offer better remediation leverage than fixing symptoms one at a time.
chokepoint
From connected evidence to qualified paths.
Attack Path Analysis combines fragmented security observations with system, identity, and authority context to construct evidence-qualified paths and structured remediation outputs.
transformation engine
brand licensing
Choose the brand and product-control model.
The licensing spectrum ranges from AI Security LLC-branded delivery through co-branding and private label to a deeply partner-branded white-label experience.
branding spectrum
What changes and what remains governed.
Brand, domain, packaging, and customer experience may shift to the partner while evidence semantics, security boundaries, release integrity, and entitlement remain governed.
operating model
From branded pilot to partner-operated offer.
Branding is only one part of launch; the operating path also needs tenant provisioning, offer configuration, acceptance, support, upgrades, and commercial governance.
governed lifecycle
Licensing and Control Spectrum
Branded, co-branded, white-label, private-label, and embedded models vary in presentation control, product ownership, and operating responsibility.
ordered spectrum
consulting
Consulting partner delivery model.
The consulting partner owns advisory context, stakeholder management, and implementation while AI Security LLC supplies repeatable technical workflows, evidence, and productized delivery assets.
operating model
From assessment through remediation and retest.
A consulting engagement should connect discovery, evidence, findings, recommendations, implementation, retest, and residual risk rather than ending at report delivery.
governed lifecycle
One client engagement pilot.
A bounded engagement proves joint delivery, evidence quality, role boundaries, remediation usefulness, and the ability to repeat the work under the consulting partner relationship.
pilot contract
Consulting-to-Product Handoff
A consulting engagement should leave reusable tests, evidence, controls, and productized capability rather than only a static report.
transformation engine
partners hub
The AI Security LLC partner ecosystem.
Different partner types connect through distinct commercial motions while sharing a common capability, evidence, interoperability, and support foundation.
ecosystem map
Choose the partner motion.
The right program depends on who owns the product and who owns the customer and delivery.
quadrant matrix
From first conversation to a supported partner motion.
Every partner path begins with fit and a representative object, then advances through proof, operating-model design, launch, and support.
process rail
Partner Motion Selector
What does the partner want to own?
decision tree
Commercial Path Overview
A partner engagement moves from motion selection and representative proof through commercial fit, integration, launch, and expansion.
process rail
interoperability
Canonical contracts between heterogeneous systems.
Canonical objects isolate security and learning semantics from any one vendor transport, endpoint, or proprietary data model.
transformation engine
Protocol adapter versus vendor binding.
A portable adapter handles transport and normalization; a vendor binding adds endpoint, authentication, lifecycle, and sandbox behavior that must be independently validated.
comparison
One object, round trip, lifecycle preserved.
Interoperability is complete only when identifiers, provenance, state transitions, and retest behavior survive the return to the partner workflow.
integration round trip
Interoperability Maturity Ladder
Integration maturity progresses from file exchange to structured contracts, native round trips, and embedded workflow state.
ordered spectrum
marketplace
From partner capability to transactable listing.
A marketplace offer moves through packaging, validation, listing, discovery, commercial action, provisioning, and delivery.
process rail
Marketplace operating model.
The vendor, marketplace, AI Security LLC, and customer each own distinct responsibilities across listing, transaction, entitlement, delivery, support, and renewal.
operating model
One listing and fulfillment pilot.
A bounded pilot proves the listing metadata, commercial action, provisioning, delivery, usage signal, and renewal path for one offer.
pilot contract
Buyer Offer Map
The marketplace organizes products, services, licensing, and partner offers by the buyer problem they solve.
ecosystem map
Offer-to-Pilot Journey
A marketplace listing becomes useful when it leads cleanly from problem selection through scope, pilot, evidence, and production decision.
governed lifecycle
mssp
Keep the client relationship. Add the AI product-security delivery layer.
The MSSP owns the service, account, and operating model while AI Security LLC adds repeatable technical workflows, evidence discipline, and productized delivery assets.
capability plane
From customer onboarding to recurring assurance.
A managed service becomes scalable when scope, evidence, decisions, remediation, retest, and renewal are handled as a repeatable lifecycle.
governed lifecycle
One customer. One bounded service pilot.
A pilot should prove that AI Security LLC workflows can fit the MSSP operating model, produce customer-ready evidence, and support a repeatable recurring service.
pilot contract
Managed Delivery Lifecycle
The MSSP owns customer delivery while AI Security LLC supplies reusable capability, evidence contracts, and escalation-ready technical outputs.
governed lifecycle
oem
Choose the capability layer.
Partners can embed one bounded AI Security LLC capability or combine a controlled module set without adopting the entire AI Security LLC interface.
capability plane
From headless engine to partner-native product.
A bounded input invokes a headless AI Security LLC capability and returns a structured object that fits the partner data model, lifecycle, and interface.
transformation engine
From evaluation package to production OEM.
The path to production proves integration, evidence quality, deployment fit, entitlement, support, and a credible commercial operating model.
pilot contract
OEM Deployment Spectrum
OEM delivery can progress from hosted API access through local workers and embedded sidecars to controlled private deployment.
ordered spectrum
OEM Pilot to Production
A representative-object pilot should prove contract fit, output value, return-path fidelity, and operational ownership before production expansion.
governed lifecycle
offensive platforms
Keep the attack engine. Add the context and evidence plane.
The offensive platform retains attack generation and orchestration while AI Security LLC adds system context, authority reasoning, evidence qualification, and structured remediation outputs.
capability plane
From attack result to qualified path.
A successful attack or trace is not automatically a defensible multi-step path; context, grounding, validation, and review must remain explicit.
governed lifecycle
One representative attack. One measurable platform pilot.
A bounded pilot compares the current result with an evidence-qualified, context-enriched path and returns the useful fields to the partner platform.
pilot contract
Attack Engine and Evidence Plane
The partner keeps attack generation and operator workflow while the OEM Engine adds context, qualification, evidence, and defensible return objects.
comparison
Observed Attack Round Trip
A useful integration accepts a representative attack object, enriches and validates it, and returns a partner-native result.
integration round trip
resellers
Who sells, delivers, and supports.
The reseller owns market access and customer coordination while AI Security LLC supplies technical proof, provisioning, product delivery, and escalation support.
operating model
From lead registration through renewal.
A repeatable channel motion needs lead ownership, qualification, technical proof, commercial coordination, delivery handoff, support, and renewal rules.
governed lifecycle
One customer opportunity and handoff pilot.
A bounded opportunity proves qualification, technical support, proposal coordination, provisioning, delivery handoff, and shared account ownership.
pilot contract
Reseller Delivery Boundary
The reseller owns opportunity and relationship while delivery, support, and technical accountability remain explicitly assigned.
operating model
scanner providers
Keep the scanner. Add the finding and evidence plane.
The scanner remains the product surface while AI Security LLC adds AI-native finding depth, evidence discipline, attack-path context, and lifecycle support.
layered cutaway
From candidate to defensible finding.
Detection is only the beginning; evidence, validation, review, and claim status determine whether a result is reportable.
governed lifecycle
One representative finding. One measurable OEM pilot.
A bounded integration test proves the create, update, evidence, deduplication, rescan, remediation, and retest lifecycle.
pilot contract
Partner Findings Lifecycle
AI-native findings must return into the partner product with lifecycle state preserved through review, remediation, and rescan.
integration round trip
Generic Bridge and Native Adapter
A generic exchange proves interoperability; a native adapter preserves richer partner concepts, workflow state, and user experience.
comparison
technology partners
Where technology partners connect.
Technology partners contribute complementary data, controls, orchestration, evidence, and downstream workflows around a shared customer outcome.
ecosystem map
Discover, map, validate, and publish an integration.
A technology integration should progress from capability discovery through schema mapping and fixture validation to a supported published status.
governed lifecycle
One joint solution pilot.
A bounded pilot combines one partner capability with one AI Security LLC capability and proves a measurable customer outcome plus a product-native return.
pilot contract
verno
Keep the attack engine. Add the context and evidence plane.
The offensive platform continues generating and chaining attacks while AI Security LLC adds system context, evidence qualification, remediation chokepoints, and structured return.
capability plane
One representative attack, enriched.
A single sanitized result is enough to compare the current product object with the evidence, authority, and remediation context AI Security LLC can add.
comparison
Pilot input and return contract.
The pilot is complete only when a partner-native result returns to the intended product workflow.
pilot contract
workforce readiness
Keep the learning platform. Add the readiness and evidence layer.
The partner continues delivering learning experiences while AI Security LLC adds role models, assessments, evidence-backed scoring, and workforce reporting.
capability plane
From completion to defensible readiness claim.
Course completion, lab activity, and self-reporting become readiness evidence only after mapping, scoring, validation, and claim controls.
governed lifecycle
One role. One cohort. One measurable readiness pilot.
A bounded pilot maps one role and learner population into the partner platform, then returns evidence-backed readiness results and next actions.
pilot contract
Role-to-Cohort Scoring Flow
Role definitions, assessment evidence, calibrated scoring, and cohort reporting form one workforce-readiness pipeline.
transformation engine
map
What Map Establishes
Map turns architecture, authority, data movement, and trust boundaries into a security model that can be tested.
capability plane
Trust and Authority Topology
Security-relevant behavior emerges from relationships among identities, agents, tools, data, and approval gates.
topology map
Evidence Readiness Boundary
A useful map separates observed relationships from inferred connections and explicitly exposes evidence gaps.
grounding boundary
Map-to-Attack Handoff
Mapped assets, authority, and trust boundaries become bounded attack hypotheses and a prioritized test plan.
transformation engine
Inventory and Trace Workflow
Mapping progresses from intake through architecture capture and trace review to a reviewable security model.
process rail
attack
What Attack Tests
Attack turns mapped exposure into reproducible tests across prompts, retrieval, agents, tools, authority, and workflows.
capability plane
From Isolated Finding to Defensible Path
A defensible attack path connects observed steps, evidence, and consequences rather than merely grouping nearby findings.
comparison
Adversarial Testing Workflow
Attack testing moves from scoped hypotheses through controlled execution and evidence capture to qualified findings.
process rail
Black-Box to Gray-Box Uplift
System context turns observed behavior into better-qualified risk without replacing the partner or operator attack engine.
comparison
Attack-to-Defend Handoff
Qualified paths become prioritized control opportunities, remediation hypotheses, and retest plans.
transformation engine
defend
What Defend Changes
Defend reduces dangerous authority, strengthens control boundaries, and converts remediation into retestable system changes.
capability plane
Interrupt the Paths
The highest-value control is often the chokepoint that disrupts several plausible attack paths at once.
chokepoint
Remediation Operating Model
Security, product, platform, and engineering teams need explicit ownership for control design, implementation, and proof.
operating model
Fix, Retest, Prove
A remediation is not complete until the control change is retested and the resulting evidence closes or updates the finding.
governed lifecycle
evidence
What Evidence Makes Defensible
Evidence preserves provenance, validation state, and retest history so technical findings can support real decisions.
capability plane
Evidence Lifecycle
Evidence moves from observation through validation, remediation, retest, and controlled reuse without losing claim state.
governed lifecycle
Claim State Boundary
Observed, reproduced, grounded, inferred, rejected, and analyst-reviewed claims must remain visibly distinct.
grounding boundary
Evidence-to-Decision Stack
Executive conclusions remain trustworthy only when they stay traceable to technical evidence and replayable proof.
artifact stack
scanner
AI-Native Scanner Coverage
The scanner evaluates AI-specific code and workflow surfaces that conventional AppSec categories do not fully describe.
coverage matrix
Scanner Workflow
The scanner moves from target discovery through analysis, evidence grouping, remediation, and rescan.
process rail
Scanner Output Contract
One scan can produce machine-readable findings, reviewable evidence, path inputs, and remediation-ready artifacts.
artifact stack
AI Scanner and Conventional AppSec
AI-native analysis extends conventional code scanning by following prompts, retrieval, tools, agents, and authority through application logic.
comparison
authority graph
Authority Graph Overview
The authority graph models which identities and agents can invoke which tools, data, approvals, and consequential actions.
imported graph
Unsafe Authority Composition
Individually reasonable permissions can combine into a dangerous end-to-end authority path.
imported graph
Authority to Control Chokepoint
A small change to permission, approval, or action scope can break several unsafe authority paths.
chokepoint
attack path analysis product
Attack Path Analysis Value Proposition
Attack Path Analysis qualifies and explains multi-step risk rather than merely correlating alerts or drawing speculative paths.
evidence dossier
Attack Path Analysis Validation Pipeline
Independent validators challenge evidence, ordering, consequence, and alternative explanations before a path is accepted.
governed lifecycle
Attack Path Analysis Result Contract
The path engine returns clusters, chokepoints, evidence links, residual state, and retest-ready outputs.
artifact stack
Analyst-Reviewed Path
A proposed path becomes publishable only after evidence challenge, correction, and explicit analyst review.
grounding boundary
adversarial range
Scenario Pack Architecture
Reusable scenario packs organize prompt, retrieval, tool, authority, multimodal, and workflow abuse into repeatable tests.
ecosystem map
Evaluation-to-Regression Lifecycle
An observed failure becomes durable security capability when it is converted into a replayable regression fixture.
governed lifecycle
Human-Guided and Autonomous Modes
Analyst-guided, model-guided, and replay-driven testing serve different purposes and require different controls.
quadrant matrix
rag harness
Retrieval Trust Topology
RAG security depends on the full path from query and corpus boundaries through ranking, prompt assembly, model behavior, and output.
imported graph
RAG Failure-Mode Coverage
A useful harness separates prompt injection, corpus poisoning, cross-tenant retrieval, provenance loss, and unsafe action propagation.
coverage matrix
RAG Retest and Evidence Flow
A retrieval failure is closed only after corpus, ranking, policy, or action controls are changed and replayed.
governed lifecycle
artifact analyzer
Artifact Analyzer Overview
The analyzer turns code, binaries, browser artifacts, and protocol surfaces into structured capability and authority context.
transformation engine
Artifact to Authority Context
Raw implementation artifacts become more useful when their exposed capabilities, identities, permissions, and sinks are made explicit.
comparison
Artifact Evidence Outputs
The analyzer returns structured findings, extracted metadata, authority context, and evidence references for downstream review.
artifact stack
workbench
Connected Workbench Analysis Flow
Map, scan, test, chain, defend, and prove operate as connected capabilities rather than isolated tools.
ecosystem map
Connected Finding and Evidence Contract
Shared evidence and lifecycle states allow scanner, range, graph, Attack Path Analysis, and reporting capabilities to exchange results without losing provenance.
integration round trip
launch review
Launch Review Scope
A launch review covers the architecture, data, models, retrieval, tools, authority, and evidence needed to make a release decision.
coverage matrix
Launch Review Engagement Lifecycle
The engagement moves from intake and mapping through validation, prioritized findings, remediation planning, and retest.
governed lifecycle
Launch Review Deliverable Stack
The buyer receives traceable technical findings, launch blockers, remediation priorities, and decision-ready evidence.
artifact stack
red team
AI Red Team Scope
AI red teaming should state which prompt, retrieval, tool, agent, authority, multimodal, and workflow surfaces are actually tested.
coverage matrix
Red Team Engagement Flow
A controlled engagement moves from scoping and scenario design through execution, reproduction, evidence, and reporting.
process rail
What Qualifies as a Finding
Theoretical exposure, observed anomaly, reproduced failure, evidence-qualified path, and analyst-reviewed finding are different claim states.
grounding boundary
attack path review
Attack-Path Review Uplift
A path-qualified review adds evidence, sequencing, shared weaknesses, and remediation leverage to an ordinary finding set.
comparison
Path Validation Workflow
Candidate paths are challenged for evidence, ordering, consequence, and alternatives before they become reportable.
governed lifecycle
Chokepoint Remediation Plan
The review prioritizes controls that interrupt several qualified paths while preserving residual-risk visibility.
chokepoint
enterprise
Enterprise Operating Model
AppSec, platform, product, data, governance, and engineering teams need explicit ownership across the AI security lifecycle.
operating model
Deployment Control Spectrum
Hosted, local-worker, sidecar, and air-gapped models trade operating simplicity for customer control and isolation.
ordered spectrum
Enterprise Evidence Package
Enterprise delivery should preserve machine-readable findings, retest records, control evidence, and executive reporting in one traceable stack.
artifact stack
