# AI Security Workbench Figures
PROD-01
Connected Workbench Analysis Flow
Modeling, scanning, testing, path analysis, hardening, and evidence share stable identities, relationships, claim state, and lifecycle context rather than behaving as disconnected tools.
Ecosystem map with the AI Security Workbench at the center and Model, Analyze, Test, Qualify, Defend, and Evidence capabilities around it.
Core
AI Security Workbench
Shared contracts
- Shared evidence and lifecycle state
- Shared contracts and identifiers
- Reusable outputs and regression proof
Model
Threat Canvas • Authority Graph • Surface discovery
Analyze
Code Scanner • Tool and artifact analysis • Configuration analysis
Test
Adversarial Range • RAG and injection test packs • Runtime Trace
Qualify
Attack Path Analysis • Evidence grounding • Sequence and precondition review
Defend
Authority hardening • Control chokepoints • Retest and regression
Evidence
Provenance and claim state • Structured outputs • Partner and buyer evidence
PROD-02
Connected Finding and Evidence Contract
Shared evidence and lifecycle states allow scanner, validation, authority, path analysis, and evidence capabilities to exchange results without losing provenance.
Round-trip integration figure showing scanner, validation, authority, path analysis, remediation, retest, and evidence sharing one evidence contract.
Normalize • Enrich • Return • Retest
- 1Capability outputScanner finding • Adversarial observation • Artifact context
- 2Connected evidence contractStable identifiers • Source and provenance • Claim and lifecycle state • Entities and relationships
- 3Downstream analysisAuthority enrichment • Attack Path Analysis • Control and retest planning
- 4Lifecycle returnRemediation state • Retest result • Partner projection • Reviewed report or evidence pack
Return to stage 1 — lifecycle state preserved