Supply intelligence
ChainSentry decision workspace
Maps bounded upstream dependencies, corroborates supported public evidence, and keeps likelihood, business impact, confidence, and inventory recommendations independently inspectable.
Resilience decision platform ← Software
MainStreet Sentinel
A guided decision-support platform that helps smaller organizations strengthen supply resilience, build a practical cybersecurity baseline, and adopt AI with clear boundaries—without turning critical decisions over to a black box.
Human-reviewed decision flow
From operating facts to an auditable action
01
Owner-provided contextInventory, dependencies, controls, and declared use cases
facts 02
Evidence-aware analysisSource provenance and degradation remain visible
traceable 03
Deterministic prioritiesFormula-driven risk, coverage, and readiness rules
reproducible 04
Owner approvalMaterial changes require review and re-approval
human owned 3Readiness workstreams
14NIST-aligned starter controls
6Versioned AI micro-lessons
1Human-reviewed action program
Product capabilities
One operational view across resilience, security, and responsible AI
The product turns broad readiness questions into bounded workflows with visible assumptions, evidence, limitations, and accountable next actions.
Resilience planning
Deterministic disruption math
Turns owner-provided inventory, demand, lead-time, and recovery facts into reproducible coverage gaps and action priorities with visible formula traces.
Cyber readiness
Practical safeguards
Guides organizations through 14 starter controls aligned to all six NIST CSF 2.0 functions, with explicit implementation states and evidence-aware priorities.
AI governance
Bounded AI readiness
Combines versioned micro-lessons, deterministic assessments, use-case design, data classification, and a constrained educational coach grounded in approved sources.
Human authority
Reviewable action programs
Merges supply, cyber, and AI priorities into immediate, near-term, and continuous lanes. Material changes invalidate prior approval before a new PDF can be issued.
Trust architecture
Tenant-aware by design
Uses server-derived organization membership, role-aware commands, data minimization, safe audit events, idempotent operations, and explicit degraded states.
Operating model
Built to support a decision—not impersonate the decision-maker
01
Collect
Minimal business context and explicitly classified inputs.
02
Evaluate
Deterministic rules plus supported, provenance-aware evidence.
03
Prioritize
Immediate, near-term, and continuous actions with visible rationale.
04
Approve
Authorized human review, payload-bound approval, and export.
Engineering signals
What the implementation demonstrates
A full-stack product shaped around explainability, multi-tenant security, graceful degradation, and operational handoff.
Deterministic decisions, optional generation
Calculations, priorities, quiz keys, lifecycle transitions, approvals, and exports remain in versioned application code. Ollama is optional and has no authority to change those outcomes.
Evidence failure stays visible
Unavailable connectors do not manufacture fallback evidence. The product records degradation, excludes failed sources from scoring, and lowers confidence where appropriate.
Production-minded multi-tenancy
FastAPI authorization derives tenant context from validated membership, while object-level role checks protect review, approval, export, and privacy workflows.
Operational delivery, not a prototype shell
The workspace includes migrations, background export jobs, containerized services, automated validation, threat modeling, retention guidance, and deployment runbooks.
Selected stack
Modern web delivery with a controlled AI boundary
Next.jsReactTypeScriptFastAPIPydanticPostgreSQLSQLAlchemyAlembicRedis / RQS3-compatible storageOllama (optional)Docker Compose
Interested in resilient, human-centered AI systems?
I’m open to conversations about trustworthy product engineering, evidence-aware AI, and operational decision platforms.