Maps With Teeth
Information Continuity & Architecture for High-Constraint Human Handoffs
Designed and implemented infrastructure for navigating fragmented institutional systems, agency handoffs, verified resources, and changing state without silently losing provenance or context.
What was difficult, fragmented, or broken?
When individuals navigate traumatic, high-stress institutional systems (legal, medical, shelter, or protective services), context collapses. Critical information is fragmented across disparate intake sheets, conflicting directories, and unverified handoff notes. Handoffs between support actors result in severe context degradation, forcing repetitive disclosures of sensitive history while risking evidentiary omissions and safety leaks.
Why the obvious solution wasn't enough
Standard "resource databases" or checklist apps assume stable mental bandwidth, linear user journeys, and static directory data. In high-stress scenarios, memory retrieval is non-linear, emotional fatigue causes abandonment, and directory data becomes dangerously outdated. Generic SaaS forms treat inputs as flat fields rather than stateful records with living provenance and strict privacy boundaries.
My Role
What I Built
- A 26-route Next.js production web architecture modeling survivor-centered routing, state persistence, and resource navigation under high operational constraints.
- A canonical data architecture (`src/content/canon/`) isolating verified resource registries, handoff state machines, and disclosure rules from presentation components.
- A 47-record resource verification index with cryptographic hash verification (`SHA256SUMS.txt`).
- Continuity Contact Record & Expiring Disclosure Envelope patterns to enable controlled disclosure without persistent surveillance exposure.
- Strict architectural governance documented in repository specifications (`SITE_ARCHITECTURE.md`, `CANON_AND_SOURCES.md`, `FILE_OWNERSHIP.md`).
- Automated headless browser verification tooling (`verify_field_journal_page.js`) auditing viewport regressions, visual clipping, and network response states.
Fragmented Intake Records (Disparate Agencies & Service Providers)
│
▼ (Canonical Resource Normalization)
State-Preserving Resource Ledger (5 Governed Domains)
│
▼ (Controlled Disclosure & Boundary Enforcement)
Continuity Contact Record & Expiring Disclosure Envelopes
│
▼ (Deterministic Verification & Human-Centered Navigation)
Audited 47-Record Verified Directory across 26 Production Routes// Verification & Governance Registry (src/content/canon/)
export interface VerifiedResourceRecord {
id: string;
sourceDomain: 'legal' | 'housing' | 'healthcare' | 'advocacy' | 'crisis';
verificationHash: string; // SHA256 checksum
lastVerifiedUtc: string;
governanceStatus: 'active' | 'under_review' | 'deprecated';
disclosureEnvelope: {
retentionDays: number;
accessGate: 'explicit_consent' | 'survivor_delegated';
};
}How I Approached It
I treated survivor navigation and institutional handoffs as an information topology and trust architecture problem. I architected the platform around five distinct domain boundaries: verified resource registries, handoff continuity, controlled disclosure boundaries, safety safeguards, and immutable provenance. Every record was given immutable ownership to prevent data drift during progressive state changes.
AI is an instrument, not an authority.
Used agentic development environments as operational leverage to execute the entire product lifecycle: repository architecture decomposition, automated test script generation, layout math verification, and multi-viewport regression checks. I personally established the invariants, prompt constraints, and strict acceptance criteria, discarding speculative code and enforcing domain separation.
Key Decisions & Tradeoffs
Enforced strict semantic ownership over import convenience: components used across domains are decoupled through canonical schemas rather than brittle shared state.
Engineered Expiring Disclosure Envelopes and local-first persistence over cloud surveillance stores, ensuring user safety and dignity in high-constraint scenarios.
Prioritized deterministic resource validation rules over automated scrapers, ensuring zero outdated or misleading emergency contact records.
The Result
Shipped a clean, fully audited 26-route Next.js production application with verified route parity, 47 verified public resource records, and automated regression verification.