This reference provides an information-oriented view of SEA™ Forge governance components. Use it as a quick lookup during implementation.
| Component | Specification | Purpose | Key Commands |
|---|---|---|---|
| Policy Gateway | SDS-042 | Runtime LLM output filtering | Service configuration |
| Evidence Service | SDS-043 | Tamper-evident audit trails | Log queries |
| Semantic Debt Ledger | SDS-016 | Technical debt tracking | Debt queries |
| Spec Cross-Check | REF-019 | Specification validation | just spec-guard |
| Drift Detector | SDS-021 | Model-code divergence detection | just ci |
| SEA™ Invariants | SDS-035 | 15 non-negotiable controls | Built-in |
| Type | Purpose | Template | Validation |
|---|---|---|---|
| ADR | Architecture decisions | just /spec adr |
just spec-guard |
| PRD | Product requirements | just /spec prd |
just spec-guard |
| SDS | Service design | just /spec sds |
just sds-validate <f> |
| SEA™ DSL | Domain policies | Manual | just sea-validate <f> |
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
# Setup & Health
just setup # Install dependencies
just doctor # Environment health check
just ci # Full local CI
just ci-quick # Fast check (lint + specs)
# Spec Validation
just spec-guard # Validate all specs
just sds-validate <f> # Validate SDS YAML
just sea-validate <f> # Validate SEA-DSL
just sea-parse <f> # Generate AST from SEA-DSL
just pipeline <ctx> # Full codegen pipeline
# TDD Cycles
just cycle-start <phase> <cycle> <agent> <slug>
just cycle-complete <phase> <cycle> <agent>
just worktrees-clean
# Workflow Commands
just /spec # Create new specification
just /feature # Implement feature end-to-end
just /cycle # Execute TDD cycle
just /debug # Systematic debugging
just /review # Review PR against standards
| Industry Framework | SEA™ Forge Component | Implementation |
|---|---|---|
| NIST AI RMF - Govern | SEA™ DSL Policies | Executable policy definitions |
| NIST AI RMF - Map | SDS Documents | Context and stakeholder mapping |
| NIST AI RMF - Measure | Evidence Service | Metrics and audit trails |
| NIST AI RMF - Manage | Policy Gateway | Runtime risk management |
| ISO 42001 | SDS-035 Invariants | Management system controls |
| EU AI Act | SEA™ DSL Annotations | Risk classification |
| Risk Level | Score (Severity × Likelihood) | Action |
|---|---|---|
| Low | 0–20 | Monitor routinely |
| Moderate | 21–50 | Implement mitigation |
| High | 51–79 | Immediate mitigation; restrict use |
| Critical | 80–100 | Halt deployment; redesign |
| Category | # | Invariant | Enforcement |
|---|---|---|---|
| Identity | 1 | ifl:hash determinism |
Build-time |
| Identity | 2 | No orphan tokens | Build-time |
| Identity | 3 | Provenance tracking | Runtime |
| Governance | 4 | Forced 3-option proposals | Design-time |
| Governance | 5 | Waiver evidence required | Runtime |
| Governance | 6 | Kill switch < 1 second | Runtime |
| Economics | 7 | Burn-on-use tokens | Runtime |
| Economics | 8 | No reverse circulation | Runtime |
| Economics | 9 | Budget gate enforcement | Runtime |
| Semantic | 10 | Model singularity | Build-time |
| Semantic | 11 | Bidirectional traceability | Build-time |
| Semantic | 12 | Spec-code isomorphism | Build-time |
| Security | 13 | Tamper-evident logs | Runtime |
| Security | 14 | Key rotation policy | Runtime |
| Security | 15 | Fail-closed on unknown | Runtime |
| Filter Type | Purpose | Configuration |
|---|---|---|
| PII Detection | Identify personal data | Regex + ML patterns |
| Jailbreak Prevention | Block prompt injection | Pattern matching |
| Content Safety | Filter harmful outputs | Classification model |
| Copyright Detection | Flag protected content | Fingerprinting |
| Hallucination Check | Detect unfounded claims | Fact verification |
| Event Type | Retention | Purpose |
|---|---|---|
| API Request | 7 years | Audit trail |
| Policy Violation | 7 years | Compliance evidence |
| LLM Interaction | 7 years | Model behavior analysis |
| User Action | 7 years | Access audit |
| System Event | 1 year | Operational monitoring |
| Stage | Input | Output | Tool |
|---|---|---|---|
| Plan Compiler | ADR/PRD/SDS | Structured plan | Internal |
| Semantic Compiler | Plan + SEA™ DSL | SEA™ IR | sea-parse |
| Manifest Compiler | SEA™ IR | Generator manifests | Internal |
| Generator Runner | Manifests | Target code | just pipeline |
| Drift Detector | Code + Model | Pass/Fail | just ci |
| Quadrant | Role | Service | Governance Aspect |
|---|---|---|---|
| Soul | Value hierarchies | SDS-022 | Ethical principles |
| Mind | Intent → Artifact | SDS-027 | Transformation rules |
| Will | OODA loop | SDS-023 | Self-correction policies |
| Body | Resource management | SDS-024 | Budget and compute limits |
| Industry | Regulations | SEA™ Forge Implementation |
|---|---|---|
| Finance | GLBA, SR 11-7, SOX | Full audit trails, human-in-loop |
| Healthcare | HIPAA, FDA SaMD | PII filtering, clinical validation |
| Government | FAR/DFARS, NIST 800-53 | Supply chain docs, access controls |
| General | EU AI Act, GDPR | Risk classification, data minimization |
| Tier | Examples | SEA™ Forge Requirements |
|---|---|---|
| Prohibited | Social scoring, manipulation | Not supported |
| High-Risk | Credit, employment, safety | Full governance stack |
| Limited-Risk | Chatbots, emotion detection | Transparency disclosure |
| Minimal-Risk | Spam filters, recommendations | Standard monitoring |
| Resource | Path |
|---|---|
| System Context Map | docs/specs/00_system_context_map.md |
| Traceability Matrix | docs/specs/shared/reference/009-traceability-matrix.md |
| SEA™ DSL Reference | docs/reference/sea_dsl_language_reference.yml |
| SEA™ DSL Idioms | docs/reference/sea_dsl_language_idioms.yml |
| Agent Instructions | .github/copilot-instructions.md |
| Component | Path |
|---|---|
| Policy Gateway | docs/specs/shared/sds/042-policy-gateway.md |
| Evidence Service | docs/specs/shared/sds/043-evidence-service.md |
| SEA™ Invariants | docs/specs/shared/sds/035-system-invariants.md |
| Semantic Debt | docs/specs/shared/sds/016-semantic-debt-ledger.md |
| Drift Detector | docs/specs/shared/sds/021-drift-detector.md |
| Task | Inputs | Outputs |
|---|---|---|
| Governance setup | Environment, policies | Configured stack |
| Spec authoring | Requirements | ADR, PRD, SDS |
| SEA™ DSL authoring | Business rules | Policy files |
| Code generation | Specs, SEA™ DSL | Generated code |
| Validation | All specs | Pass/Fail report |
| Deployment | Validated code | Running service |
| Monitoring | Service metrics | Dashboards, alerts |
| Template | Location | Purpose |
|---|---|---|
| ADR | just /spec adr |
Architecture decisions |
| PRD | just /spec prd |
Product requirements |
| SDS | just /spec sds |
Service design |
| Risk Register | Handbook §2.4.2 | Risk tracking |
| Readiness Assessment | Handbook §2.4.1 | Capability assessment |
| Last Updated: January 2026 | Version: 1.0.0 |