LegalTech scaffold

A complete program, not a fragment: it compiles as written. Copy it, rename the placeholder identifiers to your domain, and run axon check — the compiler will tell you what your renaming broke. An MCP client can also generate it through the axon.compose tool by naming the legaltech domain.

// AXON LegalTech scaffold — multi-tenant SaaS for legal-technology
// products (contract automation, e-discovery, IP portfolio mgmt,
// matter management).
//
// Differs from `legal.axon` (in-house counsel pattern): this scaffold
// targets PRODUCT deployments — multi-tenant isolation, audit-heavy,
// capability-gated endpoints, replay-token binding for write paths,
// product-grade shield with PrivilegeShield + audit categories.

// ── Regulated data types ──────────────────────────────────────────

type ContractDocument {
    document_id:    Text
    tenant_id:      Text
    party_count:    Int
    jurisdiction:   Text
    body:           Text
}

type ClauseFinding {
    clause_id:        Text
    category:         Text
    // `severity` is a reserved AXON keyword (shield.severity:) — use
    // a non-keyword field name for type bodies.
    severity_class:   Text
    rationale:        Text
}

type ContractAnalysis {
    findings_summary: Text
    citations:        Text
}

type AnalyzeRequest { doc: ContractDocument }
type AssessLiabilityRequest { doc: ContractDocument }

// ── Identity + grounding ──────────────────────────────────────────

persona ContractAnalyst {
    domain: ["contract-law", "commercial-litigation", "intellectual-property"]
    tone: precise
    confidence_threshold: 0.85
    cite_sources: true
}

context LegalProduct {
    memory: session
    language: "en"
    depth: exhaustive
    max_tokens: 4096
    temperature: 0.2
}

anchor NoLegalAdvice {
    require: source_citation
    confidence_floor: 0.85
    unknown_response: "Insufficient citation evidence — escalate to a licensed attorney."
    on_violation: raise UnauthorisedLegalAdviceError
}

// ── Shields ───────────────────────────────────────────────────────

shield PrivilegeShield {
    scan:       [prompt_injection, pii_leak, data_exfil, model_theft]
    on_breach:  quarantine
    severity:   high
    compliance: [SOC2]
}

// ── Flows ─────────────────────────────────────────────────────────

flow AnalyzeContract(doc: ContractDocument) -> FlowEnvelope<ContractAnalysis> {
    step Extract {
        given: doc
        ask: "Extract every clause + its plain-language summary, with citation anchors."
        output: FlowEnvelope<ContractAnalysis>
    }
    return Extract.output
}

flow AssessLiability(doc: ContractDocument) -> FlowEnvelope<ContractAnalysis> {
    step Score {
        given: doc
        ask: "Score liability exposure per clause; cite jurisdictional precedents."
        output: FlowEnvelope<ContractAnalysis>
    }
    return Score.output
}

// ── HTTP boundaries ───────────────────────────────────────────────

axonendpoint ContractAnalysisAPI {
    method:      post
    path:        "/v1/contracts/{tenant_id}/analyze"
    body:        AnalyzeRequest
    execute:     AnalyzeContract
    output:      FlowEnvelope<ContractAnalysis>
    shield:      PrivilegeShield
    backend:     auto
    compliance:  [SOC2]
    retries:     1
    timeout:     30s
    requires:    ["legaltech.contracts.read"]
}

axonendpoint LiabilityAssessmentAPI {
    method:      post
    path:        "/v1/contracts/{tenant_id}/liability"
    body:        AssessLiabilityRequest
    execute:     AssessLiability
    output:      FlowEnvelope<ContractAnalysis>
    shield:      PrivilegeShield
    backend:     auto
    compliance:  [SOC2]
    retries:     1
    timeout:     60s
    requires:    ["legaltech.contracts.assess"]
}