feat(C1): TrustExportService — canonical trust delta format and export endpoint

[mdproctor]

Context

Trust scores are currently per-deployment. An agent that builds strong security-review trust in one CaseHub deployment starts from Beta(1,1) = 0.5 in every other deployment. TrustExportService enables deployments to publish their trust data so other deployments can import it.

Part of: #51 (Group C epic)
Best after: Group B (#60, #61, #62) to export capability-scoped and multi-dimensional scores
Blocks: C2 (TrustImportService), D1 (trust bootstrapping)

What to implement

Canonical export format

{
  "exportedAt": "2026-04-28T14:30:00Z",
  "exportingDeployment": "acme-corp-prod",
  "actors": [
    {
      "actorId": "claude:security-reviewer@v1",
      "actorType": "AGENT",
      "globalScore": {
        "alpha": 47.3,
        "beta": 8.1,
        "trustScore": 0.853,
        "attestationCount": 55
      },
      "capabilityScores": [
        {
          "capabilityTag": "security-review",
          "alpha": 38.2,
          "beta": 4.1,
          "trustScore": 0.903,
          "attestationCount": 42
        }
      ],
      "dimensionScores": [
        {
          "dimension": "review-thoroughness",
          "score": 0.78,
          "sampleCount": 42
        }
      ]
    }
  ]
}

TrustExportService CDI bean

@ApplicationScoped
public class TrustExportService {

    /** Export all actors with scores above a minimum threshold. */
    public TrustExportPayload exportAll(double minTrustScore) { ... }

    /** Export a specific actor. */
    public Optional<TrustExportPayload> exportActor(String actorId) { ... }

    /** Export actors with changes since a given timestamp (delta export). */
    public TrustExportPayload exportDelta(Instant since) { ... }
}

Optional REST endpoint

GET /ledger/trust/export?minScore=0.5&format=json    (full export)
GET /ledger/trust/export/delta?since=2026-04-01T00:00:00Z  (delta)
GET /ledger/trust/export/actor/{actorId}

Config: casehub.ledger.trust.export.rest-enabled=true

Tests

• Export includes all actors meeting minScore threshold
• Export excludes actors below threshold
• Delta export only includes actors with score changes since since
• Format is valid JSON matching the canonical schema
• Capability and dimension scores included when present

Acceptance criteria

• TrustExportPayload record matching canonical format
• TrustExportService with full, delta, and per-actor export
• Optional REST endpoint behind config flag
• Canonical format handles actors with no capability/dimension scores (backward-compatible)
• Tests covering threshold filtering, delta export, and format validity

Refs #51

[mdproctor]

Completed last session — code shipped in commits on main.