kind: uapf.package
id: lv.pppa.exchange-governance
name: MI apmaiņas ekosistēmas pārvaldība
description: |
  Level-4 UAPF process package encoding the governance of the AI LV Exchange
  (AI LV Exchange MCP Node Profile v0.2, Part II) as executable, versioned,
  forkable models rather than prose.

  BPMN — the admission process: application, evidence gathering, source-
  independence check, domain-control and key-control challenges, assurance-level
  determination, participation agreement, credential issuance, publication of
  the reliance statement.

  DMN — four decision tables carrying the algorithm the profile currently
  states in prose:
    * avota-neatkariba          — §14.2 / §17.2 verified vs relied-upon facts,
                                  and the exclusion of circularity. Legal facts
                                  MUST come from Uzņēmumu reģistrs and MUST NOT
                                  come from the applicant or from the applicant's
                                  own product — which binds hardest precisely when
                                  the applicant is a commercial provider of
                                  company-register data.
    * noteikt-ticamibas-limeni  — §6.1 / §17 assurance level AL0..AL3.
    * piekluves-limena-prasibas — §6.3 minimum-requirements matrix (T0..T3 to
                                  AL, MCPF level, mandate, entitlement, audit,
                                  status-cache), plus §6.4: a legally significant
                                  answer is never T0 regardless of the secrecy of
                                  the datum.
    * atsaukuma-pamatojums      — §15 admissibility of an adverse decision:
                                  exhaustive grounds, 30-day notice, cure,
                                  5-working-day emergency review, two distinct
                                  officers. Fail-closed by default.

  CMMN — the withdrawal / suspension / appeal case: discretionary and mandatory
  work in an unstructured matter, with the two-officer decision and the
  conflict-free appeal as first-class human tasks.

  Nothing here infers. This package is the deterministic governance of an AI
  system, not an AI system itself (Regulation (EU) 2024/1689, Art. 3(1) and
  Recital 12). It is in scope of the sandbox because it governs admission to
  the trust and identity plane of the Process-as-Code AI system.

level: 4
version: "0.1.0"

requires_capabilities:
  - pppa.evidence.fetch@1+
  - pppa.domain.challenge@1+
  - pppa.key.challenge@1+
  - pppa.credential.issue@1+
  - pppa.credential.revoke@1+
  - event.emit@1+

profiles_supported:
  - uapf-ip-orchestrated

guardrails: resources/guardrails.yaml

includes: []
dependencies:
  "AI LV Exchange MCP Node Profile": "0.2"

cornerstones:
  bpmn: true
  dmn: true
  cmmn: true
  resources: true

paths:
  bpmn: bpmn
  dmn: dmn
  cmmn: cmmn
  resources: resources
  metadata: metadata
  algorithms: algorithms

algorithm_cards: false

exposure:
  mcp:
    enabled: true
    runnable: true
    exposedEntrypoints:
      - "Process_DalibniekaUznemsana"
      - "Case_AkreditacijasAtsaukums"
    exposedArtifacts:
      - manifest
      - bpmn
      - dmn
      - cmmn
      - docs

owners:
  - type: org
    id: pppa-lv
    contact: info@pppa.lv

lifecycle: draft