Skip to content

Agent Charter

Purpose

Defines what automation agents (Cursor, VS Code agents, Copilot, Antigravity, GitHub Actions, future Ara AO) may and may not do under OjuIDC governance.

Authority

agent_authority:
  may:
    - analyze
    - classify
    - summarize
    - document
    - lint
    - validate
    - create_prs
    - create_issues
    - build_docs
    - archive_evidence
  may_not:
    - authorize_governance
    - authorize_ws2
    - approve_retirement
    - transfer_fsmo
    - modify_active_directory
    - demote_domain_controller
    - sign_attestations
    - create_manifests_as_authorization
    - mutate_progress_ledger_as_authorization

Detailed prohibitions: prohibited-actions.md.

Prohibited Assumptions

prohibited_assumptions:
  - idc.olutech.systems_exists_as_production
  - akoda_retired
  - ws2_authorized
  - fsmo_transferred
  - workload_cutover_complete

Current vs Planned

current_state:
  ad_domain: ojuidc.olutech.systems
  writable_dc: akoda-dc01
  rodc: irin-dc01
  akoda_retirement: blocked
  ws1_endpoint: READY_FOR_HUMAN_REVIEW

planned_state:
  agbara: intended_primary_workstation
  oluso: intended_persistence_platform
  docs_portal: docs.iwa.olutechsys.com

under_evaluation:
  - idc.olutech.systems
  - separate_training_forest

Required Outputs

Every agent deliverable should include:

  • findings
  • evidence
  • assumptions
  • risks
  • dependencies
  • rollback
  • recommendations