OpenAI Agents Onboarding Playbook

Purpose

This playbook is the correctness-first onboarding path for new teams using OpenAI Agents with Decision Gate.

Readiness is based on deterministic correctness and artifact validity, not elapsed calendar time.

Canonical Inputs

• Protocol contract: LLM-native playbook
• Onboarding fixture pack: examples/agentic/onboarding/basic
• End-to-end runner: examples/frameworks/openai_agents_live_loop.py

End-to-End Flow

1. Capability discovery:
   • decision_gate_providers_list
   • decision_gate_provider_contract_get
   • decision_gate_provider_check_schema_get
2. Authoring artifacts:
   • claim_inventory
   • capability_matrix
   • claim_condition_map
3. Fast loop:
   • decision_gate_schemas_register
   • decision_gate_precheck
4. Live boundary:
   • decision_gate_scenario_define
   • decision_gate_scenario_start
   • decision_gate_scenario_next
5. Verification:
   • decision_gate_runpack_export
   • decision_gate_runpack_verify

Runbook

1. Start Decision Gate MCP endpoint.
2. Ensure OpenAI Agents adapter dependencies are installed.
3. Run the onboarding loop:

python3 examples/frameworks/openai_agents_live_loop.py \
  --fixture-dir examples/agentic/onboarding/basic

4. Inspect output JSON for required artifacts and verification status.

Correctness Acceptance

Onboarding is complete only if all checks pass:

1. Required artifacts are present and structurally valid:
   • claim_inventory
   • capability_matrix
   • claim_condition_map
   • enforcement_verdict
2. Live boundary outcome is an allowed pass state for your boundary policy.
3. runpack_verify.status is pass.
4. Blocking scenarios emit explicit blocking reasons.

Failure Taxonomy

Classify failures into deterministic buckets:

• capability_mismatch
• schema_mismatch
• comparator_or_type_mismatch
• lane_requirement_unmet
• runpack_verify_failed

Do not use an unclassified unknown bucket for release decisions.