Manage Contact Descriptive Flexfields Setup Task Guidance

Manage Contact Descriptive Flexfields Setup Task Guidance

Purpose and Boundary

Use this skill to export, import, migrate, validate, or resume status for the complete FSM setup package of Manage Contact Descriptive Flexfields. The fixed task code is HZ_MANAGE_CONTACT_DESCRIPTIVE_FLEXFIELDS. Use this skill when planning, sequencing, exporting, importing, or validating setup data for Manage Contact Descriptive Flexfields under Trading Community Foundation.

Do not create, update, delete, deploy, or inspect individual flexfields, contexts, segments, value sets, or attributes. Do not use direct descriptive flexfield REST CRUD, task-local HTTP, loose CSV, templates, filters, or partial package migration. The shared FSM exporter owns export validation and execution. Core Data Loader owns every import, including adapter, endpoint, payload, submission, status, and retry-policy selection.

Fixed Task Context

• Pillar: CX
• Offering: Customer Data Management
• Functional area: Trading Community Foundation
• Functional area code: ORA_HZ_CDM_FOUNDATION
• Task name: Manage Contact Descriptive Flexfields
• Task code: HZ_MANAGE_CONTACT_DESCRIPTIVE_FLEXFIELDS
• Export/import supported: Y
• CSV export/import supported: Y
• Data Loader import mode: FSM

Never accept a caller override for the task name or code. Use the parent skill for broader functional-area sequencing.

Oracle Source Boundary

Use these Oracle sources for deployment and migration guardrails:

• How Flexfields Work with Oracle Application Cloud Architecture
• Overview of Flexfield Configuration
• Business Objects for Moving Common Reference Objects

Oracle states that only Deployed or Deployed to Sandbox flexfields are eligible to be moved. Treat that as Fusion migration behavior, not as a manual status-evidence prerequisite for starting an FSM export. Do not ask the user for source deployment status, observation time, or an evidence reference before export; rely on the export process and its ProcessLog to report exportability. Descriptive-flexfield sandbox testing remains required before manual publication. Import can change deployment status and can submit affected flexfields for redeployment; neither an accepted import nor automatic redeployment proves that target deployment and functional verification passed. Do not apply extensible-flexfield background-deployment queue behavior here.

Supported Operations

1. export: authorize and run one complete, unfiltered, non-template task export.
2. import: validate and authorize one complete, unchanged FSM ZIP and run one import.
3. migrate: finish and persist source export, then perform separate target preflight and approval before import.
4. validate: validate fixed task context, operation-specific routing, package identity when supplied, and approval readiness without submission.
5. status: resume polling or retrieve results for one persisted ProcessId; never create another process.

Export and Import Adapter Contract

For export, preserve the existing export route: validate the fixed task and whole-task scope, display one exact task-level export preview, obtain explicit approval for that environment and process creation, and invoke shared.setup-export.task-level-export directly. Do not route export through core.data-loader, approved-loader-payload, or executionRoute=data-loader. An export never requires a pre-existing ZIP path or SHA-256.

Use this fixed export final data:

textCopy code
operation=export
task_code=HZ_MANAGE_CONTACT_DESCRIPTIVE_FLEXFIELDS
task_name=Manage Contact Descriptive Flexfields
offering_name=Customer Data Management
functional_area_name=Trading Community Foundation
template_only=false
export_criteria=[]
output_dir=project/{project_id}/exports/customer-data-management/trading-community-foundation/manage-contact-descriptive-flexfields/
execution_adapter=shared.setup-export.task-level-export
wait_seconds=300
poll_interval=10

After export approval, invoke shared.setup-export.task-level-export exactly once. The shared exporter owns the export POST, ProcessId, polling, FileContent, and ProcessLog behavior. After download, compute and persist the exported ZIP's SHA-256. Do not implement task-local export HTTP.

For import, route exclusively through core.data-loader. Require one unchanged ZIP and supply only this task context to Core Data Loader:

textCopy code
operation=import
task_code=HZ_MANAGE_CONTACT_DESCRIPTIVE_FLEXFIELDS
task_name=Manage Contact Descriptive Flexfields
setup_data_zip_path=<validated ZIP path>
setup_data_zip_sha256=<SHA-256 of the exact ZIP bytes>

Require a valid ZIP signature and exactly one case-sensitive member basename ASM_SETUP_CSV_METADATA.xml. Require its single task identity to equal the fixed task code. Preserve the archive byte for byte. Do not construct or change an import payload or request a revision-mismatch override.

Let Core Data Loader prepare the final immutable artifact, invoke its data-load Task Hook, build the approval preview, and obtain the user's separate explicit approval. The Task Hook validation envelope must remain bound to the exact ZIP SHA-256. Core Data Loader then owns UCM upload, initiation-payload construction, endpoint and FSM adapter selection, accepted-request reporting, status diagnostics, and retry policy. This pillar skill must not select or invoke an import adapter, endpoint, routing field, payload builder, or retry behavior.

Post-Import Deployment Verification

After a terminally successful import, require target deployment evidence and a functional verification result:

• Return import_succeeded_deployment_verified only when target status is Deployed or Deployed to Sandbox and the contact DFF functional test passed in the named target environment.
• Return import_succeeded_deployment_pending when import succeeded but target status or functional verification is missing, nonterminal, Edited, Patched, or Error. Report the manual validation, sandbox deployment/test, publication, or redeployment action required; do not perform it.
• Return import_failed only when the FSM import process fails. Retrieve and persist ProcessLog when available.

Never collapse import acceptance, import process success, automatic redeployment submission, target deployment status, UI rendering, and functional verification into one success claim.

Preflight and Approval

Before any live export or import:

1. Resolve a valid project, exact environment, fixed task context, durable output paths, and task-local compiled rules.
2. Authenticate separately. If a cache-managed token fails on a read-only preflight, refresh once and repeat only that GET. A second rejection blocks submission.
3. Confirm live task discovery returns the fixed task code and CSVExportImportSupportedFlag=Y. Do not reject solely because offering or functional-area fields are absent in the live row.
4. For export, validate the fixed whole-task non-template contract. Do not ask for deployment status, a ZIP path, or a SHA-256 before export.
5. For import, validate the unchanged archive, exact metadata identity, hash, and target, then hand control exclusively to Core Data Loader.
6. For export, display one redacted task-level preview containing the environment, task, whole-task scope, output path, adapter, execution options, and one-process impact. Bind approval to those values and then call the shared exporter directly.
7. For import, let Core Data Loader display the immutable preview and obtain approval bound to its environment, task, ZIP/hash, validation envelope, and impact. Any changed import-bound value restarts Core Data Loader validation, Task Hook, preview, and approval.

Submission, Polling, and Reconciliation

Persist an external identifier returned by the owning exporter or Core Data Loader immediately in an append-only run record. Delegate polling, log retrieval, and status semantics to that owner. On timeout, preserve the identifier and safe resume stage; never submit a replacement process.

Never automatically retry an export or import mutation after 401, 403, timeout, connection closure, or an uncertain response. For import, follow only Core Data Loader's reconciliation and status contract. For export, reconcile using read-only process lookup for the fixed task and run-start window:

• One matching process: persist and resume its ProcessId.
• No matching process: record not_created; a new POST needs fresh approval.
• Multiple or unidentifiable matches: record ambiguous and fail closed.

Diagnostics and Results

Classify failures as input_validation, catalog_or_metadata_mismatch, auth_expired_or_invalid, authorization_or_privilege, transport_or_environment, fusion_api_error, process_timeout, process_failed, artifact_persistence, approval_blocker, or deployment_verification.

Report the exact task, environment, operation, owning execution skill, ProcessId, terminal or last-known process status, artifact paths, ZIP SHA-256 when applicable, deployment-verification result, warnings, blockers, and one safe next action. Never print or persist tokens, secrets, Authorization headers, token-cache contents, full base64 payloads, or sensitive environment values.

Refusal Rules

Refuse:

• Direct descriptive flexfield REST CRUD or a task-local endpoint/script.
• Individual context, segment, value-set, attribute, deployment, or UI changes.
• Template/spec exports, export criteria, filters, or partial migration.
• Loose CSV, an edited/repacked ZIP, missing or ambiguous metadata identity, revision mismatch override, task-code override, or changed approved artifact.
• Task-local import adapter, endpoint, routing-field, payload, or retry selection.
• A new submission while a previous response or ProcessId is uncertain.
• A complete-success claim without target deployment and functional evidence.

Golden Scenarios

Ownership and Review

• Owner: CX Customer Data Management pillar team.
• Functional reviewer: CX Customer Data Management Pillar Lead or Trading Community Foundation SME.
• Framework reviewer: FSCR A-Team/core reviewer.
• QA reviewer: required for FSM routing, approval, deployment, or verification changes.

Track token usage for generated project artifacts and skill-run records using the manifest policy.