FSCR — Oracle Fusion Setup and Configuration Skills Registry

FSCR is separately released sample and reference material for planning and configuring Oracle Fusion Cloud implementations. It is not an Oracle Fusion Cloud product feature and is not covered by Fusion Cloud product support.

Use FSCR only with an Oracle Fusion Cloud environment that you are authorized to access. You must configure your own environment connection, supply your own authorized credentials, and have the permissions required for each operation. FSCR does not include a Fusion environment or credentials. Review generated plans, requests, and changes before running them in your environment.

This repository contains the proposed skills taxonomy for AI-enabled Oracle Fusion Cloud implementation acceleration.

FSCR is released under the Universal Permissive License (UPL), Version 1.0. See LICENSE.txt for the full text.

Developers choosing to distribute a binary implementation of this project are responsible for obtaining and providing all required licenses and copyright notices for the third-party code used in order to ensure compliance with their respective open source licenses.

The repository is organized to support:

1. End-to-end implementation orchestration.
2. Reusable core Fusion automation services.
3. Shared cross-pillar implementation methodology.
4. Independent ERP, HCM, SCM, and CX pillar contributions.
5. A local implementation project workspace that persists state and artifacts across sessions.
6. Standardized schemas, governance, and skill metadata.

Start with FSCR Onboarding for an introduction to FSCR, an overview of how its skills and project workspace fit together, and guidance on any login or authentication question or issue.

The onboarding journey guides you through:

1. Creating a public OAuth client (Mobile Application) in your Fusion instance's Identity Domain, using the IAM setup steps and diagrams.
2. Populating core/auth/fusionEnv.properties with your environment, Client ID, callback, and service scopes.
3. Understanding first-time Fusion environment enrollment in ~/.fusionaim/oauth-trust.json, PKCE browser login, and access-token acquisition and reuse through the authentication flow guide.
4. Troubleshooting configuration, enrollment, callback, scope, and access issues.

Example requests include “Introduce me to FSCR”, “Help me set up the Mobile Application”, “Help me populate fusionEnv.properties”, and “Why is my login failing?”. Onboarding provides guidance; requested enrollment and login execution use Fusion OAuth Authentication and its workflow checks.

To start or resume implementation work, begin with the Implementation Router.

FSCR skills require Oracle Fusion Applications Release 26D with the 16 October 2026 Weekly Maintenance Pack for Fusion Applications Update 26D, or a higher release or maintenance pack.

Customers should use Python 3.11 or newer when running repository-provided automation scripts.

Minimum compatible Python version:

Python 3.10+

Recommended customer baseline:

Python 3.11 or 3.12

Several scripts use modern Python type syntax such as str | None and list[dict[str, Any]], so Python 3.9 or older is not supported.

FSCR has been tested with the GPT 5.6 Terra Medium model. Performance varies by model; select a model whose reasoning capability is appropriate for the activity being pursued in FSCR.

For the full repository taxonomy and logical skill architecture, see SKILLS_TAXONOMY.md.

When users clone this repository for an implementation, the project folder is their local workspace. Guided skills should not depend on chat history for continuity. They should persist their outputs as project artifacts and update project/{project_id}/state/project-state.json so the implementation can resume across sessions.

Generated artifacts are organized under project/{project_id}/artifacts by mirroring the skill structure:

project/{project_id}/artifacts/{router|core|shared|pillars}/...

Examples:

project/{project_id}/artifacts/router/project-objective-classifier/
project/{project_id}/artifacts/shared/project-intake/scope-definition/
project/{project_id}/artifacts/shared/implementation-planning/offering-task-map-builder/
project/{project_id}/artifacts/pillars/scm/product-management/
project/{project_id}/artifacts/pillars/scm/product-management/item-organizations/

Configuration exports are stored under project/{project_id}/exports, curated baselines under project/{project_id}/baselines, design decisions under project/{project_id}/decisions, and configuration blueprints under project/{project_id}/blueprints.

When a skill is revisited, major changes create a new artifact version. Minor changes update the current artifact only after user confirmation.

Oracle-contributed canonical skills under router/, core/, shared/, and pillars/, and content under governance/, schemas/, and knowledge/, are read-only by default. Changes require explicit maintainer approval and repository review. Keep implementation-specific artifacts and customizations in project/{project_id}; a user request alone is not maintainer approval.

Pillar teams should contribute under their owning pillar folder. Reusable cross-pillar behavior should be proposed under shared. Technical execution primitives should be proposed under core.

Each skill should include a manifest that declares its inputs, outputs, dependencies, execution mode, owner, and applicability. Skills that produce project artifacts should also declare persists_project_state, state_updates, and workspace_outputs metadata.

This repository contains sample skills, scripts, reference material, and workspace templates. Review each workflow and its prerequisites before using it with your authorized Fusion environment.

This is a scaffold repository. Teams should add skill implementation files, manifests, test fixtures, knowledge artifacts, and project workspace artifact templates within the established folders.