Part 2 showed where the skill and artifact files live in a project. After the OAuth setup and Pro-Code Builder setup, you can bring those files, the Fusion AI Studio extension, a coding assistant, and the CLI together in VS Code.

Install the Fusion AI Studio extension

Download aistudio-extension.zip from the Oracle Fusion AI Studio extension package for the release branch that matches your Fusion environment. The link opens the release-26C branch. If your environment is on a different release, switch to the matching release branch in the repository and open its extensions folder. Extract the .vsix file. In VS Code, open Extensions, choose the … menu, select Install from VSIX…, and open the extracted file. Reload VS Code if prompted.

In Extensions → Installed, look for Fusion AI Studio by oracle. Open its Features → Commands view to see what this installed version supports. The example shown here was installed from a VSIX; check your own installed version against the package for your release.

One workspace, several ways to work

VS Code area Use it for
Explorer Find the local app, workflow, and other artifact files.
Fusion AI Studio visual editor Open an app or workflow file on a canvas.
Command Palette Find extension actions such as Fetch from Server.
Coding assistant Inspect related local files, plan an edit, and explain what changed.
Integrated terminal Run a known AI Studio CLI check and read its result.

You can add a coding assistant extension alongside Fusion AI Studio. The example workspace has Codex and Claude Code installed; GitHub Copilot and Amazon Q Developer are other VS Code options. Use the assistant available under your organization’s account and data rules.

Fetch an artifact from the server

Open Show and Run Commands with Ctrl+Shift+P and type Fusion. Select Fusion AI Studio: Fetch from Server. The first picker asks for an artifact type. Choose App to retrieve an Agentic App, or Workflow for a workflow file. Then search by name and read the status beside each result. The same name can appear as both [DRAFT] and [PUBLISHED].

In this example, select COE Compliance and Security [DRAFT]. Its local app file opens from src/apps in the AI Studio App Editor. You can run Fetch from Server again, choose Workflow, and open a related .wf file from src/workflows.

COE Audit Passport file open in the Fusion AI Studio Workflow Editor inside VS Code.
Figure 1. A fetched workflow file opens in the Fusion AI Studio visual editor inside VS Code. The canvas toolbar and Debug action remain in the same workspace as the app and other project files.

The visual editor also lets you make and save a local change to an app or workflow. Review the result in the editor and validate the local file before requesting a server save. A local save and a server DRAFT update are separate steps.

Ask a coding assistant to inspect the local files

Open the app beside your coding assistant and start with a read-only request:

Inspect coe_compliance_and_security.apps. List each workflow code referenced by the app, its matching local .wf file, and whether anything is missing. Do not edit files, fetch artifacts, run commands, or save anything to Fusion.

Compliance and Security app editor beside a Codex chat with a read-only request to inspect its local workflow files.
Figure 2. Keep the app editor and coding assistant side by side in VS Code.
Compliance and Security app canvas beside a Codex table matching four app panels to local draft workflow files.
Figure 3. The assistant’s local inventory maps the app to four DRAFT workflow files and reports no missing files. This is an inventory check; validation is the next step.

Run a local check in the integrated terminal

Open … → Terminal → New Terminal. Check that the prompt is at the AI Studio project root, where .agents/skills/aistudio and src are available. Run a validation command for the file you want to check:

VS Code More Actions menu open to Terminal and New Terminal.
Figure 4. Open a terminal without leaving the app editor.
node .agents/skills/aistudio/scripts/aistudio.js validate-workflow --file src/workflows/coe_audit_passport.wf
VS Code terminal showing the Passport workflow validation result with zero errors and a separate tests-not-created notice.
Figure 5. The local Passport file passes validation with zero errors. The separate Tests Not Created notice means generated workflow tests still need attention.

Read the result below any Node warning. In this capture, ok: true and errorCount: 0 confirm that the local file passed validation. That result does not confirm that a server DRAFT was updated. For more commands and when to use them, see the AI Studio CLI Command Map.

Key takeaways

  • Fetch the artifact type you need from the Fusion command list, and check whether you selected DRAFT or PUBLISHED.
  • Open the resulting local file in the Fusion visual editor, with related files in the same Explorer.
  • Use a coding assistant for a read-only inventory before asking for changes, and use the integrated terminal to validate a known file.
  • Keep local edits, local validation, generated tests, and server saves distinct when reviewing progress.

Previous: Pro-Code Builder Experience  |  Next: The Development Lifecycle with the AI Studio CLI
Back to the Learning Path for Fusion AI Agent Studio CLI