AI FDE Modes and capabilities

AI FDE separates what it’s doing (the mode) from how it does it (the capabilities), letting the agent stay focused on the right documentation, tools, and context for each task.

Modes define the broad task at hand. You can pick a mode manually, let the agent infer it from your input, or have the agent switch modes automatically as the work evolves. Selecting the appropriate mode loads relevant documentation, grants access to the correct tools, and prevents the agent from using unrelated resources (e.g., a Python‑transform mode won’t load React‑app tooling).

Available modes

• Data integration – Build or modify data pipelines (Python transforms or Pipeline Builder).
• Data connection – Create, manage, and debug Data Connection sources, egress policies, etc.
• Ontology editing – Create or update objects, links, and actions that compose your ontology.
• Functions editing – Write Foundry functions in Logic, TypeScript, or Python.
• Exploration – Read‑only investigation to understand existing platform assets before making changes.
• Governance – Audit permissions, access control, markings, and data‑protection settings.
• Machine learning – Train, evaluate, deploy, and tune models (classification, regression, time‑series forecasting, custom predictive modeling).
• OSDK React – Build React applications or custom widgets that connect to Foundry data.
• Platform Q&A – Ask general questions about how Foundry works.

AI FDE mode selector above the input field with additional configuration options.

Capabilities are granular abilities the agent can invoke across any mode. They map to specific tools and can be turned on or off as the task progresses.

These govern how the agent manages itself and interacts with you:

• Change mode – Switch to a different mode mid‑task without manual intervention.
• Request clarification – Pose multiple‑choice or free‑text questions when more information is needed.
• Generate plan – Draft a step‑by‑step plan for review before execution, useful for ambiguous or multi‑step tasks.
• Load documentation – Pull up Foundry docs on demand.
• Manage context / Manage capabilities – Clean up working memory and enable/disable capabilities as the task evolves.

These perform concrete actions inside Foundry:

• Filesystem – Create folders, browse resources, move items.
• Notepad – Read, create, and update Notepad documents.
• Solution design – Create and edit solution‑design diagrams.
• Execute actions – Run actions against ontology objects.

Note: Capabilities can be enabled or disabled at any time. The Manage capabilities agent capability allows the agent to toggle them automatically during a workflow.

• Modes set the high‑level task context and load the appropriate tools and documentation.
• Capabilities are fine‑grained actions that the agent can call regardless of mode.
• The agent can change mode and manage capabilities on the fly, keeping the workflow fluid.
• Mode‑specific configuration options (e.g., selecting a language for function editing) further tailor the agent’s behavior.
• Both agent and domain capabilities can be toggled dynamically, ensuring the agent only uses what’s needed for the current step.

the AI FDE Mode selector above the input field with additional configuration options.