Feature hasn't been suggested before.
Describe the enhancement you want to request
Users currently have to repeat the same instructions at the start of every new conversation (e.g., "always write tests", "use TypeScript", "follow the project conventions"). A preset instructions system would let users save reusable prompts and apply them to any session.
Suggested approach:
-
Preset management UI: Add a section in the settings or chat panel where users can create, edit, and delete preset instruction sets.
-
Quick apply: In the chat input area, add a dropdown or button to select and apply a preset. The preset text is prepended to the next user message or injected as a system-style instruction.
-
Default preset: Allow users to set a default preset that is automatically applied to all new sessions.
-
Sharing: Allow exporting/importing presets as JSON files so users can share useful instruction sets.
This is different from #22141 (custom instructions passed to the agent every session) because this focuses on user-facing preset management with a UI, rather than programmatic agent instructions.
Feature hasn't been suggested before.
Describe the enhancement you want to request
Users currently have to repeat the same instructions at the start of every new conversation (e.g., "always write tests", "use TypeScript", "follow the project conventions"). A preset instructions system would let users save reusable prompts and apply them to any session.
Suggested approach:
Preset management UI: Add a section in the settings or chat panel where users can create, edit, and delete preset instruction sets.
Quick apply: In the chat input area, add a dropdown or button to select and apply a preset. The preset text is prepended to the next user message or injected as a system-style instruction.
Default preset: Allow users to set a default preset that is automatically applied to all new sessions.
Sharing: Allow exporting/importing presets as JSON files so users can share useful instruction sets.
This is different from #22141 (custom instructions passed to the agent every session) because this focuses on user-facing preset management with a UI, rather than programmatic agent instructions.