Documentation

Use the Active Profile

Use the active instruction profile toggle when inserting prompts, pins, and scratchpads.

Profile workflow2 min read

Current guide

Use the Active Profile

Use the active profile

This is the working end of instruction profiles: the Use active instruction profile toggle in the pop-up, and what happens at insertion when it is on. Everything upstream is preparation; this is where the standing rules actually reach your prompts, and where you decide, chat by chat, whether they should. This page covers the toggle, the attach behaviour, and the judgment of when to run with the profile and when to insert clean.

Prerequisites:

The toggle#

  1. Click the pinned Cognardo Workbench icon to open the pop-up.
  2. Find Use active instruction profile under PREFERENCES.
  3. Set it:
    1. On: the active profile's instructions attach to the prompts you insert.
    2. Off: prompts insert clean. The active profile stays selected and waits; flipping the toggle back resumes it with nothing to re-select.

The toggle is the day-to-day switch, and the library is the mode switch: flip the toggle when a single task should run clean, change the active profile when the mode of work changes. See manage profiles.

What happens at insertion#

With the toggle on, inserting a prompt carries the active profile with it:

  1. Insert a prompt by any library route: Insert Prompt/Pin in the sidebar, or the // shortcut. Rerun carries the same attach behaviour. See insert a prompt from the sidebar and insert a prompt with the shortcut.
  2. The active profile's instructions attach to the prompt as it lands, joined by a line break with nothing else marking the boundary, and the combined text sits in the chat input.
  3. Nothing is sent. Scroll to the end of the input to see exactly what was added, adjust either part, then send it yourself.

The saved prompt is never touched: the profile attaches to the copy landing in the input, and the library keeps the prompt exactly as written. A prompt with {{variable}} placeholders fills them first, through the Fill Prompt Variables dialog, and the profile rides on the completed text.

When on, when off#

  • On is the default for a mode: The point of a profile is consistency, so once the active profile matches the work, leave the toggle on and let every inserted prompt carry the rules.
  • Off for the exception: A one-off task where the standing rules would mislead (a different audience, a format the profile forbids) is a toggle flip, not a profile edit. Insert clean, send, flip it back.
  • Wrong rules attaching is a mode problem: If the instructions arriving are right for last week's work, the fix is the library: mark the right profile active. See manage profiles.

FAQ#

Does the toggle change which profile is active?+

No. The toggle decides whether the active profile applies; the library decides which profile that is. Off and back on, the same profile resumes.

How do I check what will attach before I send?+

Scroll to the end of the chat input before sending. The profile's instructions are appended after a line break with no label of their own, so scrolling is the only way to see exactly what joined the prompt.

Does the profile attach when I type a prompt by hand?+

No. Profiles ride on insertion from your library and the shortcut. Hand-typed text inserts nothing and attaches nothing.

Does applying a profile use AI credits?+

No. Profiles attach as part of insertion, and insertion is free on every plan. Credits apply only to Enhance actions.

Does the toggle behave the same on claude.ai?+

Yes. One toggle and one active profile serve chatgpt.com and claude.ai, and the attach behaves identically on both.