There are two main issues to address in this bug ticket:
1) Instruction Defence
In our addInfoMessage function (MainComponent.tsx) we add the following info message in the chat for the user:
"{defence.name} defence activated"
(e.g. "prompt enclosure defence activated")
Since the name of the "Instruction Defence" contains the word "defence", this results in:
"instruction defence defence activated"
Only using "Instruction" might not be very clear. Perhaps something like "Pre-Prompt Instruction" would work.
Discuss with the team which wording to use before implementing this.
Acceptance Criteria
Given the user is in the sandbox environment
When they toggle the Information Defence on/off
Then they get a message in the chat without duplicated words
2) Prompt Enclosure
Further more, the "Prompt Enclosure" heading under which these defence can be found is confusing since the Instruction Defence is not enclosing the prompt, but prepending the prompt, so a different name (e.g "Prompt Transformation") would be better.
Discuss with the team which wording to use before implementing this.
Acceptance Criteria
The name for the section "Prompt Enclosure" is changed to something more accurate
Reproduction steps
Steps to reproduce the behaviour:
Go to the "Prompt Enclosure" defence configuration
Toggle the "Instruction Defence" on.
See error in chat box "instruction defence defence activated"
Expected behaviour
Clear names, no duplicating words
Screenshots
Additional context
Add any other context about the problem here.
Acceptance criteria
GIVEN some setup
WHEN an action happens
THEN the expected outcome
Bug report
Description
There are two main issues to address in this bug ticket:
1) Instruction Defence
"{defence.name} defence activated"(e.g. "prompt enclosure defence activated")Only using "Instruction" might not be very clear. Perhaps something like "Pre-Prompt Instruction" would work. Discuss with the team which wording to use before implementing this.
Acceptance Criteria
2) Prompt Enclosure
Further more, the "Prompt Enclosure" heading under which these defence can be found is confusing since the Instruction Defence is not enclosing the prompt, but prepending the prompt, so a different name (e.g "Prompt Transformation") would be better. Discuss with the team which wording to use before implementing this.
Acceptance Criteria
Reproduction steps
Steps to reproduce the behaviour:
Expected behaviour
Clear names, no duplicating words
Screenshots
Additional context
Add any other context about the problem here.
Acceptance criteria
GIVEN some setup WHEN an action happens THEN the expected outcome