Improve prompt package using real LLM output failures #85

Open
opened 2026-05-23 00:57:29 -04:00 by JackFrostbyte · 0 comments

Objective

Improve the LLM prompt package based on actual output from external LLM tests.

Scope

Review where external LLMs misunderstood:

  • Device profile constraints.
  • Aura profile intent.
  • Schema requirements.
  • Reserved zone restrictions.
  • LED index ranges.
  • Output-only JSON requirements.
  • Supported effect types.

Then improve the prompt package to reduce those failures.

Acceptance criteria

  • Prompt package is clearer and more reliable.
  • Prompt explicitly addresses common failure patterns.
  • Prompt still avoids asking the LLM for Python code.
  • Prompt output requirements remain data-only JSON.
# Objective Improve the LLM prompt package based on actual output from external LLM tests. # Scope Review where external LLMs misunderstood: - Device profile constraints. - Aura profile intent. - Schema requirements. - Reserved zone restrictions. - LED index ranges. - Output-only JSON requirements. - Supported effect types. Then improve the prompt package to reduce those failures. # Acceptance criteria - Prompt package is clearer and more reliable. - Prompt explicitly addresses common failure patterns. - Prompt still avoids asking the LLM for Python code. - Prompt output requirements remain data-only JSON.
JackFrostbyte added reference develop 2026-05-23 00:57:33 -04:00
Sign in to join this conversation.
No labels
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
JackFrostbyte/rgb-aura#85
No description provided.