System Prompt Builder

Composed a system prompt with 6 section(s): Role, Objective, Capabilities, Constraints, Tone, Output format.

How it works

• Placed the role as the opening instruction.
• Added each provided section (objective, capabilities, constraints, tone) in a fixed order.
• Closed with the selected output-format instruction.

Methodology

1. Open with the role as the first instruction so it frames the task.
2. Add the objective, then optional capability and constraint bullet lists (split on newlines or semicolons), then a tone note.
3. Close with a fixed output-format instruction drawn from a closed set (prose, Markdown or JSON).

Usage

1. Enter the role and the objective.
2. Optionally add capabilities, constraints and a tone, one per line.
3. Pick the output format and copy the composed system prompt.

Examples

These are examples, not current values: they use fixed inputs, while the interactive tool computes live results in your browser.

A concise reviewer

{
  "role": "a senior code reviewer",
  "objective": "Review a diff and list concrete issues.",
  "capabilities": "Read code in many languages\nExplain the reasoning briefly",
  "constraints": "Be specific and cite line numbers\nDo not rewrite the whole file",
  "tone": "direct and technical",
  "outputFormat": "markdown"
}

A JSON extractor

{
  "role": "a data extractor",
  "objective": "Extract the order fields from the text.",
  "outputFormat": "json"
}

Limitations

Structure over prose

A system prompt is easier to maintain when role, objective, constraints and output format are separate sections. This builder applies one stable order so prompts across a project stay comparable.