Day4

GPTClaudeGemini··420 copies·updated 2026-07-14
day4.prompt
# Day 4 - Output Evaluation (60 minutes)

Goal: Measure output quality with a repeatable rubric, then improve the prompt deliberately.

## Inputs

- Prompt template: `prompts/template.md`
- Rubric: `evals/day4-rubric.md`
- Output log: `examples/day4-output.md`

## Step 1 (20 min): Generate baseline output

1. Run `prompts/template.md` with your weekly context.
2. Save the output as "Baseline Output" in `examples/day4-output.md`.

## Step 2 (20 min): Score baseline output

Using `evals/day4-rubric.md`, score each criterion (1-5):
- Specificity
- Actionability
- Constraint Fit
- Clarity
- Risk Awareness
- Progress Measurability

Record the total score and 2-3 short observations.

## Step 3 (15 min): Improve one prompt element

Change exactly one part of `prompts/template.md`:
- output format, or
- constraints, or
- success criteria

Rerun the prompt and save as "Improved Output".

## Step 4 (5 min): Rescore and compare

Score the improved output with the same rubric.
Document score change and one lesson learned.

## Suggested Commit

when to use it

Community prompt sourced from the open-source GitHub repo vramana-7/cursor-skill-builder (no explicit license). A "Day4" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.

tags

roleplaycommunitygeneral

source

vramana-7/cursor-skill-builder · no explicit license