skills.rest
tomoya0318
tomoya0318COMMUNITY
Explore all skills by this author →
𝕏@tomoyan1018553GitHub

empirical-prompt-tuning

Run prompts with a subagent to surface ambiguities and guide refinement.

Updated Mar 3, 2026
Search Tags:#evaluation#iteration#subagent#bias#prompt-engineering#prompt
One-click install
npx skills add https://github.com/tomoya0318/dotfiles --skill empirical-prompt-tuning-tomoya0318
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: empirical-prompt-tuning
Source: https://github.com/tomoya0318/dotfiles/tree/main/dot_claude/skills/empirical-prompt-tuning
Command: npx skills add https://github.com/tomoya0318/dotfiles --skill empirical-prompt-tuning-tomoya0318

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Prompt quality is hard to judge from the writer's perspective. This Skill uses a bias-free agent to execute prompts and collect objective, bidirectional feedback, enabling iterative improvements until the desired reliability is reached.

Core Features & Use Cases

  • Bias-free evaluation: Run prompts with a subagent to surface ambiguities and measure execution against a predefined checklist.
  • Iterative refinement: Pair executor self-reports with instruction-side metrics to drive focused improvements.
  • Operational workflow: Provides a structured process (iteration, evaluation, and differences application) for prompt and skill hardening after new prompts or changes or when agent behavior deviates.

Quick Start

Dispatch a new subagent to read the target prompt, execute a defined scenario, and return a structured report that guides iterative improvements.

Frequently Asked Questions about empirical-prompt-tuning

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I evaluate and improve prompt quality without writer bias?▼

Bias-free prompt evaluation uses a dispatchable subagent to execute scenarios against a predefined requirements checklist, returning structured reports with tool usage metrics and self-reported observations to guide iterative refinement.

What is the best way to refine prompts when agent behavior deviates from expectations?▼

Refining deviating agent behavior requires dispatching a subagent to execute the target prompt, collecting objective bidirectional feedback through self-reports and instruction-side metrics, and applying the differences iteratively until desired reliability is reached.

How do I set up bias-free agent evaluation for prompt engineering?▼

Setting up bias-free agent evaluation requires a dispatchable subagent workflow, a clear requirements checklist for execution measurement, and a structured report format to capture tool usage metrics and self-reported observations.

When do I need to use subagent prompt evaluation in my workflow?▼

Subagent prompt evaluation is needed when creating new prompts, heavily revising existing ones, or when agent behavior deviates from expectations, to surface ambiguities and drive focused improvements until reliability is achieved.

Can I use iterative prompt tuning to measure execution against a requirements checklist?▼

Iterative prompt tuning measures prompt execution by dispatching a subagent against a predefined requirements checklist, pairing executor self-reports with instruction-side metrics to surface ambiguities and apply focused improvements systematically.

Related Skills

Explore More Software Engineering Skills→
ResearAIResearAI
3.3k

baseline

Standardize a reference object for quest evaluation with plan and checklist requirements.

Community
Advanced
doodledooddoodledood
17

author-voice

Create and refine AUTHOR_VOICE.md documents through iterative feedback cycles.

Community
Intermediate
gmliaogmliao
1

subagent-driven-development

Dispatch a fresh subagent per task with two-stage spec and code reviews.

Community
Intermediate
sori883sori883

empirical-prompt-tuning

Evaluate and iteratively refine agent prompts using executor feedback and system metrics.

Community
Advanced