skills.rest
rakibulism
rakibulismCOMMUNITY
Explore all skills by this author →
Website𝕏@rakibulismGitHub

skill-creator

Create and refine agent skills with evals and benchmark loops.

2|1|Updated May 17, 2026
Search Tags:#prompt engineering#skill creation#benchmarking#iteration#evals#trigger optimization
One-click install
npx skills add https://github.com/rakibulism/agent-skills-os --skill skill-creator-rakibulism
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/rakibulism/agent-skills-os/tree/main/skills/skill-creator
Command: npx skills add https://github.com/rakibulism/agent-skills-os --skill skill-creator-rakibulism

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and assets (resource) components.

What problem does it solve?

This Skill helps you design, draft, test, evaluate, and refine new skills so they trigger reliably and produce better results over time.

Core Features & Use Cases

  • Skill creation: Turn an idea or workflow into a structured SKILL.md with strong triggering guidance.
  • Testing and evaluation: Build realistic eval prompts, run comparisons, grade outputs, and review benchmark results.
  • Iterative improvement: Use user feedback, transcripts, and performance data to revise instructions, add tools, and reduce failure modes.
  • Description optimization: Improve the frontmatter description to increase the chance the right skill activates in the right context.

Quick Start

Use the skill-creator skill to help me design, test, and improve a new skill for my workflow.

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I create and test agent skills with reliable trigger optimization?▼

You create and test agent skills by drafting structured instructions, generating eval prompts, and running benchmark loops to optimize trigger descriptions. This skill handles the full creation cycle by turning workflows into self-contained skills with strong triggering guidance.

Why does my skill underperform across realistic prompts and edge cases?▼

Your skill underperforms when instructions lack rigor or trigger descriptions are weak. This skill diagnoses failure modes by running comparison loops, grading outputs, and applying iterative revision using transcript data and feedback review.

What is the best way to improve skill trigger descriptions for better context activation?▼

The best way to improve trigger descriptions is to optimize frontmatter metadata through benchmark testing. This skill refines descriptions by testing activation reliability across realistic prompts and edge cases, ensuring the right skill triggers in the right context.

Can I use prompt engineering to refine an existing skill and reduce failure modes?▼

Yes, you can refine an existing skill by applying prompt engineering techniques to revise instructions and add tools. This skill supports iterative improvement by leveraging user feedback, transcripts, and performance data to reduce failure modes.

Do I need test-set creation scripts to evaluate skill performance before deployment?▼

You need test-set creation scripts to build realistic eval prompts and grade outputs effectively. This skill provides scripts and reviewer tooling to automate test-set generation, run grading, and aggregate benchmark results for evaluation.

Related Skills

Explore More Education & Research Skills→
jdonohoojdonohoo
15

new-idea

Create a standardized discovery folder structure with input and output directories.

Community
Basic
vnmoorthyvnmoorthy
2

canary

Monitor live apps for console errors and performance regressions post-deploy.

Community
Advanced
NodeJSmithNodeJSmith
1

mine.skill-eval

Automate evaluation and comparison of AI skill variants via structured grading and statistical analysis.

Community
Advanced
CeeKay1901CeeKay1901

skill-creator

Draft, test, and refine reusable skills with evaluation and trigger optimization.

Community
Advanced