af-fix-by-benchmarks

Identify and fix skill failure root causes using benchmark-driven analysis.

3|Updated Oct 5, 2025

Search Tags:#automation #verification #benchmarking #subagent #root-cause-analysis #skill-improvement

Ready to run

npx skills add https://github.com/korchasa/ide-rules --skill af-fix-by-benchmarks

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: af-fix-by-benchmarks
Source: https://github.com/korchasa/ide-rules/tree/main/.cursor/skills/af-fix-by-benchmarks
Command: npx skills add https://github.com/korchasa/ide-rules --skill af-fix-by-benchmarks

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill provides a structured, benchmark-driven approach to diagnosing and fixing failures in AssistFlow skills, ensuring issues are identified and mitigated with evidence-based reasoning.

Core Features & Use Cases

Systematic benchmarking workflow to identify root causes of skill failures.
Guided analysis via a benchmark-runner to reproduce, observe, and document failures.
Actionable fixes with argumentation and verification plans to ensure stability.

Quick Start

Run the benchmark for the target skill and review the benchmark-runner results to begin root-cause analysis.

Frequently Asked Questions about af-fix-by-benchmarks

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

FAQPage Schema

How do I identify the root cause of failing skill benchmarks?▼

Root cause analysis for failing skill benchmarks requires executing benchmark scenarios to reproduce failures and collecting evidence from the results to pinpoint the underlying issue.

What is benchmark-driven verification for skill failures?▼

Benchmark-driven verification for skill failures uses a structured workflow to execute scenarios, collect evidence, and propose fixes with argumentation and verification plans to ensure stability.

How do I fix failing skills using benchmark scenarios?▼

To fix failing skills using benchmark scenarios, execute the benchmark runner, collect evidence from the results, and present a proposed fix with argumentation and a verification plan for user approval.

Can I automate skill failure diagnosis and verification?▼

You can automate skill failure diagnosis using a subagent workflow that executes benchmark scenarios, collects evidence, and generates verification plans to ensure stability.

Does the benchmark-driven workflow require approval before applying fixes?▼

Yes, the benchmark-driven workflow pauses for user approval before implementing changes, presenting findings with a proposed fix, argumentation, and a verification plan.