How does the OCA cycle work?

The OCA cycle stands for Observe (collect data), Codify (document patterns), and Automate (convert methodologies into automated checks or tools), creating a continuous loop of improvement.

When should I avoid using this skill?

Avoid using this skill for simple, one-time ad-hoc tasks or trivial processes that don't require reusability or systematic standardization.

What does the knowledge-extractor subagent do?

Once a methodology experiment converges, the knowledge-extractor automatically transforms the results into reusable Claude Code skills, patterns, and principles, saving hours of manual documentation.

BAIME (Bootstrapped AI Methodology Engineering) is a unified framework for LLM-based development that integrates the OCA cycle, empirical validation, and dual-layer value functions to build systematic engineering methodologies.

What are dual-layer value functions?

They are quantitative metrics used to evaluate an iteration: V_instance measures domain-specific task quality (like test coverage), while V_meta measures how reusable and complete the methodology itself is.

Methodology Bootstrapping (BAIME)

Name: Methodology Bootstrapping (BAIME)
Author: yaleh

byyaleh

•

Productivity & Workflow

Develops and validates systematic software engineering methodologies using iterative Observe-Codify-Automate cycles.

Methodology Bootstrapping leverages the Bootstrapped AI Methodology Engineering (BAIME) framework to evolve ad-hoc development practices into structured, automated, and reusable methodologies. By utilizing the OCA (Observe-Codify-Automate) cycle, this skill helps developers create robust testing strategies, CI/CD pipelines, and observability systems with empirical validation. It employs dual-layer value functions to measure both immediate results and long-term transferability, ensuring that every workflow improvement is data-driven, scalable, and optimized for AI-assisted development.

Key Features

01Iterative OCA (Observe-Codify-Automate) cycle for systematic process evolution

0215 GitHub stars

03Dual-layer value functions to measure both task quality and methodology reusability

04Specialized subagents for automated prompt design, iteration execution, and knowledge extraction

05Empirical validation framework using data-driven convergence criteria

06Automated transformation of experiment findings into distributable Claude Code skills

Use Cases

01Building standardized high-coverage testing strategies for complex codebases

02Developing reusable error recovery, observability, and technical debt management patterns

03Optimizing CI/CD pipelines through systematic observation and bottleneck removal

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yaleh/meta-cc methodology-bootstrapping

For use in Claude.ai and ChatGPT

Download Skill