Does this skill modify core system files?

No. Halmoni includes a 'Remember' phase that rejects any changes to protected components like bead schemas, command names, or integration protocols.

How does Kimchi Halmoni track historical changes?

It automatically performs patch version bumps in YAML frontmatter, creates entries in a skills/CHANGELOG.md file, and generates automated git commits.

Can I review changes before they are applied to my code?

Yes, using the --dry-run flag allows you to see all proposed adjustments without applying them, while --taste-only provides observations without proposals.

What is the primary purpose of Kimchi Halmoni?

It serves as a self-improvement system that observes execution outcomes or user feedback to propose and apply incremental improvements to the AI's skills and personas.

Kimchi Halmoni

Name: Kimchi Halmoni
Author: Tromml

byTromml

•

Productivity & Workflow

Implements a self-optimizing feedback loop for AI agents by observing execution results and proposing incremental improvements to core skills and personas.

Kimchi Halmoni is an autonomous refinement engine designed to evolve the Kimchi framework through continuous observation and iteration. By analyzing 'bead' executions, git logs, or direct user feedback, it identifies performance gaps and proposes precise, version-controlled adjustments to skills, validators, and personas. The system mimics a master artisan's process—taste, adjust, and remember—ensuring that your AI's behavior becomes more reliable and sophisticated over time while protecting core architectural integrity.