NeoLabHQ/context-engineering-kit v2.2.0

on GitHub

What's Changed

SADD plugin:

• Add a meta-judge and judge agent that together implement the LLM-as-a-Meta-Judge pattern. Meta judge generates high-quality checklists and rubrics that essentially work as a target specification, against which the judge is capable of reviewing task results
• Integrate the meta-judge and judge in the majority of SADD plugin workflows

DDD plugin:

• Add rules with Clean Architecture, DDD, SOLID, Functional Programming, and other patterns examples - will be automatically added to the context, when
• removed software-architecture skill - probably never was invoked by agent

Code Review plugin:

• Update local review command with issues impact/confidence filtering to eliminate halicunations
• Add support for --min-impact param to pr-review and local-review commands and set it by default to high

Customize Agent Plugin:

• Add create-rule command

MCP plugin:

• fix: replace unsupported --version flag with -help in codemap setup skill by @acsezen in #63

New Contributors

• @acsezen made their first contribution in #63

Full Changelog: v2.1.1...v2.2.0