Enforces disciplined skill discovery and systematic workflow management for all Claude Code interactions.
The using-harness skill serves as the foundational governance layer for the Harness framework within Claude Code. It establishes a strict protocol requiring the AI to check for and invoke relevant specialized skills before responding to any user prompt or starting a task. By providing a structured framework for skill priority, anti-rationalization checks, and standardized user communication via the AskUserQuestion tool, it ensures that best practices and domain-specific patterns are consistently applied while maintaining project integrity through automated backlog management.
Key Features
01Priority-based execution ordering for process and implementation skills
020 GitHub stars
03Mandatory skill discovery protocol for all tasks
04Red-flag detection to prevent skipping disciplined workflows
05Automated tracking of deferred items in .artifacts/ directories
06Structured AskUserQuestion integration to reduce user cognitive load
Use Cases
01Maintaining a persistent project roadmap by automatically capturing out-of-scope ideas during development.
02Standardizing communication during complex architectural decisions using structured selection tools.
03Initializing a new project phase to ensure all relevant domain skills are identified and loaded.
