Codify behavioral guidelines for code agents to enforce cautious, goal-driven execution.
npx skills add https://github.com/fatihkan/badi --skill badi-discipline
Please help me install this Agent Skill. Skill: badi-discipline Source: https://github.com/fatihkan/badi/tree/main/_bootstrap/badi-skills/skills/badi-discipline Command: npx skills add https://github.com/fatihkan/badi --skill badi-discipline
Behavioral guidelines, distilled from Badi's 21 subagents, 12 hooks, and CLAUDE.md, to guide AI coding agents toward cautious, goal-driven execution and minimize scope creep by enforcing explicit source citations and safe practices.
Follow the eight principles before starting a coding task to ensure cautious, goal-driven execution.
High-intent search queries and answers about installing and using this skill.
Yak-shave detection in coding tasks is an approach that identifies and stops unnecessary prerequisite chains before they derail your main objective. It enforces task discipline by surfacing side-issues separately instead of allowing scope creep.
To enforce task discipline in AI code reviews, follow structured principles like task-board discipline and simplicity first. This reduces wasted effort, improves traceability, and ensures the agent surfaces side-issues separately rather than expanding scope.
A destructive action gate in coding is a mandatory safety check that prevents AI agents from executing harmful actions without explicit approval. It enforces cautious execution by requiring safety checks before destructive changes are applied.
To apply structured task discipline to AI coding, start by following eight guiding principles including think before coding, simplicity first, and surgical changes. These principles enforce cautious, goal-driven execution before starting any feature development or refactoring task.
microsoftCreates a shadow knowledge base from source code and markdown files for analysis and understanding.
tiankong0101-byteEnforce cautious assumptions, minimal implementations, and precise code changes in AI coding workflows.
2300031147Enforce safety boundaries for destructive commands and directory-scoped edits.
kmshihab7878Scaffold new projects with Claude Code configuration for multiple stacks.
