Coordinates Ze repository design decisions grounded in repo rules and maintains a running decision log.
npx skills add https://github.com/ze-software/ze --skill ze-design
Please help me install this Agent Skill. Skill: ze-design Source: https://github.com/ze-software/ze/tree/main/.codex/skills/ze-design Command: npx skills add https://github.com/ze-software/ze --skill ze-design
Ze design coordinates decisions for repository planning and ensures all recommendations stay aligned with existing constraints and code patterns, reducing drift and rework.
Propose a design approach for the requested feature in Ze that strictly follows repo rules and existing code patterns, documenting each decision in a running log.
High-intent search queries and answers about installing and using this skill.
You can ground architectural design decisions in repository rules by binding every recommendation to configuration files like CLAUDE.md and AGENT.md. This ensures alignment with existing constraints and code patterns, reducing drift and rework.
Maintaining a decision log for feature design reviews involves documenting each architectural choice in a persistent running log. This provides traceability for policy alignment and allows reuse of past decisions in future specifications.
Break down complex architectural discussions by evaluating one decision at a time with explicit trade-offs. This stepwise approach ensures each recommendation is verifiable and grounded in established repository constraints.
Yes, you can apply repo guidance to evaluate existing code patterns before proposing a feature design. It checks recommendations against applicable rules within the codebase to ensure new features do not violate established policies.
Repository planning experiences drift when recommendations are not strictly bound to applicable rules and existing code patterns. Applying grounded design reviews with a persistent decision log resolves this by enforcing policy alignment.
You should avoid a stepwise approach for architectural design reviews when immediate, holistic decisions are required without explicit trade-off analysis, as this method specifically breaks problems into single, testable decisions.
mgratzerCreate CLAUDE.md, AGENTS.md, docs/, README, and CHANGELOG.md for Forge projects.
mymx2Enforce a specification and planning phase before code implementation.
VoDaiLoczAutomate linting and static analysis of code in CI pipelines.
quangphu1912Detects circular dependencies, architectural violations, and conflicts in module graphs via static code analysis.
