What problem does it solve?
Many agents proceed to act or ask clarifying questions without first checking whether a dedicated Cadence skill applies, leading to inconsistent workflows and missed safeguards. This Skill ensures agents always check for and invoke relevant Cadence skills before any response or action, establishing disciplined, discoverable behavior across conversations.
Core Features & Use Cases
- Mandatory Invocation Rule: Require invoking the Skill tool (or loading SKILL.md) whenever there is even a small chance a skill applies, before any reply or clarifying question.
- Priority & Safety Rules: Define precedence so user explicit instructions override skills, and skills override the default system prompt.
- Platform Adaptation: Map Claude Code tool names to Codex equivalents and explain how to spawn subagents, wait for results, and close agents.
- Use Cases: Starting a conversation to determine workflows, entering planning or implementation phases, dispatching subagents, and ensuring TDD or debugging skills are applied when appropriate.
Quick Start
Invoke the using-cadence skill before answering any user message to check for and load applicable Cadence skills.