claude-code-features

Commands handle simple, manual, repeatable tasks via slash commands. Skills provide contextual domain knowledge for complex guidance. Hooks automate deterministic validation or logging around tool usage. Choose Commands for fixed workflows, Skills for knowledge-heavy decisions, and Hooks for event-driven enforcement.

Skills auto-load contextual domain knowledge into the current conversation without polluting context. Agents are independent sub-agents that handle large tasks and data processing separately. Use Skills for guidance and Agents to isolate heavy computation or large datasets.

Rules are modular constraints and policies applied conditionally or globally to enforce standards across workflows. Hooks are event-driven shell automation that runs before or after tool usage for validation and logging. Use Rules for policy constraints and Hooks for deterministic automation triggers.

Yes, Commands inject explicit, manual, repeatable prompts via slash commands for fixed workflows. They're ideal for standardized tasks you run frequently without needing decision-making or complex reasoning.

Hooks provide event-driven shell-based automation that executes before or after tool usage for validation, logging, or policy enforcement. They enable deterministic automation without modifying core conversation flow.

Rules apply modular constraints and policies conditionally or globally to enforce standards. Skills provide contextual domain knowledge. Use Rules for project-wide constraints and Standards, Skills for complex guidance and decision support.