skills-auto-activation

Auto-activation failures often stem from weak keyword matching or missing context. Use a hook-based activation system with a skill-rules.json catalog that applies keyword, intent, file, and content triggers to reliably activate skills across large projects and multi-repo environments where standard cues fail.

Yes. A custom hook-based activation pipeline lets you define deterministic rules in skill-rules.json that trigger based on keywords, detected intent, file patterns, or content analysis, ensuring skills load automatically even when keyword cues fail or in cross-session workflows.

Configure explicit activation rules and a CLAUDE.md reference, or set up a custom hook system with a configurable skill-rules.json catalog. This enforces deterministic activation across multi-repo environments with a maxSkills cap and optional debugging support.

The pipeline reads skill-rules.json to match incoming prompts against keyword, intent, file, and content triggers. When a match occurs, it activates the corresponding skill, respecting a maxSkills limit. This approach scales reliably in large projects where implicit activation often fails.

Yes. The hook system includes optional debugging support that logs trigger evaluation and activation decisions, helping you identify whether keyword, intent, file, or content rules are matching correctly and refine your skill-rules.json configuration.