What problem does it solve? AI agents working on a project often lack awareness of which skills and project conventions are available, leading to missed context and inconsistent behavior. This Skill scans user and project skill directories plus convention files to build a single centralized registry that orchestrators and sub-agents can load on demand. ## Core Features & Use Cases - Skill Discovery: Scans user-level directories (Claude, Gemini, Cursor, OpenCode, Antigravity) and project-level directories for SKILL.md files, extracting their triggers. - Convention Indexing: Detects project convention files such as AGENTS.md, GEMINI.md, .cursorrules, and .agent/rules, following referenced files recursively. - Flexible Persistence: Writes the registry to .juar/skill-registry.md and optionally persists it to Engram memory depending on the artifact storage mode (engram, openspec, hybrid, none). - Use Case: After installing new skills in a project, run this Skill so the orchestrator and all sub-agents immediately know which skills exist, their triggers, and where to load them from. ## Quick Start Ask the agent to scan all skill directories and project conventions and generate the centralized skill registry index in .juar/skill-registry.md.
