expert-lookup

What problem does it solve? When an AI agent takes over a business module or troubleshoots a domain issue, it often re-reads code from scratch even though curated expert knowledge packages already exist. This Skill locates those pre-built expert assets (generated by expert-team and stored in .module-experts/) so prior knowledge is reused instead of rediscovered. ## Core Features & Use Cases - Keyword-based expert discovery: Matches task keywords against INDEX.md entries (including sub-experts and topics) across project-level and user-level stores, then verifies hits by actually reading agent.md or topic.md. - Intent-based layered loading: Loads contract-layer docs (C0-C4) for usage tasks, implementation docs plus source code for diagnosis or modification tasks, and T0 topic overviews for cross-expert architecture questions. - Multi-expert collaboration: Loads up to three experts simultaneously in collaborative or parallel mode when a task spans multiple business domains. - Use-driven maintenance: Performs restricted incremental updates when contracts drift from code, auto-appends new matching keywords to INDEX.md, and checks asset freshness via git baseline commit comparison. - Use Case: You ask "why isn't the alert triggering and how do I query the logs" — the Skill matches both the alert-processing expert and the log-query expert, loads their contract layers, and solves the problem from the experts' perspective. ## Quick Start Ask the AI "is there an existing expert for the alert module" or "find the expert for this module" to trigger a lookup and load the matching expert assets.