What problem does it solve? AEM as a Cloud Service projects accumulate code-quality and correctness issues — outdated Maven dependencies, discouraged @Inject usage in Sling Models, unbounded queries, missing HTTP timeouts, and legacy listener/scheduler patterns — that cause build failures, OOM outages, and silent runtime bugs. This Skill detects and fixes those issues entirely against the local workspace, with no external services or network calls. ## Core Features & Use Cases - Deterministic local detection: A bundled Java analyzer (scripts/analyze.sh) parses the workspace once and runs per-pattern detectors, emitting structured JSON findings with file, line, and snippet. - Expert pattern skills: Routing to specialized remediation guides covering outdated Maven dependencies, @Inject to injector-specific annotation migration, unbounded query triage, outbound call timeouts, Sling Scheduler, ResourceChangeListener, event migration, replication, and AssetManager modernization. - Surgical, reviewable fixes: Applies minimal edits on a git branch or in place, verifies with mvn compile, never commits or pushes, and enforces one pattern per session with resumable batching for large migrations. - Use Case: Ask it to "scan this AEM project for issues" — it runs every detector, reports findings across all built patterns, then applies a chosen fix (e.g., modernizing Sling Models) after you confirm. ## Quick Start Ask the assistant to scan this AEM project for code-quality issues and report what it finds.