What problem does it solve? Teams generating AI skills with tools like /learn need a consistent way to verify those skills conform to the Synthos five-section standard, contain no leaked credentials, and migrate cleanly across a dual-repository architecture. This Skill automates that audit so non-conforming or unsafe skills are caught before deployment. ## Core Features & Use Cases - Four-block to five-section mapping: Maps /learn-generated skill blocks (思想/原理/流程/方法) to the Synthos five-section method and records invocation hierarchy differences (P2 vs P0-P1). - Privacy scan grading: Greps skill content for API keys, tokens, and passwords, then remediates findings via environment-variable substitution and re-scans. - Path trap and dual-repo audit: Detects broken symlinks, repairs permissions, and verifies migration coverage across skills/private/ and skills/ directories. - Use Case: Given three private skills pending migration, produce a coverage report showing 2/3 migrated, one blocked by a hardcoded API key with the exact grep location and remediation steps. ## Quick Start Audit the /learn-generated xhs-content skill against the Synthos five-section standard and report mapping results, hierarchy differences, and privacy scan findings.
