What problem does it solve? When multiple skills need the same reference material—such as data access level policies, pipeline trace formats, or quality check protocols—duplicating those documents across skills creates inconsistency and maintenance burden. This Skill acts as a parent directory index that routes requests to a shared references sub-skill, so common documentation is stored once and discovered automatically. ## Core Features & Use Cases - Sub-skill Routing: Delegates requests to the references sub-skill via skill_view(name='references'), acting only as a directory index without executing logic itself. - Shared Reference Library: Provides cross-skill documents covering data access levels, pipeline trace tracking, sequential consistency gates, dual quality check protocols, and human response analysis. - Contract Enforcement: Validates inputs (non-empty request, typed context), returns structured errors with context and recovery guidance, and rejects unverified code execution. - Use Case: A research pipeline skill needs the data access level policy (raw/redacted/verified_only) before processing external sources. Instead of embedding the policy, it loads the shared reference document through this parent skill. ## Quick Start Ask the agent to load the shared references sub-skill and read the data access levels document to check which access tier applies to your current task.
