For anyone managing an evolving knowledge base in Omni Apps, transforming disparate notes into a structured, queryable system is a common need. The schema-author skill is designed for this purpose. This AI agent evolves your brain's type taxonomy by adding page types, link verbs, and prefixes, systematically turning unstructured notes into a queryable system where content becomes discoverable through schema-aware routing.
It helps organize large bodies of information that might have grown organically over time, providing a path to greater clarity and utility for your content.
How the Skill Operates
The schema-author skill executes its taxonomy refinement process through a sequence of seven distinct phases, ensuring a thorough and controlled evolution of your brain's structure. First, it confirms the active pack to understand the current context of operations. Second, it assesses current coverage by reviewing existing schema stats, providing a baseline of your brain's organization. Third, the agent detects orphaned pages—content that lacks defined types—identifying areas ripe for classification.
Next, the skill uses ML refinement to propose new types based on the identified needs and patterns within your corpus. This data-driven approach suggests page types, link verbs, and prefixes that best fit your content. Once proposals are generated, it moves to apply the mutations, integrating the new schema elements into your brain's structure. These changes are performed with atomic and idempotent operations, meaning each operation completes entirely or not at all, and can be repeated without causing unintended effects.
Following the schema modification, it proceeds to backfill existing pages via sync, assigning the newly defined types to relevant content. This ensures that previously untyped or ambiguously typed pages are brought into the new taxonomy. Finally, the process concludes by verifying the results, confirming that the new schema has been correctly applied and the content is accurately categorized. This methodical approach ensures robustness and accuracy in schema evolution.
Applying Schema-author
This agent is designed to trigger on specific requests and reports. For instance, you might initiate an operation with a simple request like 'add a researcher type', prompting the skill to propose and integrate a new page type. It also responds to diagnostic reports such as '4000 untyped pages under meetings/', which indicates a significant area where structured types are needed. Furthermore, you can prompt the skill with 'propose types from my corpus', allowing it to analyze your existing content and suggest suitable classifications.
Operations include schema add-type for direct type creation, lint for reviewing proposed changes, sync --apply for committing backfills, and batch mutations for larger-scale modifications. For local CLI usage, the system assumes trust, enabling straightforward execution of commands. When operating remotely through an MCP (Multi-Cloud Platform), OAuth admin scope is required to ensure secure and authorized access for schema modifications.
What It Is Not For
It is important to clarify the boundaries of the schema-author skill to ensure effective use. This tool is not for filing individual pages; for that purpose, you would use brain-taxonomist, which is designed for granular content classification. The skill also does not perform schema validation inside another iteration loop; its focus is on evolving the schema itself, not on validating content against an existing, stable schema repeatedly within a workflow.
Furthermore, this agent is not intended for querying expert knowledge directly. Its function is to structure the knowledge within your brain, making it discoverable through schema-aware routing, rather than serving as an oracle for direct answers. Understanding these distinctions helps maximize the utility of the skill within your Omni Apps environment.
FAQ
Q: What core problem does this skill address? A: It addresses the challenge of evolving an unstructured knowledge base into a queryable, organized system by refining the brain's type taxonomy.
Q: How does the skill ensure changes are safe? A: All operations are atomic and idempotent, meaning they are either fully completed or not started, and can be rerun without unintended side effects.
Q: Can I use it to classify just one page?
A: No, it is not designed for filing individual pages. For that task, you should use the brain-taxonomist skill.
The schema-author skill provides a robust method for enhancing the discoverability of your content by introducing structural organization. It serves as a practical solution for maintaining a coherent and queryable knowledge base.
