What problem does it solve?
Google Drive documents contain valuable organizational knowledge—policies, definitions, and operating procedures—but that knowledge stays locked in raw docs unless someone reads, deduplicates, and distills it into a durable, searchable knowledge base.
Core Features & Use Cases
- Doc-to-Wiki Synthesis: Reads staged Google Doc raw files (page.md plus metadata.json) and writes a small set of high-quality KTX wiki pages capturing durable business knowledge.
- Deduplication and Conflict Handling: Searches existing wiki pages before writing, updates overlapping pages instead of creating duplicates, and records conflict notes when sources disagree.
- Identifier Verification: Verifies warehouse, dbt, Looker, Metabase, or MetricFlow references via discovery, entity details, and SQL probes before citing them, marking unverifiable identifiers explicitly.
- Eviction Management: Processes deleted raw paths through eviction decisions so stale knowledge is retired deliberately.
- Use Case: A memory agent receives a WorkUnit containing a Google Doc describing the company's revenue recognition policy; the skill reads the doc, finds the existing revenue wiki page, updates it with the new policy, and cites the doc title and last-modified date.
Quick Start
Load this skill when a WorkUnit contains staged Google Drive raw files under docs/** and synthesize the durable knowledge into KTX wiki pages.