What problem does it solve? AI coding agents repeatedly rediscover project constraints on every request because there is no maintained, routable project memory. This Skill converts finalized documentation into a structured .Knowledge base so agents load only the relevant topics before editing code. ## Core Features & Use Cases - Topic Generation: Splits finalized stock-docs into focused topic files under .Knowledge/topics with summaries, rules, and boundaries. - Routing Manifest Maintenance: Updates manifest-routing.json with topicPaths, taskToTopicRules, topicDependencies, and topicMetadata, plus keyword matcher shards. - Index Synchronization: Keeps .Knowledge/index.md consistent with one row per topic and clickable related-document links. - Orchestration Rules: Enforces main/sub-agent write boundaries so only the main agent writes the manifest and index. - Use Case: After finalizing an architecture document with f2s-doc-final, run this Skill to produce routable topics and matchers so future sessions match requests like "batchReScore sync or async" directly to the right topic. ## Quick Start Run f2s-kb-build with the path to a finalized document in .Knowledge/stock-docs to generate the corresponding topics, index entries, and routing matchers.
