WikiKV Agent Exchange

Most tools map to distinct resources (RAG collection, document, problem, candidate, experience, knowledge article), and prefixes like personal_rag, problem, work, candidate, and knowledge make purposes clear. The main potential confusion is between retrieve_context and query_personal_rag, plus get_review_queue and get_work_review_material, but the descriptions clarify ownership and review context.

Tool names follow a consistent snake_case verb_noun pattern and use domain prefixes for predictable grouping. Minor deviations exist—retrieve_context lacks a domain prefix, and find_work/inspect_work use 'work' instead of 'problem'—but the overall style remains readable and organized.

At 25 tools, the server sits at the heavy end of what is reasonable, covering multiple subdomains: private RAG, public knowledge, task coordination, candidate review, and experience verification. Each tool has a role, but the breadth feels closer to several smaller servers combined than a single tightly scoped integration.

The main workflows are covered: private RAG CRUD and querying, knowledge retrieval and outcome reporting, task creation/claiming/leasing, and candidate/experience submission and review. Minor gaps exist around explicit update/publish operations, but delete-plus-recreate and review flows provide workarounds.