No user-submitted related servers found.
Scored across 22 tools
The traversal tools (find_references, graph_neighborhood, trace_data_flow, trace_path) and the search pair (semantic_search vs semantic_locate) are explicitly cross-referenced and distinguished by their descriptions, which is excellent. However the mutation surface overlaps: kin_mutate ('atomically validate and commit a batch') appears to duplicate what kin_transaction_begin/stage/commit do, and no description reconciles the two paths, creating a misselection risk.
All names use snake_case and mostly follow verb_noun (find_references, list_file_entities, trace_data_flow), which is readable. But the kin_ prefix is applied inconsistently—about half the tools carry it (kin_session_start, kin_transaction_begin, kin_mutate) and half do not (get_entity_source, impact_analysis, semantic_locate)—and a few are bare noun phrases (impact_analysis, graph_neighborhood), breaking the pattern.
At 22 tools the surface is on the heavy side of the 16-25 range for a single code-graph server. Much of the count is justified by genuinely distinct concerns (search, retrieval, tracing, sessions, transactions, artifacts), but the coexistence of both kin_mutate and the four-tool transaction family suggests a redundant mutation pathway that inflates the count.
Coverage is broad and lifecycle-complete: search, retrieval, traversal, impact analysis, provenance, atomic mutations with create/update/delete/rename, plus session and graph-status introspection leave few obvious dead ends. Minor gaps remain around review/approval workflows and multi-endpoint tracing beyond single-source walks, but core workflows are covered.