Scored across 79 tools
Many tools overlap in retrieval, consolidation, and evaluation purposes: memory_recall, memory_search_fast, memory_explain_search, memory_recall_iterative, memory_associate, memory_context_build, and multiple memory_eval_* tools have unclear boundaries. Fast-path duplicates (memory_save_fast vs memory_save) and several stats/status/report tools further blur selection.
Names are mostly snake_case with clear categorical prefixes (memory_, self_, kg_, workflow_, session_), which is predictable. Minor inconsistency appears in prefix+part-of-speech patterns, e.g. memory_history (noun) vs memory_save (verb), but overall the convention is readable.
With 79 tools, the server is far beyond the ideal 3-15 range and exceeds the 50+ threshold for extreme mismatch. Many tools are redundant variants (fast paths, eval suites, debug utilities) that could be consolidated without losing core functionality.
Coverage of the memory lifecycle is extensive: save, get, update, delete, recall, relate, export, statistics, rules, graph, skills, workflows, and sessions are all present. A minor gap is the lack of an import counterpart to memory_export, and some overlapping query variants add noise rather than fill gaps.
