No user-submitted related servers found.
Scored across 89 tools
The set contains multiple clusters of overlapping tools: shared_memory_search/list/forget, memory_recall, recall_relevant, memory_graph_query, vault_search, and entroly_retrieve all involve memory retrieval; verify_provenance, verify_and_repair, verify_response, eicv_verify_claim, and eicv_suppress_hallucinations all verify generated content; record_outcome/test/command/CI/edit tools all record feedback. Descriptions distinguish them somewhat, but the volume and overlap make misselection likely.
Most names are snake_case, but the conventions are mixed: verb_noun (get_stats, create_skill), noun_verb (memory_recall, vault_query), prefixed namespaces (work_*, vault_*, shared_memory_*), and irregular names like smart_read and entroly_dashboard. The pattern is readable but not predictable across the full set.
Eighty-nine tools is far beyond the appropriate range for a single MCP server and creates an overwhelming surface. The server spans many domains—context compression, memory, vault beliefs, verification, security, work sessions, skills, and operations—each with multiple overlapping operations.
The surface is broad and covers many workflows across context optimization, memory, belief vault, verification, and work state, so it is not severely incomplete. However, there are notable gaps and inconsistencies, such as limited update/delete lifecycle for primary context fragments and a sprawling, non-uniform set of operations rather than complete CRUD per domain.