curve-mcp

Many tools have overlapping purposes, causing significant ambiguity. For example, there are multiple tools for retrieving pools (get_pools_all, get_pools_big, get_pools_small, get_pools_empty) that differ only by TVL filters, and multiple tools for supply data (get_crvusd_total_supply, get_crvusd_total_supply_number) that differ only by output format. The descriptions help somewhat, but the boundaries between tools are often unclear, leading to potential misselection.

The naming follows a mostly consistent get_verb_noun pattern with snake_case throughout, which is predictable and readable. There are minor deviations, such as get_get_all_gauges (redundant 'get_get') and inconsistent use of underscores in some names (e.g., get_facto_gauges_crv_rewards_blockchain_id), but overall the convention is maintained across the set.

With 37 tools, the count is excessive for the apparent scope of querying Curve protocol data. Many tools appear to be variations of the same core operations (e.g., multiple pool retrieval tools with different filters), creating redundancy. This heavy toolset will likely overwhelm agents and complicate selection without adding proportional value.

The tool surface appears comprehensive for querying Curve protocol data, covering pools, gauges, volumes, APYs, supplies, and other metrics across chains and registries. There are minor gaps, such as limited write operations (e.g., no tools for interacting with contracts), but for a read-only API, it provides broad coverage of the domain with few dead ends.